Scientific Article Recommendations with Mahout

Presentation by Kris Jack PhD, Senior Data Mining Engineer @Mendeley at Data Science London 23/05/12

Transcript

1. !"#$%&#'#"()*&#"+$( ,$"-..$%/0&#-%( 1#&2(302-4& 5*#6(70"89(:2; !$%#-*(;0&0(3#%#%<(=%<#%$$*
2. !"#$%&"# !(>--/(*$6$0*"2$*6(0*$(-%(&-?(-'(&2$#*(<0.$ !(@0*<$(0.-4%&(-'(*$6$0*"2(?*-/4"$/ !(A08$6(&#.$(&-(<$&(0&(120&(B-4(%$$/ !(C$+?(*$6$0*"2$*6(DB(*$"-..$%/#%<(*$+$E0%&(*$6$0*"2
3. FGH(.#++#-%I(46$*6J(&2$(KL(+0*<$6&(46$*(D06$6M (N%#E$*6#&B(-'(O0.D*#/<$ !&0%'-*/(N%#E$*6#&B 3PA N%#E$*6#&B(-'(3#"2#<0% C0*E0*/(N%#E$*6#&B N%#E$*6#&B(-'(QR'-*/ !0-(:04+-(N%#E$*6#&B P.?$*#0+(O-++$<$(@-%/-% N%#E$*6#&B(-'(=/#%D4*<2 O-*%$++(N%#E$*6#&B N%#E$*6#&B(-'(O0+#'-*%#0(0&(S$*8$+$B ,TAC()0"2$% O-+4.D#0(N%#E$*6#&B >$-*<#0(A$"2 N%#E$*6#&B(-'(T#6"-%6#% NO((!0%(;#$<- N%#E$*6#&B(-'(O0+#'-*%#0(0&(@) N%#E$*6#&B(-'(U+-*#/0 N%#E$*6#&B(-'(V-*&2(O0*-+#%0 HL.(*$6$0*"2(0*&#"+$6
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
6. !(C-1(/-$6(302-4&W6(*$"-..$%/$*(1-*8X !(C-1(1$++(/-$6(#&(?$*'-*.(-4&(-'(&2$(D-RX !(C-1(1$++(/-$6(#&(?$*'-*.(0'&$*(6-.$(&4%#%<X Y4$6&#-%6
7. 302-4&W6( ,$"-..$%/$*
8. >$%$*0&#%<(*$"-..$%/0&#-%6( &2*-4<2(.0&*#R(.4+&#?+#"0&#-% A2#6(#6(#&$.ZD06$/( *$"-..$%/0&#-%6(06( 6#.#+0*#&B(#6(D06$/(-%( #&$.69(%-&(46$*6 !"#$%&%'()$*%(!+,$'-$,%.,)$(%/!!&$0,)*$1)'!**)2/)"3!4
9. !"#$%& '())(&* +$%,-*$% .*/-0% 1023456$47 89:,$6,47 ,$6$0*"2()*&#"+$6 ,$6$0*"2$*6 89:,$6,4; 1023456$4; P%?4&([0++(46$*(?*$'$*$%"$6\
10. !"#$%& '())(&* +$%,-*$% .*/-0% 1023456$47 89:,$6,47 ,$6$0*"2()*&#"+$6 ,$6$0*"2$*6 89:,$6,4; 1023456$4; FGH3 HL3 P%?4&([0++(46$*(?*$'$*$%"$6\ ]LL3( ?*$'6
11. ,$6$0*"2( )*&#"+$6 ,$6$0*"2$*6 )++(N6$*(:*$'$*$%"$6( [#&$.(R(46$*\ FG(:*$?G(?*$'G(.0&*#R([FZ]\ KG(>$%G(6#.G(.0&*#R([^Z_\ ]G(34+&#?+B(.0&*#"$6([`ZFL\ #&$.G,$"-..$%/$*7-D
12. ,$6$0*"2( )*&#"+$6 A4*#%< )(N6$*W6(:*$'$*$%"$6 [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2$*6 )++(N6$*(:*$'$*$%"$6( [#&$.(R(46$*\ FG('(#)*$)(#+*$,&-(./$01234 KG(>$%G(6#.G(.0&*#R([^Z_\ ]G(34+&#?+B(.0&*#"$6([`ZFL\ #&$.G,$"-..$%/$*7-D
