Paper - UAL
An Ontology-Driven Case Study for the Knowledge Representation of Management Information Systems Jose A. Asensio, Nicolás Padilla, and Luis Iribarne Applied Computing Group, University of Almería, Spain {jacortes,npadilla,luis.iribarne}@ual.es
Abstract. Web-based Management Information Systems (MIS) require the use of standardized methods and techniques for their design and development. The two most used resources are Model Driven Architectures (MDA) and Ontology Driven Architectures (ODA). The MDA perspective allows us to separate the specification issues from the system architecture as well as the details from its implementation. On the other hand, ODA provides the semantic representation of knowledge domain, also independently of its implementation. This paper presents the use of ODA in the SOLERES system, an information system for environmental management (EMIS). Keywords: Management Information Systems, Ontology-Driven Architecture.
1
Introduction
Nowadays, Web-based Management Information Systems (WMIS) require the use of standardized methods and techniques for their design and development [8]. The two most used resources are Model Driven Architectures (MDA) and Ontology Driven Architectures (ODA). While MDA allows us to separate, on the one hand, the specification from the system architecture and, on the other hand, details from its implementation, ODA provides the semantic representation of knowledge domain also independently of its implementation. Thus, combined use of both resources turns out to be very powerful for the design and development of this kind of systems. Both MDA and ODA have been used in SOLERES. SOLERES is an Environmental Management Information System (EMIS) based on WMIS, which is built by using the Multi-Agent Systems (MAS) technology. This paper is focused on the use of ODA in SOLERES, particularly on its knowledge representation subsystem called SOLERESKRS. In this subsystem, the ontologies have been used in two very different contexts. On the one hand, they describe the information domain knowledge in the system; on the other hand, they model the processes and communications between different system components. The content of this article is structured as follows: Section 2 examines the ontologies used in EMIS; Section 3 presents the SOLERES-KRS case study regarding the use of ontologies; finally, Section 4 shows the conclusions. M.D. Lytras et al. (Eds.): WSKS 2011, CCIS 278, pp. 426–432, 2012. © Springer-Verlag Berlin Heidelberg 2012
An Ontology-Driven Case Study for the Knowledge Representation
2
427
Background: EMIS Ontologies
It is obvious the advance of the new technologies in WMIS. Ontology is an example of these advances. Ontology was designed to be used in applications that need to manage the information semantic. In [4], the authors present an environmental decision-support system called OntoWEDSS for wastewater management. In this system, ontology is used to model the wastewater treatment process, to provide a common vocabulary, and an explicit conceptualization that describes the semantic of the data. Another example may be found in [5], an air quality monitoring system which uses an ontology to define messages and communications concisely and unambiguously. In [3], the authors present Ecolingua, an EngMath family ontology to represent quantitative ecological data. These examples show the use of ontologies to build models that describe the entities in the given domain and characterize the relationships and constraints associated with these entities. In [13], the authors present an ontology to represent geographic data and functions related to them. To meet the need for interoperable Geographic Information System (GIS), in [1] the authors propose a Geo Ontology model design to integrate geographical information. We have also explored an ontological application in the field of geographical information retrieval [11]. A different use appears in [7], where we can find an Ontology Web Language (OWL) extension with new primitives for modeling spatial location and spatial relationships with a geographic ontology. Extensions of existing ontologies have also appeared in this knowledge domain. In [10], the authors propose a geographic ontology based on Geographic Markup Language (GML) [6], and the OWL-S profile is extended to form geographic profiles. Another case is an extension of NASAs SemanticWeb for Earth and Environmental Terminology (SWEET) ontologies that include part of the hydrogeology domain [12].
