Adoption of the Semantic Web for overcoming ... - Semantic Scholar
Expert Systems with Applications 36 (2009) 7318–7327
Contents lists available at ScienceDirect
Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa
Adoption of the Semantic Web for overcoming technical limitations of knowledge management systems Jaehun Joo a,*, Sang M. Lee b a b
Department of Information Management, Dongguk University, 707 Sukjang-dong, Gyeongju, Gyeongsangbuk-do 780-714, South Korea Department of Management, CBA 209, University of Nebraska-Lincoln, Lincoln, NE 68588-0491, USA
a r t i c l e
i n f o
Keywords: Semantic Web Knowledge management Limitation factors IT innovation Innovation adoption
a b s t r a c t New information technology is a core factor enabling innovation in knowledge management (KM). The purpose of this study is to analyze the limitations of current KM systems and to propose an approach for applying the Semantic Web to KM. We analyzed the factors that affect KM system user dissatisfaction through a survey. We found that inconvenience, search and integration were statistically significant limitation factors for system quality. On the other hand, incongruence and untrustworthiness of knowledge were significant limitation factors for knowledge quality. Finally, we suggest methods for applying the Semantic Web to KM as an alternative overcoming the technical limitations of current KM systems. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Knowledge management (KM) has been recognized as one of the most critical factors for obtaining organizational competitive advantage (Antoniou & Harmelen, 2004). New advances in information technology (IT) have supported innovations in KM. In other words, IT innovation has become a crucial factor for survival and success of organizations in the knowledge age. The Semantic Web is one of the technologies driving a paradigm shift in KM as well as supporting electronic commerce and Web services (Antoniou & Harmelen, 2004; Davies, Fensel, & Harmelen, 2003). The Semantic Web as an extension of the current Web provides information in a well-defined manner, enabling computers and people to work in cooperation (Bernes-Lee, Hendler, & Lassila, 2001). The Semantic Web handles machine-processable information, which enables communication between machines without human intervention. One very promising application area of the Semantic Web is KM. Fig. 1 shows research areas on the Semantic Web for KM. The studies are classified into four areas: developing infrastructure and architecture, killer applications, business management issues and other social issues. Most surprising research area is technical issues related to architecture design and building of infrastructure for knowledge management based on the Semantic Web (Bernes-Lee et al., 2001; D’Aquin et al., 2005; Tiwana & Ramesh, 2001). Recently, a
* Corresponding author. Tel.: +82 54 770 2346. E-mail addresses: [email protected] (J. Joo), [email protected] (S.M. Lee). 0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2008.09.005
few projects targeting killer applications as well as building infrastructure, have been conducted. The on-to-knowledge project as a typical example aimed at developing an ontology-based tool suite that efficiently processes large numbers of heterogeneous, distributed, and semistructured documents (Davies et al., 2003). Although the project contributed to building an infrastructure for applications of the Semantic Web to KM as well as development of tools, this project did not deal with business issues associated with considering part of strategic relationships between IT and management. As Orlikowski and Iacono (2001) pointed out, it is necessary to study the complex ensemble of people, culture, and technology embedded in social contexts as well as a specific organizational level. However, there are no such studies in the area of KM applying the Semantic Web because the Semantic Web itself is in its infancy. According to Hevner, March, Park, and Ram (2004) dichotomy of the research on management information systems, studies on the infrastructure and architecture and the killer applications refer to design science while the two other issues in Fig. 1 are related to behavioral science. The former has more characteristics of technology push than the latter. The dotted circle with the arrow of Fig. 1 shows the interactive relationship between technology push and demand pull. Research in the aspect of demand pull in the stage of introduction of the Semantic Web can contribute to improving the performance of IT investment like the prototyping approach improved productivity of the waterfall model in the system development methodology. However, until recently, most behavioral science studies reflecting demand pull have been conducted in the post-adoption stage of a specific technology. A study of the Semantic Web as an IT innovation in KM, dealing with business
7319
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Fig. 1. Research areas of the Semantic Web in KM.
issues together in the perspective of demand pull, is important, especially in the environment of rapid IT development. Many organizations design and use KM systems as specialized information systems where various technologies are integrated. KM systems help organizations increase their effectiveness and competitiveness. However, there are some limitations in achieving the objectives of KM due to limited ability for semantic integration. Therefore, our research procedure is first to identify barriers or limitations of existing KM systems through an empirical study to analyze user’s needs in terms of demand pull and then to discuss how to overcome the limitations by applying the Semantic Web to KM. The purpose of this study is to discuss the limitations of current KM systems and to propose an approach for applying the Semantic Web to KM. First, we analyze the factors that affect KM system user dissatisfaction through a survey. Second, we analyze how to overcome the limitations of current KM systems by applying the Semantic Web.
infrastructure and enabler of KM (Alavi & Leidner, 2001; Tanriverdi, 2005). Many organizations have introduced KM systems as a systematic way of applying IT to KM. Recently, Kulkarni, Ravindran, and Freeze (2007) presented a KM success model derived from DeLone and McLean’s IS success model (Delone & McLean, 1992) and conducted an empirical study. Most of all previous works focused on success factors of information systems including KM systems (Delone & McLean, 1992; Delone & McLean, 2003; Kulkarni et al., 2007; Rai, Lang, & Welker, 2002; Seddon, 1997). There are a few studies on barriers or limitations to information systems (Bouwman, Carlsson, MolinaCastillo, & Walden, 2007; Chircu & Kauffman, 2000; Damodaran & Olphert, 2000; Evgeniou & Cartwright, 2005). Even though business value often originates from a KM system, it does not always improve organizational performance as a lag or discrepancy exists between innovation and performance (Damanpour & Evan, 1984). This indicates possible barriers or limitation factors between the KM system and business values. We classify barriers into two types: technological factors and social/cultural factors including people as described in Benbya, Passiante, and Belbaly (2004). In this study, we focus on analyzing the technological limitations, rather than social and cultural limitations, of KM systems. We propose the research model shown in Fig. 2 by considering the characteristics of KM systems and also by referring to DeLone and McLean’s IS success model (Delone & McLean, 1992). We apply the reverse perspective of their IS success model (Delone & McLean, 1992). They proposed system quality and information quality as important factors that affect user satisfaction and organizational performance. Table 1 provides the definition of the limitation factors listed in Fig. 2, and indicates the number of items used to measure each
Limitation Factors in System Quality •Time/Space •Inconvenience •Knowledge Search •Knowledge Integration
User Dissatisfaction Limitation Factors in Knowledge Quality
2. Research design There have been many studies on the relationship between IT and KM (Edwards & Collier, 2005; Holsapple, 2005; Tanriverdi, 2005; Tsui, 2005). Tyndale (2002) considered IT as a KM tool which supported such activities as creation, sharing and transfer of knowledge. IT improves business values by playing a role as an
•Incongruence/ Incompleteness •Untrustworthiness Fig. 2. The research model.
Table 1 Definitions of KM systems limitation factors Factors and dissatisfaction Limitation factors for system quality
Time and space Inconvenience Knowledge search Knowledge integration
Limitation factors for knowledge quality
Incongruence and incompleteness of knowledge Untrustworthiness of knowledge
User dissatisfaction
Definition
Number of items
Time and space limitation in the KM system use and limitation of access methods Degree of inconvenience of the KM system use resulting from slow response and instability Limitations of keyword-based search as well as limited knowledge categorization Limitations in integration of heterogeneous systems as knowledge resources and integration of the existing KM system with Web resources Degree of incongruence or incompleteness of knowledge offered by the KM system
3 3 6 3
Degree of inaccuracy and untrustworthiness of knowledge offered by the KM system Degree of overall dissatisfaction with KM system use
6 1
4
7320
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
construct. We developed the measures by referring to the following previous work: System quality and information quality of the IS success model (Delone & McLean, 1992; Delone & McLean, 2003; Rai et al., 2002; Seddon, 1997). Barriers or limitations to IS (Chircu & Kauffman, 2000; Damodaran & Olphert, 2000; Evgeniou & Cartwright, 2005). Applications of the Semantic Web to KM (Antoniou & Harmelen, 2004; Cayzer, 2004; D’Aquin et al., 2005; Davies et al., 2003; Tiwana & Ramesh, 2001). We derived six research questions from the research model. We were unable to find studies on the relationship between limitation factors in the KM system and user dissatisfaction. Therefore, we use research questions rather than hypotheses. Q1: Do the limitations of system quality have a positive impact on user dissatisfaction with the KM system? Q1a: Does the limitation of time and space in the KM system have a positive impact on dissatisfaction with the KM system? Q1b: Does the inconvenience of the KM system have a positive impact on user dissatisfaction with the KM system? Q1c: Does the limitation of knowledge search have a positive impact on user dissatisfaction with the KM system? Q1d: Does the limitation of knowledge integration have a positive impact on user dissatisfaction with the KM system? Q2: Do the limitations of knowledge quality have a positive impact on user dissatisfaction with the KM system? Q2a: Do incongruence and incompleteness of knowledge provided by the KM system have a positive impact on user dissatisfaction with KM system? Q2b: Does untrustworthiness of knowledge provided by the KM system have a positive impact on user dissatisfaction with the KM system?
