Finding Surprising Patterns in a Time Series Database in ... - UCR CS

Report 3 Downloads 25 Views
Finding Surprising Patterns in a Time Series Database in Linear Time and Space Eamonn Keogh    

Stefano Lonardi



   

Bill ‘Yuan-chi’ Chiu    ! "

Department of Computer Science and Engineering University of California Riverside, CA 92521

ABSTRACT #%$'&)(*,+-'./&01+234'56/4'798;:6?3&5@(8A&6/U&V5]0WP>S$ 8A.E8:./P56E47f4'+U&.?A"QW5&BA,&>BA,6/+4 Š ßc÷–s845@84+0C8.?Q@5&SA,&>BA,6/+4vŠ øù–M #%$'&](*,+-'./&0ú*,&VN P'6?*,&:W8Y0F&84'6/472hP.w5&S34'6?A,6/+4t+2Dm":V8:BQ +2O8F(8ASQC+2s(*,&U 6/+P':&gA,$'&Y:B*,&BA,6/…&gA,6/0F&@:&|:S€ H 7*,+P'4'5‚0C8A,&S*,6^8.I+4u:bA./P'56/4'7v:6/&4 A,.?QC>+0F('PA,&G84 Qo(8ABA,&5F2^*,&VN P'&4>BQ - QâP:J$u[I6/A,$›8e5&SAJ86/./&V5›&B0F('6?*,6/>V8. &U8./P8A,6/+4!MG†v&f[%86?AfP'4 A,6/.¦¢ &>BA,6/+4{ù c=[I$'&4|A,$'&f*,&V85&B*Vã : 6/4 A,P'6?A,6/+4':;8-=+PAfA,$'&W(*,+-.E&B0ñ8*,&F0F+*,&i2hP././Q{5&U&./+(=&5 A,+a56/:P:BA,6/+4‰g[%&D+ ƒ&B* >+4'>./P':B*,&BA,&B4'&:J$'+:&fA,6E0F&Z:S$Ö*J84'7&Y6?A 0C8(:GA,+c_84'5g0C8A,>J$'6/4'7F:bQ0W-=+.!6/:f8:B+4':J$!M|r 0WP.?A,6/A,P'5&)+2T8.?A,&B*,4'8A,6/U&;0F&BA,$+ 5:%A,+C56/:B*,&SA,6E…B&;A,$&jA,6/0F& :+P.^5g-=&9P':./P:bA,&B*,6/4'7o-'8:S$4'6EN P'&:fŠ ‰–M

3. BACKGROUND ON STRING PROCESSING †v&GP':S$C*,&(*,&:>P*&BH&Bpo>6/&4 AZ58AJ89:bAJ$â6/4 A,&B*,48.G4+ 5&cq+A,$&B*FA,$84A,$&@*,+ +AVc%$8:C8AC./&V8:bA A[q+o>J$'6/.E5*,&4|845D&8>J$g&V57&;6E:I.E8-=&B.E&5D[I6/A,$Y8i4'+4&0F(A"Q :S$'8*J8>SA,&B*VM#%$'&OABAVcA,$'&w>B+0F('PAJ8A,6/+4i+2=A,$&w:bAJ8A,6/:bA,6/>:q+2 8./._:V8A,6/+4D+2s:+4':bAJ84 Aj6E4:bA,&V85a+2q7*,+V[I6/4'7DA,+@6E4346/A"Qg[I$'&B4 µ 7+ &:;A,+g6/43'4'6?AQ M]#%$&BQ{8*,&FA,$'&B*,&S2h+*,&F+2I./6/AB&:BA,.?Q‚A,$'&Y:+*,&cj8:B>P*&: 6/4'7 ú Ì ú ÇÅåœ | +ª 2 º„6/4{A,$&W*,&B2l&B*,&4'>B&F:bASA,+* {AJ8€&:%>8*,&Z+2s+2s85 ûbP:bA,6/4'7iA,$&)+ >B>P*&: -8:S$'86/4':ZKL:BA,6/+4tø MEÞVRSMg#%$6/:i:bA+*B&: Ë Ì KNºjRT845 ñ NK ºjRSMZ#%$'&9(*,&B(*,+ >&:B+0F('./&Bz6?AQ@5&(=&45:w+4DA,$&jA,6/0F&jAJ8€&4 ñ NK ºZRSM\`"2fA,$&{8./7+*,6?A,$0 [%+P'.E5u-=&Y6/0F('./&BH A,+t>+0F('PA,& èF 0F&4 A,&V5;8:6/4i¤s6E7P*,&Iø cA,$'&TA,6/0F&q>+0F('./&Bz6?AQf[%+P'.E5i-=&q:6/&4 A,.?Q èF & ·!&0F0C8%ø M/Þ 84'5jA,$'&:+0i-'6/48A,+*,6E8.=848./Q:+4'>6/:V8.E. ;ˆ#‡6ˆ#h Mg#%$'&C-8:@8./7+*,6?A,$0ô6/:;:J$'&V5{6E4v#s8H -'./4 & aM