13. ,$6$0*"2( )*&#"+$6 A4*#%< )(N6$*W6(:*$'$*$%"$6 [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2$*6 )++(N6$*(:*$'$*$%"$6( [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2( )*&#"+$6 ; 7 7 7 < < < < ; ; ; ; < < < < P&$.(!#.#+0*#&B( [#&$.(R(#&$.\ FG(:*$?G(?*$'G(.0&*#R([FZ]\ KG(5#6*$".,*$,&-(./$07284 ]G(34+&#?+B(.0&*#"$6([`ZFL\ #&$.G,$"-..$%/$*7-D
14. ,$6$0*"2( )*&#"+$6 A4*#%< )(N6$*W6(:*$'$*$%"$6 [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2$*6 )++(N6$*(:*$'$*$%"$6( [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2( )*&#"+$6 ; 7 7 7 < < < < ; ; ; ; < < < < P&$.(!#.#+0*#&B( [#&$.(R(#&$.\ ,$6$0*"2( )*&#"+$6 A4*#%< ,$"-..$%/0&#-%6 [#&$.(R(46$*\ = > FG(:*$?G(?*$'G(.0&*#R([FZ]\ KG(>$%G(6#.G(.0&*#R([^Z_\ 3*$9:;-.);<$,&-(.=#"$0>21?4 #&$.G,$"-..$%/$*7-D
15. C-1(1$++(/-$6 #&(1-*8X
16. 3$%/$+$B(!4<<$6&
17. ,4%%#%<(-%().0a-%W6(=+06&#"(30?(,$/4"$ ?%4@*2(%@4",*4(%@4*(,:4-0460,-
18. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL 302-4&W6 :$*'-*.0%"$
19. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ 302-4&W6 :$*'-*.0%"$
20. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ 302-4&W6 :$*'-*.0%"$
21. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ 302-4&W6 :$*'-*.0%"$
22. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ ] 302-4&W6 :$*'-*.0%"$
23. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ ] 302-4&W6 :$*'-*.0%"$
24. @$&W6(&4%$(#&d
25. 1*$@#A:=#$)(B=#"".6C$-.,# D*$E,)(BF#$G:&;.-<
26. 1*$@#A:=#$)(B=#"".6C$-.,# !(302-4&W6(*$"-..$%/$*(#6(0+*$0/B($''#"#$%& !(S4&(B-4*(/0&0(.0B(20E$(4%4640+(?*-?$*&#$6 !(C0/--?(.0B(%$$/(0(2$+?#%<(20%/ !(@$&W6(6$$(120&W6(<-#%<(-%GGG
27. A068()++-"0&#-% AB490"#,4-046023C*-* 74#*@"6*#4(CC06(-*@D4@*,3$-*49(E$%&4FG4(E($C()C*HHH
28. A068()++-"0&#-% I0)H&*-10%J$&"#(-$0%KLH,*-K M2(3#*@H2(NH,3C$-H,$O*MD 5-#$%&HE(C"*?JK,3C$-5$O*LLP )++-"0&#%<(.-*$(.0??$*6(-%(0(?$*(e-D(D06#6 I0)H&*-10%J$&"#(-$0%KLH,*-Q%-K M2(3#*@H#*@"6*H-(,R,MD %"2S*@"6*#,LP )++-"0&#%<(.-*$(*$/4"$*6(-%(0(?$*(e-D(D06#6
29. A068()++-"0&#-% AB490"#,4-046023C*-* 7F490"#, T#02474U4F<4 #*@"6*#,
30. :0*&#&#-%$*6 7F490"#,4-046023C*-*
31. :0*&#&#-%$*6 7F490"#,4-046023C*-* HI?JK HI??9K
32. Q%3"-5(23C*#H5(23C*#VQ%-W#$-()C*D4!*N-X4,(23C*#4> !"#4Q%3"-5(23C*#HS(%@025(23C*#VQ%-W#$-()C*D4!*N-XKHHHLP Q%3"-5(23C*#H/#$-*8(#-$-$0%T$C*K60%JD4,(23C*#LP 60%JH,*-8(#-$-$0%*#1C(,,K!0-(C?#@*#8(#-$-$0%*#H$%&''LP 2&&?Mbb111G?2#+#??$0/e#.0%G"-.bD+-<bKLLfbFKbKLb20/--?Z&4&-*#0+Z 6$*#$6Z#664$ZKZ<$&&#%<Z6&0*&$/Z1#&2Z"46&-.#a$/Z?0*&#&#-%#%<b