3. Analysis We interviewed four managers from two firms that adopted and used KM systems for two or more years to pretest the survey questionnaire. The primary goal of the pretest was to check content validity and proper wording of the questionnaire. We revised a few question items for clarity and made some changes in the sequence of questions based on the pretest. We investigated KM system solutions produced by Korean corporations. Two products A-wave and K-wave have been used pervasively in Korea. We selected these two products as typical KM system solutions for our questionnaire survey. We chose a sample of five firms that had used the products for more than one year. Representatives of KM teams in the five firms accepted our request for support in conducting our survey. Two hundred and fifty-six respondents from the five firms returned the questionnaire. However, 34 incomplete or invalid responses were discarded because they were incomplete or incredulous which meant there were the same responses for more than seven sequential questions. A total of 222 responses were used for statistical analysis by a tool of the software package SPSS version 12.0. We conducted an exploratory factor analysis and verified six factors as shown in Appendix A. Cronbach’s a was used to test for internal consistency. The measures were acceptable as all a-values exceeded the cut-off value of 0.6 (Nunnally, 1978). To assess construct validity, we performed a confirmatory factor analysis using AMOS version 5.0. Two main types of construct validity are commonly convergent and discriminant. The test of critical ratio values, the reliability of constructs and the average variance extracted were used as the measures for convergent validity (Bagozzi & Yi, 1988; Fornell & Larker, 1981). Table 2 shows the results of confirmatory factor analysis for assessing convergent validity. The critical ratio is referred to as the t-statistic and Waldstatistic (Stevens, 1996). The critical ratio of every measurement item exceeded the cut-off value of 1.96 (Anderson & Gerbing, 1988; Bagozzi & Yi, 1988). Construct reliability (CR) should be
Table 2 The results of confirmatory factor analysis Constructs
Items
Standardized coefficient (k)
Time/space limitation
TS1 TS2 TS3 IN1 IN2 IN3 SL1 SL2 SL3 SL4 SL5 SL6 IL1 IL2 IL3 II1 II2 II3 II4 UT1 UT2 UT3 UT4 UT5 UT6
0.656 0.853 0.527 0.582 0.645 0.762 0.775 0.738 0.741 0.736 0.752 0.786 0.564 0.782 0.644 0.618 0.536 0.619 0.774 0.731 0.748 0.720 0.606 0.699 0.737
Critical ratio (t-value)
– 6.719 6.306 Inconvenience 4.086 6.682 7.190 Knowledge search limitation – 11.119 11.176 11.086 11.367 11.984 Knowledge integration limitation – 7.107 6.571 Incongruence and incompleteness – of knowledge 6.475 7.259 8.456 Untrustworthiness of knowledge – 10.436 10.042 8.416 9.745 10.286 P 2 .h P 2 P i ð ki Þ þ ei Construct reliability ðCRÞ ¼ ð ki Þ a .hP Refer to Fornell and Larker (1981) for following formula . P i P ðki Þ2 þ ei AVE ¼ ðki Þ2
Measurement error (e)
Construct reliabilitya
AVEa
0.746 0.316 0.651 0.582 0.412 0.307 0.41 0.384 0.436 0.512 0.454 0.354 0.655 0.334 0.717 0.645 0.708 0.432 0.287 0.247 0.335 0.363 0.501 0.348 0.323
0.708
0.456
0.753
0.507
0.889
0.573
0.699
0.441
0.758
0.444
0.895
0.587
7321
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327 Table 3 Correlations between constructs versus average variances extracted Variables
Time/space limitation
Time/space limitation Inconvenience Search limitation Integration limitation Incongruence and incompleteness Untrustworthiness
0.675 0.196 0.068 0.136 0.040 0.011
Inconvenience
0.712 0.494 0.444 0.308 0.368
Search limitation
Integration limitation
Incongruence and incompleteness
0.757 0.501 0.609 0.662
0.664 0.428 0.458
0.666 0.654
Untrustworthiness
0.766
Note: Diagonal elements (in boldface) represent square roots of average variances extracted; other cells denote correlations.
Table 4 The results of regression analysis Dependent variable
Independent variables
Mean
Standard coefficient (b)
t-Value
Significance
Result
User dissatisfaction
Time/space limitation Inconvenience Search limitation Integration limitation Incongruence and incompleteness Untrustworthiness
3.05 2.31 2.61 2.80 2.64 2.39
0.026 0.139 0.248 0.206 0.245 0.390
0.465 2.444 4.371 3.627 4.309 6.866
0.642 0.015 0.000 0.000 0.000 0.000
Non-affirmative Affirmative Affirmative Affirmative Affirmative Affirmative
R2 = 0.336, adjusted R2 = 0.316, F = 17.360, significance: 0.000.
greater than 0.7 and average variance extracted (AVE) should be at least 0.5 to indicate reliable constructs (Hair, Anderson, Tatham, & Black, 1998). All CRs exceed 0.7 with the exception of knowledge integration limitation with 0.699. Three of six constructs were slightly below 0.5 in terms of the AVE. The second major type of construct validity is known as discriminant validity representing that constructs are different if the AVE is greater than their shared variance. In other word, the square root of the AVE for a given construct should be greater than the absolute value of the standardized correlation of the given construct with other constructs (Fornell & Larker, 1981). As shown in Table 3, the square root of the AVE for each pair of constructs exceeds the absolute value of the corresponding correlation. Thus, discriminant validity was satisfactory. Table 4 shows the results of multiple regression analysis. The regression model was significant at the p < 0.001 level. Six limitation factors of the model explained 31.6% of user dissatisfaction with the KM system. All factors, except the time/space limitation for Q1a, had a significant positive influence on user dissatisfaction. Namely, we found that technical factors such as search limitation, lack of integration, and inconvenience positively affected user dissatisfaction with the KM systems. The limitation factors of knowledge quality such as incongruence/incompleteness and untrustworthiness also increased user dissatisfaction at the significant level of 0.001. Accordingly, we needed a new approach or an alternative for overcoming the limitations to improving user satisfaction with the KM systems.
4. The Semantic Web as innovation of KM 4.1. Understanding the complexity of Semantic Web technologies Understanding the complexity of technology is a critical factor for its successful application to creation or enhancement of business value. Appendices B and C provide concise descriptions of technologies related to the Semantic Web. Fig. 3 shows the relationships among Semantic Web technologies through a collaboration diagram which represents the interacting relations between objects in unified modeling language (UML).
Ontology languages such as DAML + OIL and OWL, as well as RDF and RDFS, are grounded in XML syntax. OWL is based on DAML + OIL and uses DL for knowledge representation because reasoning is important for ontology design, integration and deployment. OWL-s as the successor of DAML-s is based on OWL to facilitate the automation of Web services. Tools and services related to the Semantic Web exploit ontology and query languages as well as XML. W3C developed RDF as standard for metadata and adopted OWL as a recommended standard for knowledge representation. On-to-knowledge is a project of information society technologies (IST) where various tools including OntoBroker, OntoEdit, Sesame, and OntoShare were developed. Ontoprise (www.ontoprise.de), a successor of OntoEdit, offers software solutions based on ontologies such as OntoBroker and OntoStudio. 4.2. Overcoming limitations of the KM system through the Semantic Web The findings of the empirical analysis presented in the previous section suggest that the limitation factors of a KM system are related to system quality and knowledge quality. The limitation factors of system quality are mainly related to the technology itself while the limitation factors of knowledge quality are related to people and culture (Benbya et al., 2004). In this section, we discuss how the Semantic Web support KM processes and how a KM system based on the Semantic Web offers an opportunity to overcome technical limitations of the current KM systems. Hereafter, we define the system possessing semantics and providing machine-processable capabilities by using the Semantic Web technologies as the Semantic Web-driven KM system. Fig. 4 shows the relationship between the Semantic Web-driven KM system and the KM process. Capabilities such as semantics and machine-processability, enabled by Semantic Web technologies, give the Semantic Web-driven KM system the potential for semantic integration and reducing information overload. Ultimately, the Semantic Web-driven KM system supports KM processes by overcoming the limitations of systems integration and knowledge search embedded in the existing KM system. The KM system, based on the Semantic Web, supports each KM process in innovative ways that the existing KM system is incapable of providing.
7322
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Fig. 3. Collaboration diagram of the Semantic Web technologies.
Fig. 4. The relationship between a Semantic Web-driven KM system and a KM process.
(1) Overcoming the search limitation The Semantic Web enables us to overcome the barriers to knowledge retrieval in the current KM system. All resources in the Semantic Web are represented in RDF as metadata and this representation method makes it possible for users to query and get answers as if they are using database management systems. The Semantic Web also supports RDFS and ontology which enables
semantic analysis on vocabularies contained in query and domains as well as syntactic analysis. Thus, the Semantic Web can provide accurate knowledge suitable to users. The Semantic Web also can offer context-aware knowledge to users because ontology languages, such as OWL, support reasoning functions and domain knowledge. The inference function and context-aware capability enable a Semantic Web-driven KM system to enhance the ability to search knowledge suitable for users. Internal or external documents of organizations and Web resources can be represented as a resource in RDF. A resource of RDF, a knowledge object, can be searched with an independent knowledge unit as a user searches a document in document management systems. Furthermore, a specific part or sentence of a Web page or a part of a document may be represented as a knowledge object. This capability allows the Semantic Web-driven KM system to search for a knowledge object unit rather than a document unit. (2) Overcoming the integration limitation There are three types of integration: data, application and process (Giachetti, 2004). The goal of data integration is data sharing where different systems exchange data with each other. The goal of application integration is to achieve interoperability between systems. The obstacles for integration arise at syntactic and semantic heterogeneity between different information systems or applications (Giachetti, 2004; Noy, Doan, & Halevy, 2005; Uschold & Gruninger, 2004). Until recently, the approaches for providing interoperability include standardization and middleware or mediators as well as enterprise application integration (EAI). The enterprise knowledge portal (EKP) is a system for integrating various tools for KM in the perspective of users (Benbya et al., 2004). Although the traditional integration approaches such as middleware and standardization easily integrate structured data extracted from heterogeneous databases, they have limitations when integrating unstructured data or knowledge from sources such as HTML, word processor files, and spreadsheet files. For instance, the limitation of EKP is that it does not play the role of a content integrator which automatically extracts related
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
7323
Fig. 5. Comparison of traditional and semantic integration.