3
The SOLERES-Knowledge Representation System Ontology
The SOLERES system is an information system for environmental management. The general idea of the system is the study of a framework for the integration of the aforesaid disciplines using the “Environmental information” as the application domain, specifically in ecology. SOLERES-KRS is a subsystem used to manage environmental information. Given the magnitude of the information available in the information system, and given that this information may be provided by different sources, at different times or even by different people, the environmental information can be distributed, consulted, and geographically located in different places, called Environmental Process Units (EPU). The EPU manages two local repositories of environmental information. One of them contains metadata on the information in the domain itself (i.e., basically information related to ecological and satellite image classifications): Environmental Information Map (EIM) data, or EIM documents. This information is extracted from external databases. The EIM documents have been specified by the use of an ontology in OWL [2][9], and they represent the first level of information in the KRS subsystem. The second repository contains meta-metadata:
428
J.A. Asensio, N. Padilla, and L. Iribarne
Environmental Information metaData (EID), or EID documents. These documents contain the most important metadata of the EIM documents that could be used in an information recovering service, and further, incorporate other new metadata necessary for agent management itself. To a certain extent, an EID document represents a “template” with the basic EIM document data (the meta-metadata). The EID documents have also been specified by an ontology and represent the second level of information in the KRS subsystem. Each EPU keeps its own set of EID documents locally and also registers them to a Web Trading Agent (WTA). In this way, the WTA has an overall repository with all EID documents of all EPUs in order to offer an information search service. Ontology is used in another context, too. It allows us to make the definitions of behavior and interaction protocols between the system agents. When a system agent needs another agent to perform a specific action, he builds a message where he describes such an action through an ontology and sends it to the second agent. The latter, once the message has been received, extracts the content from the ontology, performs the appropriate action and uses the same ontology to show the action result to the former. This methodology is repeated with all actions among the system agents. In next subsections, data and process ontologies used by KRS subsystem will be described. They are written in OWL notation and we formalize them through UML class diagrams. 3.1
Data Ontologies
KRS subsystem uses two data ontologies: EIM and EID ontologies. The first one contains metadata on the information in the domain itself. This ontology has widely been described in [2] and [9], so that it will not be described in this paper. The EID ontology contains meta-metadata from EIM documents. Figure 1 shows it. The ecological information domain (one type of environmental information) is related to cartographical maps and satellite images. A cartographical map stores its information in layers (Layer), each of which is identified by a set of variables (Variable). For instance, in our project we are using cartographical maps classified into 4 layers (climatology, litology, geomorphology and soils) having more than a hundred of variables (i.e., scrubland surface, among others). Something similar happens with satellite images. The information is also stored in layers, but they are called Bands here. An example of satellite image is the LANDSAT images, which have 7 bands (and no variables stored in this case). Finally, both the cartographical and satellite classification have geographical information associated and this one (Classification) is carried out at a given time (Time) by a technician or group of technicians (Technician). 3.2
Process Ontologies
Process ontologies have been modeled to make the definitions of behavior and interaction protocols between system agents, as indicated in Section 3.
An Ontology-Driven Case Study for the Knowledge Representation
429
Fig. 1. EID ontology
All ontologies have been designed in terms of concepts (domain entities), actions performed in the domain (actions that affect concepts) and predicates (expressions connected with concepts). The Admin Ontology can refer to actions that modify the main parameters of configuration of the trading service. For instance, this ontology allows us to modify the maximum number of results obtained after a query has been executed. This ontology will be used by both IMI Agent and Trading Agent. Figure 2 represents the structure of this ontology.
Fig. 2. Admin ontology
This defines three pairs of actions: SetDef_search_card/GetDef_search_card allows us to establish/get the number of registers to be located by default in a search and uses the Def_search_card concept to store such information. On the other hand, SetMax_search_card/GetMax_search_card allows us to establish/get the maximum
430
J.A. Asensio, N. Padilla, and L. Iribarne
number of registers to be located in a search and uses the Max_search_card concept to store such information. Finally, SetOffer_repos/GetOffer_repos can establish/get the address (in URL) of the file that stores the repository with the data registers used by the Trading service. The Offer_repos concept is used to store the address. The Register Ontology can refer to the management actions (creation, elimination, modification and query) about EIM and EID documents. This ontology is used in communication messages that certain agents exchange with each other: IMI Agent with Resource Agent and Resource Agent with Trading Agent. For instance, when a Resource Agent adds a new EID to his local repository, he directly registers it in the global repository of his associated Trading Agent through the use of this ontology. Figure 3 represents the structure of the Register Ontology. This ontology considers actions such as Export, Modify, Withdraw and Describe. The Export action inserts a data register previously stored in the Offer concept. The data register can be an EIM Document or an EID Document. The Modify action represents the modification of a register and uses the Offer concept to refer to a new register with the modified data and OfferId concept to represent the identifier of the data register to be modified. Lastly, the Withdraw and Describe actions represent the elimination and query of a particular data register specified in the OfferId concept.