21

~

1

Fˆ#‡6ˆ#hu6/:o4'+AC84‚8>S*,+4 Q0gMe`"AF6/:F8a('P'4e+4A,$&@2L8>BA A,$8AqA,$&Z$'&8*B4 & „qP*&F58AJ8-'8:+4:bABA,&5{[I6?A,$ A,$'&C2h&8A,P*,&@[I6E4'5+V[ ./&4'7A,$ { | 84'5eA,$&g8.E($8-=&BAF:84C-=&I6^5&4 A,6?3&V5W- Qi78A,$'&B*,6/4'7;:bAJ8A,6E:bA,6/>: 8-=+PACA,$'&g56/:bA ] ?    Öh “i , Œ@= R ŽT Âi ,  “    iŒ“) iÿ[ K;h “ƒV Ñ ) “  B%a ) Ž(| "    F”  Œ@| " ŽsŒ  ,9ŒH  Œ!+= R  )98 HH  QBO ‘á‘ Œ“ƒ  •aŒ!+"A )    &  +=  “‚Œ !0/ ”l  ¦  • "? Œ l!  Ý )   1 “*1 ) !  ›Œ'“ƒ  •{Œ !+"o T Œ'“*)ðh “ƒ ' / )** & %Ö )  =  •{^"? Œ h! YŒl! Œ%  •{XF 8 ‹ KH  ìKD(]…/nª   |Œ%/ *'  Œ &6Uh FŒ h!  As  “ŒŽFl! a A )    & F  YŒ'“ƒ  •{Œ !+"8  Œ ' % ¨ Œ'“Ϗ%o , ]Œ '=  “”\  Œ ƒ?g    )*  Œ'0=  “$?   YŒ'“ƒ  •{Œ !+"

;KJLJCM7N O; ,PQM7NB

I=

B

B

R"B

8./7+*,6?A,$'0gc:6/&4 AVM‡)+[q&U&B*w[q& $8VU&%8.?*,&V85 QW4+A,&V5;A,$8ATA,$&qA,6/0F&%>+0F('./&Bz6?AQ;+2 Fˆ#‡6ˆ#h 6/:I.E6/4&V8*)6/4@A,$'&f:+4':bABA,&V5o8j*,&B2l&B*,&4'>B&)58AJ8:B*,&V8A,6E4798 :+4:B&W58AJ8:+*& 8@*,&:bABA,6/+4+4XA,$'&F0C8z6E0iP'0Ó84'5X0F6/4'6/0WP0ÓA,+./&B*J8-'./& U8./P'&:c=845DA,+C:>&(AJ8-'./&9-=&$8VU 6/+*VMI#%$6/:Z(*J8>BH A,6/>&D6/:W*,&S2h&B*S$Ï6/4 U+./U&: ì ŸhçSïBé,è"òö œ ïBÿHSïJë'èBAf2hPA,P*,&WU8./P&:MorG4 Q|56/:B*,&B(84>BQa-=&BH A[q&&4XA,$&WA[q+DA,$8A;&Bz>&&V5:98D>&B*&C>V84{-=& 5&B4'+A,&5v8:9:BA,6/+498./7+*,6?A,$'0gc84'5fA,6/0F&:BA,6/+4@6E:G894'+A,+*,6/+P':P.?AI(*,+-./&0gMT#%$ P':c A,$'6/: 0F&BA,$+ 5CA,&4'5:qA,+i$8VU&j(=+ +*G:BA,6/U 6/A"Q c(*,+ 5P>6/4'7i0C84 Q 2L8./:V85&M¦‡)+[q&U&B*A,$&w5&B3'4'6?H A,6/+4C+2!:S${6E:;4+A;:BA,6/4'7;P'4 P:J$v8:f[w8VU&./&BA,:984'5X¤'+P*,6/&B*i5&B:B*,6/(A,+*,:c¦845{8AJ$8*J8>BA,&B*,6/…&a[I$'6/>S$ÖA,&>J$'46EN P&g[%+*,€:@-=&:bA@+4Ö[I$'6/>J$ 5'8AJ8:V8P:B&4H A