knowledge from different sources and aggregates this knowledge. EKP only plays the role of an interconnected integrator which integrates different applications and offers one access point for users. Fig. 5 shows a comparison of traditional integration and semantic integration based on the Semantic Web. Although the traditional approach enables syntactic and structural integration at the application level as well as the data level, it cannot provide semantic integration. In other words, the traditional approach allows users to share data between different systems and provides interoperability between systems by exploiting the mediator as shown in Fig. 5. The traditional approach needs n*(n 1) mediators for mapping and translating between systems in the worst case. Software agents cannot understand terms represented in different systems or process them without human intervention. In the integration approach based on the Semantic Web, the software agents understand the meanings of the terms and automatically process them by exploiting the RDF and ontologies as illustrated in the upper and right part of Fig. 5. Assume that two different systems have the same term, Lincoln that actually means different things while the two systems have different terms, stayed and lodged that mean the same thing. For example, consider the sentence, ‘‘President Lincoln visited Lincoln and stayed (or lodged) at The Husker.” The sentence is written in the RDF in the manner as shown in Fig. 5. The software agents capture the instance of the RDF and understand the meanings of the terms by referring to ontologies. President Lincoln is a type of Person while the other is a type of Region which means the name of a city. The Husker is a type of Hotel as a subclass of Accommodation and also both stayed and lodged mean the same thing. Since W3C adopted XML as the Web document standard, XML is widely used to ensure interoperability among heterogeneous systems (Giachetti, 2004; Uschold & Gruninger, 2004). The Semantic
Web, RDF, RDFS, and OWL follow XML-based syntax (Antoniou & Harmelen, 2004). In the Semantic Web, a software agent can access heterogeneous systems and provide knowledge and information suitable for users (Beydoun, Kultchitsky, & Manasseh, 2007). The Semantic Web enables software agents to extract some parts of the related knowledge from different resources and to automatically aggregate knowledge without a user’s intervention. (3) Overcoming inconvenience When the Semantic Web is combined with ubiquitous computing (Chen & Finin, 2004), users can access KM systems anytime and anywhere conveniently. The Semantic Web-driven KM systems with the support of peer-to peer (P2P) technology (Davies et al., 2003) can improve personalized KM services. Therefore, the Semantic Web enables the current KM system to overcome limitations such as time/space and inconvenience through a combination of ubiquitous networks and P2P technology. (4) Improvements of knowledge quality and support of the SECI The Semantic Web-driven KM system facilitates KM processes. The virtual or online community increasingly plays an important role in knowledge exchange. Furthermore, Semantic Web technologies can help to enable online communities to evolve, use, and manipulate more intuitively emerging content and knowledge structures (Neumann, Hogan, & MacDonaill, 2005). The introduction of semantics and ontologies to current KM systems gives a way to semantic interoperability of resources for online communities. Nonaka and Konno (1998) suggested the SECI (socialization, externalization, combination, and internalization) model for knowledge creation. Let us discuss the relationships between the Semantic Web-driven KM system and each knowledge conversion process of the SECI model. Socialization which involves the sharing of tacit knowledge among individuals is activated by exchanging of tacit knowledge through joint activities (Nonaka & Konno, 1998). The Semantic
7324
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Fig. 6. Interaction among groups for deriving demand pull.
Web-driven KM system facilitates exchange of tacit knowledge for an online community by providing capabilities of semantics and integration. In the SECI model, externalization by conversing tacit into explicit knowledge requires the expression of tacit knowledge. Semantic Web technologies including RDF, RDFS, and ontology provide ability for knowledge representation as well as reasoning for the KM system. Thus, the KM system facilitates externalization of knowledge creation. According to Nonaka and Konno (1998), combination involves the conversion of explicit knowledge into new and more complex sets of explicit knowledge. It is important to capture, disseminate, edit or process, and integrate explicit knowledge in the combination. Consider blogs as an example of explicit knowledge on the Web. Bloggers easily and conveniently bring other blogs in their blog site and reply or add new knowledge to them. Blogging is an activity for creating, capturing, editing, and integrating explicit knowledge. Cayzer (2004) discussed the blog as a killer application for the Semantic Web. The reason is that capabilities of semantic integration and automatic processing of knowledge can support blogging. Finally, internalization of the SECI model means the conversion of explicit knowledge into tacit knowledge, requiring identification or search of relevant knowledge as well as a learning and training program. The Semantic Web-driven KM system enables users to resolve the information overload problem resulting from the traditional keyword search method on the Web. Users of the Semantic Web-driven KM system can find more relevant and accurate knowledge in Web resources and improve their learning effect and knowledge quality.
5. Conclusions To identify the limitation factors of current KM systems, we proposed a research model representing the relationship between the limitation factors of KM systems and user dissatisfaction. We tested the research model by analyzing the data from users of KM systems in five different firms with advanced IT applications. The results of the empirical analysis indicated that there indeed exist limitation factors in system quality and knowledge quality in current KM systems. Inconvenience, search and integration are statistically significant limitation factors for system quality. On the other hand, incongruence and untrustworthiness of knowledge are significant limitation factors for knowledge quality. We focused on how to overcome the limitations of system quality by applying a Semantic Web-driven KM system to KM. The Semantic Web-driven KM system supports the KM process by en-
abling current KM systems to overcome its search and integration limitations in several ways. This paper is the first report on an empirical study on the analysis of limitation factors of the KM system. The measures developed for our survey can be employed to analyze limitations of other information systems. Most IS studies have focused on analysis of IS success factors. Recognition of the limitations or barriers in existing information systems could enforce demand pull forces in the context of a more powerful technology push. According to Hevner et al. (2004), behavioral science in IS research is passive with respect to technology. In most cases, behavioral studies are conducted only after the adoption of technology in industry. A lag exists between development and adoption of technology. Our study can contribute to the reduction of such a gap. How do organizations accept user’s needs derived from our empirical study to achieve an efficient and effective global KM by adopting the Semantic Web? To realize the potential of the Semantic Web in KM with minimum barriers, it is necessary for organizations to possess the capability for coordinating or controlling technology development to reflect user groups’ needs or requirements. One way to get such capability is to facilitate interactions and collaboration between technology development groups and user groups. Fig. 6 shows interactions between the Semantic Web initiator groups and user groups at macro- and micro-levels, namely inter- and intra-organizational levels, to improve the capability for controlling technology push. At the macro-level, groups such as systems integrators including IT firms, research institutions (e.g., W3C, AIS SIGSEMIS, and WebKorea), and funding institutions (e.g., NSF and IST) are typical examples of initiators leading technology push. The user groups include firms, public organizations, and behavioral scientists. The gap among groups in determining the objective and direction of technology development is minimized through users’ direct participation and a variety of communication channels in technology development, in particular, at the standardization stage. At the micro-level, the initiator groups ranging from CEO and CIO to IS project teams should inform the potential of the Semantic Web technology to user groups and provide guidelines for its application. In general, the process of interaction between both groups at the micro-level is as follows: User groups identify and recognize the concept of the Semantic Web, its capabilities and features, and its potential benefits. They find opportunities to apply the Semantic Web to their business domains or tasks. Finally, they suggest new ideas or methods for technology or system development.
7325
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Appendix A Results of factor analysis and reliability analysis Limitation factora
Itemb
Time/space limitation
I should use the KM system for only a limited time (TS1) I should use the KM system only at limited locations (TS2) I have some limitations in accessing in the KM system through various access points such as mobile devices, PDA, and the Internet (TS3) I think the KM system is unstable due to malfunctions or system errors (IN1) I think the response time of the KM system is slow (IN2) I think it is not easy to use the KM system (IN3) I think it is difficult to find the knowledge that I need in the KM system (SL1) I think the classification scheme for knowledge is not organized well in the KM system (SL2) I think the KM system provides irrelevant knowledge as search results (SL3) I think the KM system provides redundant or unnecessary knowledge as search results (SL4) I have some difficulty in finding knowledge suitable to my requirements from the KM system with only partial prior knowledge (SL5) I cannot find relevant knowledge when I use the keyword search method with synonyms (SL6) The KM system is not integrated with other information systems such as groupware and electronic document management systems (IL1) I have some difficulty in aggregating and integrating knowledge from disparate information systems (IL2) I have some difficulty in integrating knowledge from internal systems with knowledge from the Web (IL3) I think the KM system contains the knowledge irrelevant to my tasks (II1) I think knowledge provided by the KM system is redundant (II2) I think knowledge provided by the KM system is neither comprehensive nor complete enough (II3) I think knowledge provided by the KM system is too abstract (II4) I think knowledge provided by the KM system is inaccurate (UT1) I think knowledge provided by the KM system is quite different from business practices (UT2) I think knowledge provided by the KM system is not verified through practical business activities (UT3) I cannot trust people who provide knowledge to the KM system (UT4) I think knowledge provided by the KM system is not verified as trustworthy (UT5) I think knowledge provided by the KM system is not reliable (UT6)
Inconvenience
Knowledge search limitation
Knowledge integration limitation
Incongruence and incompleteness of knowledge
Untrustworthiness of knowledge
a b c
Factor loading
Cronbach’s
a
.738 (.336)c .804 (.187) .777 (.201)
0.7265
.825 .625 .618 .578 .630
0.7134
(.185) (.321) (;359) (.447) (.386)
0.8874
.598 (.352) .686 (.321) .789 (.207) .721 (.276) .779 (.245)
0.6985
.601 (.301) .656 (.315) .616 (.264) .654 (.202) .630 (.248) .559 (.385) .588 (.376) .694 (.240)
0.7419
0.8481
.633 (.306) .642 (.347) .694 (.297) .799 (.213)
Method for factor extraction; principal component. Rotation method; Varimax with Kaiser normalization. All items of questionnaire are on a scale of 1–5, where 1 = do not agree at all, 2 = agree very little, 3 = somewhat agree, 4 = agree very much, and 5 = completely agree. The highest cross-loading for an item.