Fig. 3. Register ontology
Finally, the Lookup Ontology can refer to the search action in a repository under previously established criteria. The data query would be carried out under a specific repository according to previously established policies. This action can be requested
An Ontology-Driven Case Study for the Knowledge Representation
431
by different agents: IMI Agent, Query Agent, Trading Agent and Resource Agent. For example, Query Agent can communicate with a Resource Agent to ask for information about an EIM document. Similarly, when an agent solves a query, he sends the right response by using the same Lookup ontology. Figure 4 represents the structure of the Lookup Ontology. The Query action uses QueryForm and PolicySeq concepts. The QueryForm concept represents a query expressed in a specific language, whereas the PolicySeq concept represents a set of Policies that can be established to make a query.
Fig. 4. Lookup ontology
4
Conclusions
The Ontology-Driven Architecture (ODA) provides the semantic representation for the knowledge domain in information systems. This paper has presented the use of ODA in SOLERES-KRS. This architecture uses ontology in two different contexts: (a) for representing the domain information, and (b) for defining the behavior and interaction protocols between system components (software agents). In order to represent the domain information, the EIM and EID ontologies have been developed. They represent metadata and meta-metadata from environmental information, respectively. On the other hand, three process-ontology documents (Lookup, Register and Admin) have been developed, used by agents for their cooperative activities.
432
J.A. Asensio, N. Padilla, and L. Iribarne
Acknowledgments. This work has been supported by the EU (FEDER) and the Spanish Ministry MICINN under grant of the TIN2010-15588 and TRA2009-0309 projects, and the JUNTA de ANDALUCÍA (proyecto de excelencia) under grant TIC6114 project, http://www.ual.es/acg/soleres.
References 1. An, Y., Zhao, B.: Geo Ontology Design and Comparison in Geographic Information Integration. In: Fourth International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2007, vol. 4 (2007) 2. Asensio, J., Padilla, N., Iribarne, L., Muñoz, F., Ayala, R., Cruz, M., Menenti, M.: A MDEbased Satellite Ontology for Environmental Management Systems. Information Technology, Information Systems and Knowledge Management (in press) 3. Brilhante, V.: An Ontology for Quantities in Ecology. In: Bazzan, A.L.C., Labidi, S. (eds.) SBIA 2004. LNCS (LNAI), vol. 3171, pp. 144–153. Springer, Heidelberg (2004) 4. Ceccaroni, L., Cortés, U., Sanchez-Marre, M.: OntoWEDSS: Augmenting Environmental Decision-Support Systems with Ontologies. Environmental Modelling and Software 19(9), 785–797 (2004) 5. Di Lecce, V., Pasquale, C., Piuri, V.: A Basic Ontology for Multi Agent System Communication in an Environmental Monitoring System. In: 2004 IEEE International Conference on Computational Intelligence for Measurement Systems and Applications, CIMSA 2004, pp. 45–50 (2004) 6. GML/OCG: Geography Markup Language. Open Geospatial Consortium, inc. Tech. rep. (2007), http://www.opengeospatial.org/standards/gml/ 7. Huang, M.: A New Method to Formal Description of Spatial Ontology. Information Technology and Environmental System Sciences 3, 417–421 (2008) 8. Lytras, M.: Semantic Web and Information Systems: An Agenda Based on Discourse with Community Leaders. International Journal of Semantic Web and Information Systems 1(1) (2005) 9. Padilla, N., Iribarne, L., Asensio, J.A., Muñoz, F.J., Ayala, R.: Modelling an Environmental Knowledge-Representation System. In: Lytras, M.D., Damiani, E., Tennyson, R.D. (eds.) WSKS 2008. LNCS (LNAI), vol. 5288, pp. 70–78. Springer, Heidelberg (2008) 10. Shen, J., Krishna, A., Yuan, S., Cai, K., Qin, Y.: A Pragmatic GIS-Oriented Ontology for Location Based Services. In: 19th Australian Conference on Software Engineering, ASWEC 2008, pp. 562–569 (2008) 11. Song, J., Zhu, Y., Wang, J.: A Study of Semantic Retrieval System Based on Geoontology with Spatio-Temporal Characteristic. In: Proc. of International Symposium on Distributed Computing and Applications to Business, Engineering and Science (DCABES 2007), vol. I-II, pp. 1029–1034 (2007) 12. Tripathi, A., Babaie, H.: Developing a Modular Hydrogeology Ontology by Extending the SWEET Upper-Level Ontologies. Computers and Geosciences 34(9), 1022–1033 (2008) 13. Zhan, Q., Li, D., Shao, Z.: An Architecture for Ontology-Based Geographic Information Semantic Grid Service. In: Proceedings of SPIE, vol. 6754, p. 67541U (2007)
Abstract. Web-based Management Information Systems (MIS) require the use of standardized methods and techniques for their design and development. The two most used resources are Model Driven Architectures (MDA) and Ontology Driven Architectures (ODA). The MDA perspective allows us to separate the specification issues from the system architecture as well as the details from its implementation. On the other hand, ODA provides the semantic representation of knowledge domain, also independently of its implementation. This paper presents the use of ODA in the SOLERES system, an information system for environmental management (EMIS). Keywords: Management Information Systems, Ontology-Driven Architecture.