Appendix B Basics and languages related to the Semantic Web Components
Overview
URI, NS, XML, XML schema
Unicode, uniform resource identifier (URI), namespace (NS), and eXtensible markup language (XML) are the bases of the Semantic Web architecture. URI provides a unique identifier for any object as well as the Web resource. Unicode supports the multiple languages in which information is marked up all over the world. Unicode and URI provide the standard for recognition and exchange of entities at the bottom of the Semantic Web layer cake (Bernes-Lee et al., 2001). XML enables us to write structured Web documents and plays the role of syntax for data exchange. XML schema provides a mechanism defining the grammar of XML documents. XML namespace (NS) is a collection of names, allowing resources to be uniquely identified by URI reference Resource description framework (RDF) is an abstract and conceptual framework for defining and using metadata. RDF provides a foundation for processing metadata of resources and represents the resource as a graph or object–attribute–value triple. RDF has an XML-based syntax RDF schema (RDFS) is a schema defining the terms that are used in RDF statements and gives specific meanings to them. All RDF vocabularies predefined in RDFS describe classes of resource and types of relationships between resources. Although RDFS allows the representation of some ontological knowledge, it has limitations in the representation of domain knowledge (Antoniou & Harmelen, 2004) (continued on next page)
RDF
RDFS
7326
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Appendix B (continued) Components
Overview
DAML + OIL
Ontology is a formal explicit specification of a shared conceptualization. Ontology plays a role of the thesaurus in which relations among terminologies are defined. Ontology provides higher semantics for a domain. It not only provides more expressive power for the domain, but also offers reasoning support DARPA agent markup language + ontology inference layer (DAML + OIL) is an ontology language for semantic markup of web resources and the predecessor of W3C’s web ontology language (OWL) OWL is a semantic markup language for publishing and sharing ontologies on the Web. OWL is developed as a vocabulary extension of RDF and RDFS and is derived from the DAML + OIL web ontology language. OWL provides the ability to express complicated relationships among entities. There are three species of OWL: OWL lite, OWL DL (description logic), and OWL full. OWL full is the full ontology language, which is however undecidable (Antoniou & Harmelen, 2004) RDF data query language (RDQL) is the query language extracting information from RDF graphs. The Jena tool currently exploits RDQL RDF query language (RQL) is the query language for both RDF and RDFS. RQL is a declarative language for uniformly querying RDF schemas and resource descriptions. Protégé-2000 supports RQL and Sesame supports both RQL and RDQL OWL–query language (OWL–QL) is the OWL query language that is a candidate standard language and protocol for query-answering dialogues among Semantic Web computational agents using knowledge represented in OWL (Fikes, Hayes, & Horrocks, 2004) Description logics (DL) is a logic-based knowledge representation formalism for the representation of and reasoning about knowledge and ontologies based on artificial intelligence such as frame-based systems and semantic networks F-Logic is a higher-order language for reasoning about objects, inheritance and schema Description logic programs (DLP) is an alternative of the hybrid knowledge representation integrating knowledge bases described in DL with logic programs DAML-S is a DAML-based Web Service ontology, which supplies Web Service providers with a core set of markup language constructs for describing the properties and capabilities of their Web Services OWL-S is an OWL-based Web Service ontology supporting tool and agent technology to enable automation of services on the Semantic Web
OWL
RDQL RQL
OWL–QL
DL, F-Logic, DLP
DAML-S OWL-S
Appendix C Tools, projects, and organizations related to the Semantic Web OntoEdit
OntoEdit is a graphics-based tool for supporting the development and maintenance of ontologies (Davies et al., 2003)
Protégé-2000
Protégé-2000 is an ontology and knowledge-based editor, and also an open-source tool that provides an extensible architecture for the creation of customized knowledge-based applications (Noy et al., 2001) Jena is a Java based open-source framework for building Semantic Web applications Sesame is a system for storage and query of RDF and RDFS information OntoShare is a system which facilitates and encourages ontology-based knowledge sharing between virtual communities of practice There are various parsers such as Redland and SiRPAC for parsing RDF/XML syntax OntoBroker consists of an inference engine, which can read the ontologies and process the logic represented within. It supports logic-based ontology languages such as F-Logic as well as RDF and OWL KArlsruhe ontology (KAON) is an open-source ontology management infrastructure targeted for business applications. It includes a comprehensive tool suite allowing easy ontology creation and management. It provides a framework for building ontology-based applications (Volz, Staab, Oberle, & Motik, 2003) OntoMat-Annotizer is an interactive Web page annotation tool that supports the user with the task of creating and maintaining OWL ontology Ontopia provides solutions for managing knowledge and information based on Topic Maps Cerebra offers two software products based on ontology. Cerebra Server is a set of software providing business logic inference and processing capabilities and Cerebra Construct is a modeling tool using the OWL language On-to-knowledge (ontoknowledge.semanticweb.org) is a project that aims to develop methods and tools and to facilitate knowledge management. The project was initiated by a consortium of several research institutions in EU
Jena Sesame OntoShare Parser OntoBroker KAON
OntoMatannotizer Ontopia Cerebra On-toknowledge
References Alavi, M., & Leidner, D. E. (2001). Review: Knowledge management and knowledge management systems: Conceptual foundations and research issues. MIS Quarterly, 25(1), 107–136. Anderson, J., & Gerbing, W. (1988). Structural equation modelling in practice: A review and recommended two stages approach. Psychological Bulletin, 27(1), 5–24.
Antoniou, G., & Harmelen, F. (2004). A semantic web primer. Cambridge, Mass: MIT Press. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16, 421–438. Benbya, H., Passiante, G., & Belbaly, N. A. (2004). Corporate portal: A tool for knowledge management synchronization. International Journal of Information Management, 24(3), 201–220.
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327 Bernes-Lee, T., Hendler, J., & Lassila, O. (2001). The semantic web. Scientific American, 284(5), 34–43. Beydoun, G., Kultchitsky, R. R., & Manasseh, G. (2007). Evolving semantic web with social navigation. Expert Systems with Applications, 32, 265–276. Bouwman, H., Carlsson, C., Molina-Castillo, F. J., & Walden, P. (2007). Barriers and drivers in the adoption of current and future mobile services in Finland. Telematics and Informatics, 24, 145–160. Cayzer, S. (2004). Semantic blogging and decentralized knowledge management. Communications of the ACM, 47(12), 47–52. Chen, H., & Finin, T. (2004). An ontology for context aware pervasive computing environments. Knowledge Engineering, 18(3), 197–207. Chircu, A. M., & Kauffman, R. J. (2000). Limits to value in electronic commercerelated IT investments. Journal of Management Information Systems, 17(2), 59–80. Damanpour, F., & Evan, W. M. (1984). Organizational innovation and performance: The problem of organizational lag. Administrative Science Quarterly, 29, 392–409. Damodaran, L., & Olphert, W. (2000). Barriers and facilitators to the use of knowledge management systems. Behaviour and Information Technology, 19, 405–413. D’Aquin, M., Bouthier, C., Brachais, S., Lieber, J., & Napoli, A. (2005). Knowledge editing and maintenance tools for a semantic portal in oncology. International Journal of Human–Computer Studies, 62, 619–638. Davies, N. J., Fensel, D., & Harmelen, F. V. (Eds.). (2003). Toward the semantic web: Ontology-based knowledge management.. Hoboken, NJ: John Wiley & Sons. Delone, W. H., & McLean, E. R. (1992). Information system success: The quest for the dependent variable. Information Systems Research, 3(1), 60–95. Delone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A ten-year update. Journal of Management Information Systems, 19(4), 9–30. Edwards, D. S., & Collier, P. M. (2005). Knowledge management systems: Finding a way with technology. Journal of Knowledge Management, 9(1), 113–125. Evgeniou, T., & Cartwright, P. (2005). Barriers to information management. European Management Journal, 23(3), 293–299. Fikes, R., Hayes, P., & Horrocks, I. (2004). OWL–QL – A language for deductive query answering on the semantic web. Web Semantics: Science, Service and Agents on the World Wide Web, 2(1), 19–29. Fornell, C., & Larker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Giachetti, R. E. (2004). A framework to review the information integration of the enterprise. International Journal of Production Research, 42(6), 1147–1166. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). Prentice Hall: Upper Saddle River, NJ.
7327
Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75–105. Holsapple, C. W. (2005). The inseparability of modern knowledge management and computer-based technology. Journal of Knowledge Management, 9(1), 42–52. Jena2 (2006). Jena: A semantic web framework for Java. Retrieved: October 28, 2006, from Jena Web Site: . Kulkarni, U., Ravindran, S., & Freeze, R. (2007). A Knowledge management success model: Theoretical development and empirical validation. Journal of Management Information Systems, 23(3), 309–347. Neumann, M., Hogan, D., & MacDonaill, C. (2005). Semantic social network portal for collaborative online communities. Journal of European Industrial Training, 29(6), 472–487. Nonaka, I., & Konno, N. (1998). The concept of Ba: Building foundation for knowledge creation. California Management Review, 40(3), 40–54. Noy, N. F., Doan, A., & Halevy, A. Y. (2005). Semantic integration. AI Magazine, 26(1), 7–9. Noy, N. F., Sintek, M., Decker, S., Crubezy, M., Fergerson, R. W., & Musen, M. A. (2001). Creating semantic web contents with Protégé-2000. IEEE Intelligent Systems, 16(2), 60–71. Nunnally, J. C. (1978). Psychometric theory. New York: McGraw-Hill. Orlikowski, W. J., & Iacono, C. S. (2001). Research commentary: Desperately seeking the ‘IT’ in IT research – A call to theorizing the IT artifact. Information Systems Research, 12(2), 121–134. Rai, A., Lang, S. S., & Welker, R. B. (2002). Assessing the validity of IS success models: An empirical test and theoretical analysis. Information Systems Research, 13(1), 50–69. Seddon, P. B. (1997). A respecification and extension of the DeLone and McLean Model of IS success. Information Systems Research, 8(3), 240–253. Stevens, J. (1996). Applied multivariate statistics for the social science (3rd ed.). Mahwah, New Jersey: Erlbaum. Tanriverdi, H. (2005). Information technology relatedness, knowledge management capability, and performance of multibusiness firms. MIS Quarterly, 29(2), 311–334. Tiwana, A., & Ramesh, B. (2001). Integrating knowledge on the web. IEEE Internet Computing, 5(3), 32–39. Tsui, E. (2005). The role of IT in KM: Where are we now and where are we heading. Journal of Knowledge Management, 9(1), 3–6. Tyndale, P. (2002). A taxonomy of knowledge management software tools: Origins and applications. Evaluation and Program Planning, 25(2), 183–190. Uschold, M., & Gruninger, M. (2004). Ontologies and semantics for seamless connectivity. SIGMOD Record, 33(4), 58–64. Volz, R., Staab, S., Oberle, D., & Motik, B. (2003). KAON server: A semantic web management system. In Proceedings of the WWW.
Contents lists available at ScienceDirect
Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa
Adoption of the Semantic Web for overcoming technical limitations of knowledge management systems Jaehun Joo a,*, Sang M. Lee b a b
Department of Information Management, Dongguk University, 707 Sukjang-dong, Gyeongju, Gyeongsangbuk-do 780-714, South Korea Department of Management, CBA 209, University of Nebraska-Lincoln, Lincoln, NE 68588-0491, USA
a r t i c l e
i n f o
Keywords: Semantic Web Knowledge management Limitation factors IT innovation Innovation adoption
a b s t r a c t New information technology is a core factor enabling innovation in knowledge management (KM). The purpose of this study is to analyze the limitations of current KM systems and to propose an approach for applying the Semantic Web to KM. We analyzed the factors that affect KM system user dissatisfaction through a survey. We found that inconvenience, search and integration were statistically significant limitation factors for system quality. On the other hand, incongruence and untrustworthiness of knowledge were significant limitation factors for knowledge quality. Finally, we suggest methods for applying the Semantic Web to KM as an alternative overcoming the technical limitations of current KM systems. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Knowledge management (KM) has been recognized as one of the most critical factors for obtaining organizational competitive advantage (Antoniou & Harmelen, 2004). New advances in information technology (IT) have supported innovations in KM. In other words, IT innovation has become a crucial factor for survival and success of organizations in the knowledge age. The Semantic Web is one of the technologies driving a paradigm shift in KM as well as supporting electronic commerce and Web services (Antoniou & Harmelen, 2004; Davies, Fensel, & Harmelen, 2003). The Semantic Web as an extension of the current Web provides information in a well-defined manner, enabling computers and people to work in cooperation (Bernes-Lee, Hendler, & Lassila, 2001). The Semantic Web handles machine-processable information, which enables communication between machines without human intervention. One very promising application area of the Semantic Web is KM. Fig. 1 shows research areas on the Semantic Web for KM. The studies are classified into four areas: developing infrastructure and architecture, killer applications, business management issues and other social issues. Most surprising research area is technical issues related to architecture design and building of infrastructure for knowledge management based on the Semantic Web (Bernes-Lee et al., 2001; D’Aquin et al., 2005; Tiwana & Ramesh, 2001). Recently, a
* Corresponding author. Tel.: +82 54 770 2346. E-mail addresses: [email protected] (J. Joo), [email protected] (S.M. Lee). 0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2008.09.005
few projects targeting killer applications as well as building infrastructure, have been conducted. The on-to-knowledge project as a typical example aimed at developing an ontology-based tool suite that efficiently processes large numbers of heterogeneous, distributed, and semistructured documents (Davies et al., 2003). Although the project contributed to building an infrastructure for applications of the Semantic Web to KM as well as development of tools, this project did not deal with business issues associated with considering part of strategic relationships between IT and management. As Orlikowski and Iacono (2001) pointed out, it is necessary to study the complex ensemble of people, culture, and technology embedded in social contexts as well as a specific organizational level. However, there are no such studies in the area of KM applying the Semantic Web because the Semantic Web itself is in its infancy. According to Hevner, March, Park, and Ram (2004) dichotomy of the research on management information systems, studies on the infrastructure and architecture and the killer applications refer to design science while the two other issues in Fig. 1 are related to behavioral science. The former has more characteristics of technology push than the latter. The dotted circle with the arrow of Fig. 1 shows the interactive relationship between technology push and demand pull. Research in the aspect of demand pull in the stage of introduction of the Semantic Web can contribute to improving the performance of IT investment like the prototyping approach improved productivity of the waterfall model in the system development methodology. However, until recently, most behavioral science studies reflecting demand pull have been conducted in the post-adoption stage of a specific technology. A study of the Semantic Web as an IT innovation in KM, dealing with business
7319
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Fig. 1. Research areas of the Semantic Web in KM.
issues together in the perspective of demand pull, is important, especially in the environment of rapid IT development. Many organizations design and use KM systems as specialized information systems where various technologies are integrated. KM systems help organizations increase their effectiveness and competitiveness. However, there are some limitations in achieving the objectives of KM due to limited ability for semantic integration. Therefore, our research procedure is first to identify barriers or limitations of existing KM systems through an empirical study to analyze user’s needs in terms of demand pull and then to discuss how to overcome the limitations by applying the Semantic Web to KM. The purpose of this study is to discuss the limitations of current KM systems and to propose an approach for applying the Semantic Web to KM. First, we analyze the factors that affect KM system user dissatisfaction through a survey. Second, we analyze how to overcome the limitations of current KM systems by applying the Semantic Web.
infrastructure and enabler of KM (Alavi & Leidner, 2001; Tanriverdi, 2005). Many organizations have introduced KM systems as a systematic way of applying IT to KM. Recently, Kulkarni, Ravindran, and Freeze (2007) presented a KM success model derived from DeLone and McLean’s IS success model (Delone & McLean, 1992) and conducted an empirical study. Most of all previous works focused on success factors of information systems including KM systems (Delone & McLean, 1992; Delone & McLean, 2003; Kulkarni et al., 2007; Rai, Lang, & Welker, 2002; Seddon, 1997). There are a few studies on barriers or limitations to information systems (Bouwman, Carlsson, MolinaCastillo, & Walden, 2007; Chircu & Kauffman, 2000; Damodaran & Olphert, 2000; Evgeniou & Cartwright, 2005). Even though business value often originates from a KM system, it does not always improve organizational performance as a lag or discrepancy exists between innovation and performance (Damanpour & Evan, 1984). This indicates possible barriers or limitation factors between the KM system and business values. We classify barriers into two types: technological factors and social/cultural factors including people as described in Benbya, Passiante, and Belbaly (2004). In this study, we focus on analyzing the technological limitations, rather than social and cultural limitations, of KM systems. We propose the research model shown in Fig. 2 by considering the characteristics of KM systems and also by referring to DeLone and McLean’s IS success model (Delone & McLean, 1992). We apply the reverse perspective of their IS success model (Delone & McLean, 1992). They proposed system quality and information quality as important factors that affect user satisfaction and organizational performance. Table 1 provides the definition of the limitation factors listed in Fig. 2, and indicates the number of items used to measure each
Limitation Factors in System Quality •Time/Space •Inconvenience •Knowledge Search •Knowledge Integration
User Dissatisfaction Limitation Factors in Knowledge Quality
2. Research design There have been many studies on the relationship between IT and KM (Edwards & Collier, 2005; Holsapple, 2005; Tanriverdi, 2005; Tsui, 2005). Tyndale (2002) considered IT as a KM tool which supported such activities as creation, sharing and transfer of knowledge. IT improves business values by playing a role as an
•Incongruence/ Incompleteness •Untrustworthiness Fig. 2. The research model.
Table 1 Definitions of KM systems limitation factors Factors and dissatisfaction Limitation factors for system quality
Time and space Inconvenience Knowledge search Knowledge integration
Limitation factors for knowledge quality
Incongruence and incompleteness of knowledge Untrustworthiness of knowledge
User dissatisfaction
Definition
Number of items
Time and space limitation in the KM system use and limitation of access methods Degree of inconvenience of the KM system use resulting from slow response and instability Limitations of keyword-based search as well as limited knowledge categorization Limitations in integration of heterogeneous systems as knowledge resources and integration of the existing KM system with Web resources Degree of incongruence or incompleteness of knowledge offered by the KM system
3 3 6 3
Degree of inaccuracy and untrustworthiness of knowledge offered by the KM system Degree of overall dissatisfaction with KM system use
6 1
4
7320
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
construct. We developed the measures by referring to the following previous work: System quality and information quality of the IS success model (Delone & McLean, 1992; Delone & McLean, 2003; Rai et al., 2002; Seddon, 1997). Barriers or limitations to IS (Chircu & Kauffman, 2000; Damodaran & Olphert, 2000; Evgeniou & Cartwright, 2005). Applications of the Semantic Web to KM (Antoniou & Harmelen, 2004; Cayzer, 2004; D’Aquin et al., 2005; Davies et al., 2003; Tiwana & Ramesh, 2001). We derived six research questions from the research model. We were unable to find studies on the relationship between limitation factors in the KM system and user dissatisfaction. Therefore, we use research questions rather than hypotheses. Q1: Do the limitations of system quality have a positive impact on user dissatisfaction with the KM system? Q1a: Does the limitation of time and space in the KM system have a positive impact on dissatisfaction with the KM system? Q1b: Does the inconvenience of the KM system have a positive impact on user dissatisfaction with the KM system? Q1c: Does the limitation of knowledge search have a positive impact on user dissatisfaction with the KM system? Q1d: Does the limitation of knowledge integration have a positive impact on user dissatisfaction with the KM system? Q2: Do the limitations of knowledge quality have a positive impact on user dissatisfaction with the KM system? Q2a: Do incongruence and incompleteness of knowledge provided by the KM system have a positive impact on user dissatisfaction with KM system? Q2b: Does untrustworthiness of knowledge provided by the KM system have a positive impact on user dissatisfaction with the KM system?