1
Introduction
Nowadays, Web-based Management Information Systems (WMIS) require the use of standardized methods and techniques for their design and development [8]. The two most used resources are Model Driven Architectures (MDA) and Ontology Driven Architectures (ODA). While MDA allows us to separate, on the one hand, the specification from the system architecture and, on the other hand, details from its implementation, ODA provides the semantic representation of knowledge domain also independently of its implementation. Thus, combined use of both resources turns out to be very powerful for the design and development of this kind of systems. Both MDA and ODA have been used in SOLERES. SOLERES is an Environmental Management Information System (EMIS) based on WMIS, which is built by using the Multi-Agent Systems (MAS) technology. This paper is focused on the use of ODA in SOLERES, particularly on its knowledge representation subsystem called SOLERESKRS. In this subsystem, the ontologies have been used in two very different contexts. On the one hand, they describe the information domain knowledge in the system; on the other hand, they model the processes and communications between different system components. The content of this article is structured as follows: Section 2 examines the ontologies used in EMIS; Section 3 presents the SOLERES-KRS case study regarding the use of ontologies; finally, Section 4 shows the conclusions. M.D. Lytras et al. (Eds.): WSKS 2011, CCIS 278, pp. 426–432, 2012. © Springer-Verlag Berlin Heidelberg 2012
An Ontology-Driven Case Study for the Knowledge Representation
2
427
Background: EMIS Ontologies
It is obvious the advance of the new technologies in WMIS. Ontology is an example of these advances. Ontology was designed to be used in applications that need to manage the information semantic. In [4], the authors present an environmental decision-support system called OntoWEDSS for wastewater management. In this system, ontology is used to model the wastewater treatment process, to provide a common vocabulary, and an explicit conceptualization that describes the semantic of the data. Another example may be found in [5], an air quality monitoring system which uses an ontology to define messages and communications concisely and unambiguously. In [3], the authors present Ecolingua, an EngMath family ontology to represent quantitative ecological data. These examples show the use of ontologies to build models that describe the entities in the given domain and characterize the relationships and constraints associated with these entities. In [13], the authors present an ontology to represent geographic data and functions related to them. To meet the need for interoperable Geographic Information System (GIS), in [1] the authors propose a Geo Ontology model design to integrate geographical information. We have also explored an ontological application in the field of geographical information retrieval [11]. A different use appears in [7], where we can find an Ontology Web Language (OWL) extension with new primitives for modeling spatial location and spatial relationships with a geographic ontology. Extensions of existing ontologies have also appeared in this knowledge domain. In [10], the authors propose a geographic ontology based on Geographic Markup Language (GML) [6], and the OWL-S profile is extended to form geographic profiles. Another case is an extension of NASAs SemanticWeb for Earth and Environmental Terminology (SWEET) ontologies that include part of the hydrogeology domain [12].