3. Analysis We interviewed four managers from two firms that adopted and used KM systems for two or more years to pretest the survey questionnaire. The primary goal of the pretest was to check content validity and proper wording of the questionnaire. We revised a few question items for clarity and made some changes in the sequence of questions based on the pretest. We investigated KM system solutions produced by Korean corporations. Two products A-wave and K-wave have been used pervasively in Korea. We selected these two products as typical KM system solutions for our questionnaire survey. We chose a sample of five firms that had used the products for more than one year. Representatives of KM teams in the five firms accepted our request for support in conducting our survey. Two hundred and fifty-six respondents from the five firms returned the questionnaire. However, 34 incomplete or invalid responses were discarded because they were incomplete or incredulous which meant there were the same responses for more than seven sequential questions. A total of 222 responses were used for statistical analysis by a tool of the software package SPSS version 12.0. We conducted an exploratory factor analysis and verified six factors as shown in Appendix A. Cronbach’s a was used to test for internal consistency. The measures were acceptable as all a-values exceeded the cut-off value of 0.6 (Nunnally, 1978). To assess construct validity, we performed a confirmatory factor analysis using AMOS version 5.0. Two main types of construct validity are commonly convergent and discriminant. The test of critical ratio values, the reliability of constructs and the average variance extracted were used as the measures for convergent validity (Bagozzi & Yi, 1988; Fornell & Larker, 1981). Table 2 shows the results of confirmatory factor analysis for assessing convergent validity. The critical ratio is referred to as the t-statistic and Waldstatistic (Stevens, 1996). The critical ratio of every measurement item exceeded the cut-off value of 1.96 (Anderson & Gerbing, 1988; Bagozzi & Yi, 1988). Construct reliability (CR) should be
Table 2 The results of confirmatory factor analysis Constructs
Items
Standardized coefficient (k)
Time/space limitation
TS1 TS2 TS3 IN1 IN2 IN3 SL1 SL2 SL3 SL4 SL5 SL6 IL1 IL2 IL3 II1 II2 II3 II4 UT1 UT2 UT3 UT4 UT5 UT6
0.656 0.853 0.527 0.582 0.645 0.762 0.775 0.738 0.741 0.736 0.752 0.786 0.564 0.782 0.644 0.618 0.536 0.619 0.774 0.731 0.748 0.720 0.606 0.699 0.737
Critical ratio (t-value)
– 6.719 6.306 Inconvenience 4.086 6.682 7.190 Knowledge search limitation – 11.119 11.176 11.086 11.367 11.984 Knowledge integration limitation – 7.107 6.571 Incongruence and incompleteness – of knowledge 6.475 7.259 8.456 Untrustworthiness of knowledge – 10.436 10.042 8.416 9.745 10.286 P 2 .h P 2 P i ð ki Þ þ ei Construct reliability ðCRÞ ¼ ð ki Þ a .hP Refer to Fornell and Larker (1981) for following formula . P i P ðki Þ2 þ ei AVE ¼ ðki Þ2
Measurement error (e)
Construct reliabilitya
AVEa
0.746 0.316 0.651 0.582 0.412 0.307 0.41 0.384 0.436 0.512 0.454 0.354 0.655 0.334 0.717 0.645 0.708 0.432 0.287 0.247 0.335 0.363 0.501 0.348 0.323
0.708
0.456
0.753
0.507
0.889
0.573
0.699
0.441
0.758
0.444
0.895
0.587
7321
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327 Table 3 Correlations between constructs versus average variances extracted Variables
Time/space limitation
Time/space limitation Inconvenience Search limitation Integration limitation Incongruence and incompleteness Untrustworthiness
0.675 0.196 0.068 0.136 0.040 0.011
Inconvenience
0.712 0.494 0.444 0.308 0.368
Search limitation
Integration limitation
Incongruence and incompleteness
0.757 0.501 0.609 0.662
0.664 0.428 0.458
0.666 0.654
Untrustworthiness
0.766
Note: Diagonal elements (in boldface) represent square roots of average variances extracted; other cells denote correlations.
Table 4 The results of regression analysis Dependent variable
Independent variables
Mean
Standard coefficient (b)
t-Value
Significance
Result
User dissatisfaction
Time/space limitation Inconvenience Search limitation Integration limitation Incongruence and incompleteness Untrustworthiness
3.05 2.31 2.61 2.80 2.64 2.39
0.026 0.139 0.248 0.206 0.245 0.390
0.465 2.444 4.371 3.627 4.309 6.866
0.642 0.015 0.000 0.000 0.000 0.000
Non-affirmative Affirmative Affirmative Affirmative Affirmative Affirmative
R2 = 0.336, adjusted R2 = 0.316, F = 17.360, significance: 0.000.
greater than 0.7 and average variance extracted (AVE) should be at least 0.5 to indicate reliable constructs (Hair, Anderson, Tatham, & Black, 1998). All CRs exceed 0.7 with the exception of knowledge integration limitation with 0.699. Three of six constructs were slightly below 0.5 in terms of the AVE. The second major type of construct validity is known as discriminant validity representing that constructs are different if the AVE is greater than their shared variance. In other word, the square root of the AVE for a given construct should be greater than the absolute value of the standardized correlation of the given construct with other constructs (Fornell & Larker, 1981). As shown in Table 3, the square root of the AVE for each pair of constructs exceeds the absolute value of the corresponding correlation. Thus, discriminant validity was satisfactory. Table 4 shows the results of multiple regression analysis. The regression model was significant at the p < 0.001 level. Six limitation factors of the model explained 31.6% of user dissatisfaction with the KM system. All factors, except the time/space limitation for Q1a, had a significant positive influence on user dissatisfaction. Namely, we found that technical factors such as search limitation, lack of integration, and inconvenience positively affected user dissatisfaction with the KM systems. The limitation factors of knowledge quality such as incongruence/incompleteness and untrustworthiness also increased user dissatisfaction at the significant level of 0.001. Accordingly, we needed a new approach or an alternative for overcoming the limitations to improving user satisfaction with the KM systems.
4. The Semantic Web as innovation of KM 4.1. Understanding the complexity of Semantic Web technologies Understanding the complexity of technology is a critical factor for its successful application to creation or enhancement of business value. Appendices B and C provide concise descriptions of technologies related to the Semantic Web. Fig. 3 shows the relationships among Semantic Web technologies through a collaboration diagram which represents the interacting relations between objects in unified modeling language (UML).
Ontology languages such as DAML + OIL and OWL, as well as RDF and RDFS, are grounded in XML syntax. OWL is based on DAML + OIL and uses DL for knowledge representation because reasoning is important for ontology design, integration and deployment. OWL-s as the successor of DAML-s is based on OWL to facilitate the automation of Web services. Tools and services related to the Semantic Web exploit ontology and query languages as well as XML. W3C developed RDF as standard for metadata and adopted OWL as a recommended standard for knowledge representation. On-to-knowledge is a project of information society technologies (IST) where various tools including OntoBroker, OntoEdit, Sesame, and OntoShare were developed. Ontoprise (www.ontoprise.de), a successor of OntoEdit, offers software solutions based on ontologies such as OntoBroker and OntoStudio. 4.2. Overcoming limitations of the KM system through the Semantic Web The findings of the empirical analysis presented in the previous section suggest that the limitation factors of a KM system are related to system quality and knowledge quality. The limitation factors of system quality are mainly related to the technology itself while the limitation factors of knowledge quality are related to people and culture (Benbya et al., 2004). In this section, we discuss how the Semantic Web support KM processes and how a KM system based on the Semantic Web offers an opportunity to overcome technical limitations of the current KM systems. Hereafter, we define the system possessing semantics and providing machine-processable capabilities by using the Semantic Web technologies as the Semantic Web-driven KM system. Fig. 4 shows the relationship between the Semantic Web-driven KM system and the KM process. Capabilities such as semantics and machine-processability, enabled by Semantic Web technologies, give the Semantic Web-driven KM system the potential for semantic integration and reducing information overload. Ultimately, the Semantic Web-driven KM system supports KM processes by overcoming the limitations of systems integration and knowledge search embedded in the existing KM system. The KM system, based on the Semantic Web, supports each KM process in innovative ways that the existing KM system is incapable of providing.
7322
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Fig. 3. Collaboration diagram of the Semantic Web technologies.
Fig. 4. The relationship between a Semantic Web-driven KM system and a KM process.
(1) Overcoming the search limitation The Semantic Web enables us to overcome the barriers to knowledge retrieval in the current KM system. All resources in the Semantic Web are represented in RDF as metadata and this representation method makes it possible for users to query and get answers as if they are using database management systems. The Semantic Web also supports RDFS and ontology which enables
semantic analysis on vocabularies contained in query and domains as well as syntactic analysis. Thus, the Semantic Web can provide accurate knowledge suitable to users. The Semantic Web also can offer context-aware knowledge to users because ontology languages, such as OWL, support reasoning functions and domain knowledge. The inference function and context-aware capability enable a Semantic Web-driven KM system to enhance the ability to search knowledge suitable for users. Internal or external documents of organizations and Web resources can be represented as a resource in RDF. A resource of RDF, a knowledge object, can be searched with an independent knowledge unit as a user searches a document in document management systems. Furthermore, a specific part or sentence of a Web page or a part of a document may be represented as a knowledge object. This capability allows the Semantic Web-driven KM system to search for a knowledge object unit rather than a document unit. (2) Overcoming the integration limitation There are three types of integration: data, application and process (Giachetti, 2004). The goal of data integration is data sharing where different systems exchange data with each other. The goal of application integration is to achieve interoperability between systems. The obstacles for integration arise at syntactic and semantic heterogeneity between different information systems or applications (Giachetti, 2004; Noy, Doan, & Halevy, 2005; Uschold & Gruninger, 2004). Until recently, the approaches for providing interoperability include standardization and middleware or mediators as well as enterprise application integration (EAI). The enterprise knowledge portal (EKP) is a system for integrating various tools for KM in the perspective of users (Benbya et al., 2004). Although the traditional integration approaches such as middleware and standardization easily integrate structured data extracted from heterogeneous databases, they have limitations when integrating unstructured data or knowledge from sources such as HTML, word processor files, and spreadsheet files. For instance, the limitation of EKP is that it does not play the role of a content integrator which automatically extracts related
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
7323
Fig. 5. Comparison of traditional and semantic integration.