3
The SOLERES-Knowledge Representation System Ontology
The SOLERES system is an information system for environmental management. The general idea of the system is the study of a framework for the integration of the aforesaid disciplines using the “Environmental information” as the application domain, specifically in ecology. SOLERES-KRS is a subsystem used to manage environmental information. Given the magnitude of the information available in the information system, and given that this information may be provided by different sources, at different times or even by different people, the environmental information can be distributed, consulted, and geographically located in different places, called Environmental Process Units (EPU). The EPU manages two local repositories of environmental information. One of them contains metadata on the information in the domain itself (i.e., basically information related to ecological and satellite image classifications): Environmental Information Map (EIM) data, or EIM documents. This information is extracted from external databases. The EIM documents have been specified by the use of an ontology in OWL [2][9], and they represent the first level of information in the KRS subsystem. The second repository contains meta-metadata:
428
J.A. Asensio, N. Padilla, and L. Iribarne
Environmental Information metaData (EID), or EID documents. These documents contain the most important metadata of the EIM documents that could be used in an information recovering service, and further, incorporate other new metadata necessary for agent management itself. To a certain extent, an EID document represents a “template” with the basic EIM document data (the meta-metadata). The EID documents have also been specified by an ontology and represent the second level of information in the KRS subsystem. Each EPU keeps its own set of EID documents locally and also registers them to a Web Trading Agent (WTA). In this way, the WTA has an overall repository with all EID documents of all EPUs in order to offer an information search service. Ontology is used in another context, too. It allows us to make the definitions of behavior and interaction protocols between the system agents. When a system agent needs another agent to perform a specific action, he builds a message where he describes such an action through an ontology and sends it to the second agent. The latter, once the message has been received, extracts the content from the ontology, performs the appropriate action and uses the same ontology to show the action result to the former. This methodology is repeated with all actions among the system agents. In next subsections, data and process ontologies used by KRS subsystem will be described. They are written in OWL notation and we formalize them through UML class diagrams. 3.1
Data Ontologies
KRS subsystem uses two data ontologies: EIM and EID ontologies. The first one contains metadata on the information in the domain itself. This ontology has widely been described in [2] and [9], so that it will not be described in this paper. The EID ontology contains meta-metadata from EIM documents. Figure 1 shows it. The ecological information domain (one type of environmental information) is related to cartographical maps and satellite images. A cartographical map stores its information in layers (Layer), each of which is identified by a set of variables (Variable). For instance, in our project we are using cartographical maps classified into 4 layers (climatology, litology, geomorphology and soils) having more than a hundred of variables (i.e., scrubland surface, among others). Something similar happens with satellite images. The information is also stored in layers, but they are called Bands here. An example of satellite image is the LANDSAT images, which have 7 bands (and no variables stored in this case). Finally, both the cartographical and satellite classification have geographical information associated and this one (Classification) is carried out at a given time (Time) by a technician or group of technicians (Technician). 3.2
Process Ontologies
Process ontologies have been modeled to make the definitions of behavior and interaction protocols between system agents, as indicated in Section 3.
An Ontology-Driven Case Study for the Knowledge Representation
429
Fig. 1. EID ontology
All ontologies have been designed in terms of concepts (domain entities), actions performed in the domain (actions that affect concepts) and predicates (expressions connected with concepts). The Admin Ontology can refer to actions that modify the main parameters of configuration of the trading service. For instance, this ontology allows us to modify the maximum number of results obtained after a query has been executed. This ontology will be used by both IMI Agent and Trading Agent. Figure 2 represents the structure of this ontology.
Fig. 2. Admin ontology
This defines three pairs of actions: SetDef_search_card/GetDef_search_card allows us to establish/get the number of registers to be located by default in a search and uses the Def_search_card concept to store such information. On the other hand, SetMax_search_card/GetMax_search_card allows us to establish/get the maximum
430
J.A. Asensio, N. Padilla, and L. Iribarne
number of registers to be located in a search and uses the Max_search_card concept to store such information. Finally, SetOffer_repos/GetOffer_repos can establish/get the address (in URL) of the file that stores the repository with the data registers used by the Trading service. The Offer_repos concept is used to store the address. The Register Ontology can refer to the management actions (creation, elimination, modification and query) about EIM and EID documents. This ontology is used in communication messages that certain agents exchange with each other: IMI Agent with Resource Agent and Resource Agent with Trading Agent. For instance, when a Resource Agent adds a new EID to his local repository, he directly registers it in the global repository of his associated Trading Agent through the use of this ontology. Figure 3 represents the structure of the Register Ontology. This ontology considers actions such as Export, Modify, Withdraw and Describe. The Export action inserts a data register previously stored in the Offer concept. The data register can be an EIM Document or an EID Document. The Modify action represents the modification of a register and uses the Offer concept to refer to a new register with the modified data and OfferId concept to represent the identifier of the data register to be modified. Lastly, the Withdraw and Describe actions represent the elimination and query of a particular data register specified in the OfferId concept.
Fig. 3. Register ontology
Finally, the Lookup Ontology can refer to the search action in a repository under previously established criteria. The data query would be carried out under a specific repository according to previously established policies. This action can be requested
An Ontology-Driven Case Study for the Knowledge Representation
431
by different agents: IMI Agent, Query Agent, Trading Agent and Resource Agent. For example, Query Agent can communicate with a Resource Agent to ask for information about an EIM document. Similarly, when an agent solves a query, he sends the right response by using the same Lookup ontology. Figure 4 represents the structure of the Lookup Ontology. The Query action uses QueryForm and PolicySeq concepts. The QueryForm concept represents a query expressed in a specific language, whereas the PolicySeq concept represents a set of Policies that can be established to make a query.
Fig. 4. Lookup ontology
4
Conclusions
The Ontology-Driven Architecture (ODA) provides the semantic representation for the knowledge domain in information systems. This paper has presented the use of ODA in SOLERES-KRS. This architecture uses ontology in two different contexts: (a) for representing the domain information, and (b) for defining the behavior and interaction protocols between system components (software agents). In order to represent the domain information, the EIM and EID ontologies have been developed. They represent metadata and meta-metadata from environmental information, respectively. On the other hand, three process-ontology documents (Lookup, Register and Admin) have been developed, used by agents for their cooperative activities.
432
J.A. Asensio, N. Padilla, and L. Iribarne
Acknowledgments. This work has been supported by the EU (FEDER) and the Spanish Ministry MICINN under grant of the TIN2010-15588 and TRA2009-0309 projects, and the JUNTA de ANDALUCÍA (proyecto de excelencia) under grant TIC6114 project, http://www.ual.es/acg/soleres.
References 1. An, Y., Zhao, B.: Geo Ontology Design and Comparison in Geographic Information Integration. In: Fourth International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2007, vol. 4 (2007) 2. Asensio, J., Padilla, N., Iribarne, L., Muñoz, F., Ayala, R., Cruz, M., Menenti, M.: A MDEbased Satellite Ontology for Environmental Management Systems. Information Technology, Information Systems and Knowledge Management (in press) 3. Brilhante, V.: An Ontology for Quantities in Ecology. In: Bazzan, A.L.C., Labidi, S. (eds.) SBIA 2004. LNCS (LNAI), vol. 3171, pp. 144–153. Springer, Heidelberg (2004) 4. Ceccaroni, L., Cortés, U., Sanchez-Marre, M.: OntoWEDSS: Augmenting Environmental Decision-Support Systems with Ontologies. Environmental Modelling and Software 19(9), 785–797 (2004) 5. Di Lecce, V., Pasquale, C., Piuri, V.: A Basic Ontology for Multi Agent System Communication in an Environmental Monitoring System. In: 2004 IEEE International Conference on Computational Intelligence for Measurement Systems and Applications, CIMSA 2004, pp. 45–50 (2004) 6. GML/OCG: Geography Markup Language. Open Geospatial Consortium, inc. Tech. rep. (2007), http://www.opengeospatial.org/standards/gml/ 7. Huang, M.: A New Method to Formal Description of Spatial Ontology. Information Technology and Environmental System Sciences 3, 417–421 (2008) 8. Lytras, M.: Semantic Web and Information Systems: An Agenda Based on Discourse with Community Leaders. International Journal of Semantic Web and Information Systems 1(1) (2005) 9. Padilla, N., Iribarne, L., Asensio, J.A., Muñoz, F.J., Ayala, R.: Modelling an Environmental Knowledge-Representation System. In: Lytras, M.D., Damiani, E., Tennyson, R.D. (eds.) WSKS 2008. LNCS (LNAI), vol. 5288, pp. 70–78. Springer, Heidelberg (2008) 10. Shen, J., Krishna, A., Yuan, S., Cai, K., Qin, Y.: A Pragmatic GIS-Oriented Ontology for Location Based Services. In: 19th Australian Conference on Software Engineering, ASWEC 2008, pp. 562–569 (2008) 11. Song, J., Zhu, Y., Wang, J.: A Study of Semantic Retrieval System Based on Geoontology with Spatio-Temporal Characteristic. In: Proc. of International Symposium on Distributed Computing and Applications to Business, Engineering and Science (DCABES 2007), vol. I-II, pp. 1029–1034 (2007) 12. Tripathi, A., Babaie, H.: Developing a Modular Hydrogeology Ontology by Extending the SWEET Upper-Level Ontologies. Computers and Geosciences 34(9), 1022–1033 (2008) 13. Zhan, Q., Li, D., Shao, Z.: An Architecture for Ontology-Based Geographic Information Semantic Grid Service. In: Proceedings of SPIE, vol. 6754, p. 67541U (2007)