knowledge from different sources and aggregates this knowledge. EKP only plays the role of an interconnected integrator which integrates different applications and offers one access point for users. Fig. 5 shows a comparison of traditional integration and semantic integration based on the Semantic Web. Although the traditional approach enables syntactic and structural integration at the application level as well as the data level, it cannot provide semantic integration. In other words, the traditional approach allows users to share data between different systems and provides interoperability between systems by exploiting the mediator as shown in Fig. 5. The traditional approach needs n*(n 1) mediators for mapping and translating between systems in the worst case. Software agents cannot understand terms represented in different systems or process them without human intervention. In the integration approach based on the Semantic Web, the software agents understand the meanings of the terms and automatically process them by exploiting the RDF and ontologies as illustrated in the upper and right part of Fig. 5. Assume that two different systems have the same term, Lincoln that actually means different things while the two systems have different terms, stayed and lodged that mean the same thing. For example, consider the sentence, ‘‘President Lincoln visited Lincoln and stayed (or lodged) at The Husker.” The sentence is written in the RDF in the manner as shown in Fig. 5. The software agents capture the instance of the RDF and understand the meanings of the terms by referring to ontologies. President Lincoln is a type of Person while the other is a type of Region which means the name of a city. The Husker is a type of Hotel as a subclass of Accommodation and also both stayed and lodged mean the same thing. Since W3C adopted XML as the Web document standard, XML is widely used to ensure interoperability among heterogeneous systems (Giachetti, 2004; Uschold & Gruninger, 2004). The Semantic
Web, RDF, RDFS, and OWL follow XML-based syntax (Antoniou & Harmelen, 2004). In the Semantic Web, a software agent can access heterogeneous systems and provide knowledge and information suitable for users (Beydoun, Kultchitsky, & Manasseh, 2007). The Semantic Web enables software agents to extract some parts of the related knowledge from different resources and to automatically aggregate knowledge without a user’s intervention. (3) Overcoming inconvenience When the Semantic Web is combined with ubiquitous computing (Chen & Finin, 2004), users can access KM systems anytime and anywhere conveniently. The Semantic Web-driven KM systems with the support of peer-to peer (P2P) technology (Davies et al., 2003) can improve personalized KM services. Therefore, the Semantic Web enables the current KM system to overcome limitations such as time/space and inconvenience through a combination of ubiquitous networks and P2P technology. (4) Improvements of knowledge quality and support of the SECI The Semantic Web-driven KM system facilitates KM processes. The virtual or online community increasingly plays an important role in knowledge exchange. Furthermore, Semantic Web technologies can help to enable online communities to evolve, use, and manipulate more intuitively emerging content and knowledge structures (Neumann, Hogan, & MacDonaill, 2005). The introduction of semantics and ontologies to current KM systems gives a way to semantic interoperability of resources for online communities. Nonaka and Konno (1998) suggested the SECI (socialization, externalization, combination, and internalization) model for knowledge creation. Let us discuss the relationships between the Semantic Web-driven KM system and each knowledge conversion process of the SECI model. Socialization which involves the sharing of tacit knowledge among individuals is activated by exchanging of tacit knowledge through joint activities (Nonaka & Konno, 1998). The Semantic
7324
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Fig. 6. Interaction among groups for deriving demand pull.
Web-driven KM system facilitates exchange of tacit knowledge for an online community by providing capabilities of semantics and integration. In the SECI model, externalization by conversing tacit into explicit knowledge requires the expression of tacit knowledge. Semantic Web technologies including RDF, RDFS, and ontology provide ability for knowledge representation as well as reasoning for the KM system. Thus, the KM system facilitates externalization of knowledge creation. According to Nonaka and Konno (1998), combination involves the conversion of explicit knowledge into new and more complex sets of explicit knowledge. It is important to capture, disseminate, edit or process, and integrate explicit knowledge in the combination. Consider blogs as an example of explicit knowledge on the Web. Bloggers easily and conveniently bring other blogs in their blog site and reply or add new knowledge to them. Blogging is an activity for creating, capturing, editing, and integrating explicit knowledge. Cayzer (2004) discussed the blog as a killer application for the Semantic Web. The reason is that capabilities of semantic integration and automatic processing of knowledge can support blogging. Finally, internalization of the SECI model means the conversion of explicit knowledge into tacit knowledge, requiring identification or search of relevant knowledge as well as a learning and training program. The Semantic Web-driven KM system enables users to resolve the information overload problem resulting from the traditional keyword search method on the Web. Users of the Semantic Web-driven KM system can find more relevant and accurate knowledge in Web resources and improve their learning effect and knowledge quality.
5. Conclusions To identify the limitation factors of current KM systems, we proposed a research model representing the relationship between the limitation factors of KM systems and user dissatisfaction. We tested the research model by analyzing the data from users of KM systems in five different firms with advanced IT applications. The results of the empirical analysis indicated that there indeed exist limitation factors in system quality and knowledge quality in current KM systems. Inconvenience, search and integration are statistically significant limitation factors for system quality. On the other hand, incongruence and untrustworthiness of knowledge are significant limitation factors for knowledge quality. We focused on how to overcome the limitations of system quality by applying a Semantic Web-driven KM system to KM. The Semantic Web-driven KM system supports the KM process by en-
abling current KM systems to overcome its search and integration limitations in several ways. This paper is the first report on an empirical study on the analysis of limitation factors of the KM system. The measures developed for our survey can be employed to analyze limitations of other information systems. Most IS studies have focused on analysis of IS success factors. Recognition of the limitations or barriers in existing information systems could enforce demand pull forces in the context of a more powerful technology push. According to Hevner et al. (2004), behavioral science in IS research is passive with respect to technology. In most cases, behavioral studies are conducted only after the adoption of technology in industry. A lag exists between development and adoption of technology. Our study can contribute to the reduction of such a gap. How do organizations accept user’s needs derived from our empirical study to achieve an efficient and effective global KM by adopting the Semantic Web? To realize the potential of the Semantic Web in KM with minimum barriers, it is necessary for organizations to possess the capability for coordinating or controlling technology development to reflect user groups’ needs or requirements. One way to get such capability is to facilitate interactions and collaboration between technology development groups and user groups. Fig. 6 shows interactions between the Semantic Web initiator groups and user groups at macro- and micro-levels, namely inter- and intra-organizational levels, to improve the capability for controlling technology push. At the macro-level, groups such as systems integrators including IT firms, research institutions (e.g., W3C, AIS SIGSEMIS, and WebKorea), and funding institutions (e.g., NSF and IST) are typical examples of initiators leading technology push. The user groups include firms, public organizations, and behavioral scientists. The gap among groups in determining the objective and direction of technology development is minimized through users’ direct participation and a variety of communication channels in technology development, in particular, at the standardization stage. At the micro-level, the initiator groups ranging from CEO and CIO to IS project teams should inform the potential of the Semantic Web technology to user groups and provide guidelines for its application. In general, the process of interaction between both groups at the micro-level is as follows: User groups identify and recognize the concept of the Semantic Web, its capabilities and features, and its potential benefits. They find opportunities to apply the Semantic Web to their business domains or tasks. Finally, they suggest new ideas or methods for technology or system development.
7325
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Appendix A Results of factor analysis and reliability analysis Limitation factora
Itemb
Time/space limitation
I should use the KM system for only a limited time (TS1) I should use the KM system only at limited locations (TS2) I have some limitations in accessing in the KM system through various access points such as mobile devices, PDA, and the Internet (TS3) I think the KM system is unstable due to malfunctions or system errors (IN1) I think the response time of the KM system is slow (IN2) I think it is not easy to use the KM system (IN3) I think it is difficult to find the knowledge that I need in the KM system (SL1) I think the classification scheme for knowledge is not organized well in the KM system (SL2) I think the KM system provides irrelevant knowledge as search results (SL3) I think the KM system provides redundant or unnecessary knowledge as search results (SL4) I have some difficulty in finding knowledge suitable to my requirements from the KM system with only partial prior knowledge (SL5) I cannot find relevant knowledge when I use the keyword search method with synonyms (SL6) The KM system is not integrated with other information systems such as groupware and electronic document management systems (IL1) I have some difficulty in aggregating and integrating knowledge from disparate information systems (IL2) I have some difficulty in integrating knowledge from internal systems with knowledge from the Web (IL3) I think the KM system contains the knowledge irrelevant to my tasks (II1) I think knowledge provided by the KM system is redundant (II2) I think knowledge provided by the KM system is neither comprehensive nor complete enough (II3) I think knowledge provided by the KM system is too abstract (II4) I think knowledge provided by the KM system is inaccurate (UT1) I think knowledge provided by the KM system is quite different from business practices (UT2) I think knowledge provided by the KM system is not verified through practical business activities (UT3) I cannot trust people who provide knowledge to the KM system (UT4) I think knowledge provided by the KM system is not verified as trustworthy (UT5) I think knowledge provided by the KM system is not reliable (UT6)
Inconvenience
Knowledge search limitation
Knowledge integration limitation
Incongruence and incompleteness of knowledge
Untrustworthiness of knowledge
a b c
Factor loading
Cronbach’s
a
.738 (.336)c .804 (.187) .777 (.201)
0.7265
.825 .625 .618 .578 .630
0.7134
(.185) (.321) (;359) (.447) (.386)
0.8874
.598 (.352) .686 (.321) .789 (.207) .721 (.276) .779 (.245)
0.6985
.601 (.301) .656 (.315) .616 (.264) .654 (.202) .630 (.248) .559 (.385) .588 (.376) .694 (.240)
0.7419
0.8481
.633 (.306) .642 (.347) .694 (.297) .799 (.213)
Method for factor extraction; principal component. Rotation method; Varimax with Kaiser normalization. All items of questionnaire are on a scale of 1–5, where 1 = do not agree at all, 2 = agree very little, 3 = somewhat agree, 4 = agree very much, and 5 = completely agree. The highest cross-loading for an item.
Appendix B Basics and languages related to the Semantic Web Components
Overview
URI, NS, XML, XML schema
Unicode, uniform resource identifier (URI), namespace (NS), and eXtensible markup language (XML) are the bases of the Semantic Web architecture. URI provides a unique identifier for any object as well as the Web resource. Unicode supports the multiple languages in which information is marked up all over the world. Unicode and URI provide the standard for recognition and exchange of entities at the bottom of the Semantic Web layer cake (Bernes-Lee et al., 2001). XML enables us to write structured Web documents and plays the role of syntax for data exchange. XML schema provides a mechanism defining the grammar of XML documents. XML namespace (NS) is a collection of names, allowing resources to be uniquely identified by URI reference Resource description framework (RDF) is an abstract and conceptual framework for defining and using metadata. RDF provides a foundation for processing metadata of resources and represents the resource as a graph or object–attribute–value triple. RDF has an XML-based syntax RDF schema (RDFS) is a schema defining the terms that are used in RDF statements and gives specific meanings to them. All RDF vocabularies predefined in RDFS describe classes of resource and types of relationships between resources. Although RDFS allows the representation of some ontological knowledge, it has limitations in the representation of domain knowledge (Antoniou & Harmelen, 2004) (continued on next page)
RDF
RDFS
7326
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327
Appendix B (continued) Components
Overview
DAML + OIL
Ontology is a formal explicit specification of a shared conceptualization. Ontology plays a role of the thesaurus in which relations among terminologies are defined. Ontology provides higher semantics for a domain. It not only provides more expressive power for the domain, but also offers reasoning support DARPA agent markup language + ontology inference layer (DAML + OIL) is an ontology language for semantic markup of web resources and the predecessor of W3C’s web ontology language (OWL) OWL is a semantic markup language for publishing and sharing ontologies on the Web. OWL is developed as a vocabulary extension of RDF and RDFS and is derived from the DAML + OIL web ontology language. OWL provides the ability to express complicated relationships among entities. There are three species of OWL: OWL lite, OWL DL (description logic), and OWL full. OWL full is the full ontology language, which is however undecidable (Antoniou & Harmelen, 2004) RDF data query language (RDQL) is the query language extracting information from RDF graphs. The Jena tool currently exploits RDQL RDF query language (RQL) is the query language for both RDF and RDFS. RQL is a declarative language for uniformly querying RDF schemas and resource descriptions. Protégé-2000 supports RQL and Sesame supports both RQL and RDQL OWL–query language (OWL–QL) is the OWL query language that is a candidate standard language and protocol for query-answering dialogues among Semantic Web computational agents using knowledge represented in OWL (Fikes, Hayes, & Horrocks, 2004) Description logics (DL) is a logic-based knowledge representation formalism for the representation of and reasoning about knowledge and ontologies based on artificial intelligence such as frame-based systems and semantic networks F-Logic is a higher-order language for reasoning about objects, inheritance and schema Description logic programs (DLP) is an alternative of the hybrid knowledge representation integrating knowledge bases described in DL with logic programs DAML-S is a DAML-based Web Service ontology, which supplies Web Service providers with a core set of markup language constructs for describing the properties and capabilities of their Web Services OWL-S is an OWL-based Web Service ontology supporting tool and agent technology to enable automation of services on the Semantic Web
OWL
RDQL RQL
OWL–QL
DL, F-Logic, DLP
DAML-S OWL-S
Appendix C Tools, projects, and organizations related to the Semantic Web OntoEdit
OntoEdit is a graphics-based tool for supporting the development and maintenance of ontologies (Davies et al., 2003)
Protégé-2000
Protégé-2000 is an ontology and knowledge-based editor, and also an open-source tool that provides an extensible architecture for the creation of customized knowledge-based applications (Noy et al., 2001) Jena is a Java based open-source framework for building Semantic Web applications Sesame is a system for storage and query of RDF and RDFS information OntoShare is a system which facilitates and encourages ontology-based knowledge sharing between virtual communities of practice There are various parsers such as Redland and SiRPAC for parsing RDF/XML syntax OntoBroker consists of an inference engine, which can read the ontologies and process the logic represented within. It supports logic-based ontology languages such as F-Logic as well as RDF and OWL KArlsruhe ontology (KAON) is an open-source ontology management infrastructure targeted for business applications. It includes a comprehensive tool suite allowing easy ontology creation and management. It provides a framework for building ontology-based applications (Volz, Staab, Oberle, & Motik, 2003) OntoMat-Annotizer is an interactive Web page annotation tool that supports the user with the task of creating and maintaining OWL ontology Ontopia provides solutions for managing knowledge and information based on Topic Maps Cerebra offers two software products based on ontology. Cerebra Server is a set of software providing business logic inference and processing capabilities and Cerebra Construct is a modeling tool using the OWL language On-to-knowledge (ontoknowledge.semanticweb.org) is a project that aims to develop methods and tools and to facilitate knowledge management. The project was initiated by a consortium of several research institutions in EU
Jena Sesame OntoShare Parser OntoBroker KAON
OntoMatannotizer Ontopia Cerebra On-toknowledge
References Alavi, M., & Leidner, D. E. (2001). Review: Knowledge management and knowledge management systems: Conceptual foundations and research issues. MIS Quarterly, 25(1), 107–136. Anderson, J., & Gerbing, W. (1988). Structural equation modelling in practice: A review and recommended two stages approach. Psychological Bulletin, 27(1), 5–24.
Antoniou, G., & Harmelen, F. (2004). A semantic web primer. Cambridge, Mass: MIT Press. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16, 421–438. Benbya, H., Passiante, G., & Belbaly, N. A. (2004). Corporate portal: A tool for knowledge management synchronization. International Journal of Information Management, 24(3), 201–220.
J. Joo, S.M. Lee / Expert Systems with Applications 36 (2009) 7318–7327 Bernes-Lee, T., Hendler, J., & Lassila, O. (2001). The semantic web. Scientific American, 284(5), 34–43. Beydoun, G., Kultchitsky, R. R., & Manasseh, G. (2007). Evolving semantic web with social navigation. Expert Systems with Applications, 32, 265–276. Bouwman, H., Carlsson, C., Molina-Castillo, F. J., & Walden, P. (2007). Barriers and drivers in the adoption of current and future mobile services in Finland. Telematics and Informatics, 24, 145–160. Cayzer, S. (2004). Semantic blogging and decentralized knowledge management. Communications of the ACM, 47(12), 47–52. Chen, H., & Finin, T. (2004). An ontology for context aware pervasive computing environments. Knowledge Engineering, 18(3), 197–207. Chircu, A. M., & Kauffman, R. J. (2000). Limits to value in electronic commercerelated IT investments. Journal of Management Information Systems, 17(2), 59–80. Damanpour, F., & Evan, W. M. (1984). Organizational innovation and performance: The problem of organizational lag. Administrative Science Quarterly, 29, 392–409. Damodaran, L., & Olphert, W. (2000). Barriers and facilitators to the use of knowledge management systems. Behaviour and Information Technology, 19, 405–413. D’Aquin, M., Bouthier, C., Brachais, S., Lieber, J., & Napoli, A. (2005). Knowledge editing and maintenance tools for a semantic portal in oncology. International Journal of Human–Computer Studies, 62, 619–638. Davies, N. J., Fensel, D., & Harmelen, F. V. (Eds.). (2003). Toward the semantic web: Ontology-based knowledge management.. Hoboken, NJ: John Wiley & Sons. Delone, W. H., & McLean, E. R. (1992). Information system success: The quest for the dependent variable. Information Systems Research, 3(1), 60–95. Delone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A ten-year update. Journal of Management Information Systems, 19(4), 9–30. Edwards, D. S., & Collier, P. M. (2005). Knowledge management systems: Finding a way with technology. Journal of Knowledge Management, 9(1), 113–125. Evgeniou, T., & Cartwright, P. (2005). Barriers to information management. European Management Journal, 23(3), 293–299. Fikes, R., Hayes, P., & Horrocks, I. (2004). OWL–QL – A language for deductive query answering on the semantic web. Web Semantics: Science, Service and Agents on the World Wide Web, 2(1), 19–29. Fornell, C., & Larker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Giachetti, R. E. (2004). A framework to review the information integration of the enterprise. International Journal of Production Research, 42(6), 1147–1166. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). Prentice Hall: Upper Saddle River, NJ.
7327
Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75–105. Holsapple, C. W. (2005). The inseparability of modern knowledge management and computer-based technology. Journal of Knowledge Management, 9(1), 42–52. Jena2 (2006). Jena: A semantic web framework for Java. Retrieved: October 28, 2006, from Jena Web Site: . Kulkarni, U., Ravindran, S., & Freeze, R. (2007). A Knowledge management success model: Theoretical development and empirical validation. Journal of Management Information Systems, 23(3), 309–347. Neumann, M., Hogan, D., & MacDonaill, C. (2005). Semantic social network portal for collaborative online communities. Journal of European Industrial Training, 29(6), 472–487. Nonaka, I., & Konno, N. (1998). The concept of Ba: Building foundation for knowledge creation. California Management Review, 40(3), 40–54. Noy, N. F., Doan, A., & Halevy, A. Y. (2005). Semantic integration. AI Magazine, 26(1), 7–9. Noy, N. F., Sintek, M., Decker, S., Crubezy, M., Fergerson, R. W., & Musen, M. A. (2001). Creating semantic web contents with Protégé-2000. IEEE Intelligent Systems, 16(2), 60–71. Nunnally, J. C. (1978). Psychometric theory. New York: McGraw-Hill. Orlikowski, W. J., & Iacono, C. S. (2001). Research commentary: Desperately seeking the ‘IT’ in IT research – A call to theorizing the IT artifact. Information Systems Research, 12(2), 121–134. Rai, A., Lang, S. S., & Welker, R. B. (2002). Assessing the validity of IS success models: An empirical test and theoretical analysis. Information Systems Research, 13(1), 50–69. Seddon, P. B. (1997). A respecification and extension of the DeLone and McLean Model of IS success. Information Systems Research, 8(3), 240–253. Stevens, J. (1996). Applied multivariate statistics for the social science (3rd ed.). Mahwah, New Jersey: Erlbaum. Tanriverdi, H. (2005). Information technology relatedness, knowledge management capability, and performance of multibusiness firms. MIS Quarterly, 29(2), 311–334. Tiwana, A., & Ramesh, B. (2001). Integrating knowledge on the web. IEEE Internet Computing, 5(3), 32–39. Tsui, E. (2005). The role of IT in KM: Where are we now and where are we heading. Journal of Knowledge Management, 9(1), 3–6. Tyndale, P. (2002). A taxonomy of knowledge management software tools: Origins and applications. Evaluation and Program Planning, 25(2), 183–190. Uschold, M., & Gruninger, M. (2004). Ontologies and semantics for seamless connectivity. SIGMOD Record, 33(4), 58–64. Volz, R., Staab, S., Oberle, D., & Motik, B. (2003). KAON server: A semantic web management system. In Proceedings of the WWW.