33. :0*&#&#-%$*6 7F490"#,4-046023C*-* ;490"#, +E*%C:4 @$,-#$)"-*@
34. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ ] 302-4&W6 :$*'-*.0%"$
35. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH ] 302-4&W6 :$*'-*.0%"$
36. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH ] 27*1J 083L4 302-4&W6 :$*'-*.0%"$
37. D*$E,)(BF#$G:&;.-< !(302-4&(?*-E#/$6(#&$.ZD06$/(OU !(T$(20E$(.0%B(.-*$(#&$.6(&20%(46$*6 !(AB?#"0++B9(46$*ZD06$/(#6(.-*$(0??*-?*#0&$ ! !-(+$&W6(.08$(-%$d
38. ,$6$0*"2( )*&#"+$6 A4*#%< )(N6$*W6(:*$'$*$%"$6 [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2$*6 )++(N6$*(:*$'$*$%"$6( [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2( )*&#"+$6 ; 7 7 7 < < < < ; ; ; ; < < < < P&$.(!#.#+0*#&B( [#&$.(R(#&$.\ ,$6$0*"2( )*&#"+$6 A4*#%< ,$"-..$%/0&#-%6 [#&$.(R(46$*\ = > FG(:*$?G(?*$'G(.0&*#R([FZ]\ KG(>$%G(6#.G(.0&*#R([^Z_\ ]G(34+&#?+B(.0&*#"$6([`ZFL\ #&$.G,$"-..$%/$*7-D
39. ,$6$0*"2( )*&#"+$6 A4*#%< )(N6$*W6(:*$'$*$%"$6 [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2$*6 )++(N6$*(:*$'$*$%"$6( [#&$.(R(46$*\ ,$6$0*"2( )*&#"+$6 ,$6$0*"2( )*&#"+$6 ; 7 7 7 < < < < ; ; ; ; < < < < P&$.(!#.#+0*#&B( [#&$.(R(#&$.\ ,$6$0*"2( )*&#"+$6 A4*#%< ,$"-..$%/0&#-%6 [#&$.(R(46$*\ = > FG(:*$?G(?*$'G(.0&*#R([FZ]\ KG(>$%G(6#.G(.0&*#R([^Z_\ ]G(34+&#?+B(.0&*#"$6([`ZFL\ #&$.G,$"-..$%/$*7-D :"#( !"#($M.,.;&(.-<$0:"#($/$:"#(4 @#"#&(=N#(" @#"#&(=N#("
40. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH ] 302-4&W6 :$*'-*.0%"$
41. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH Q*#<G(46$*ZD06$/ !F59(KGH ] 302-4&W6 :$*'-*.0%"$
42. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH Q*#<G(46$*ZD06$/ !F59(KGH ] 21*7J 0IOL4 P1$08>L4 302-4&W6 :$*'-*.0%"$
43. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH Q*#<G(46$*ZD06$/ !F59(KGH ] O46&G(46$*ZD06$/ !LG]59(KGH 302-4&W6 :$*'-*.0%"$
44. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH Q*#<G(46$*ZD06$/ !F59(KGH ] O46&G(46$*ZD06$/ !LG]59(KGH 2?*>J 0>?L4 302-4&W6 :$*'-*.0%"$ 27*1J 083L4
45. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH Q*#<G(46$*ZD06$/ !F59(KGH ] O46&G(46$*ZD06$/ !LG]59(KGH 28*DJ 0QIL4 302-4&W6 :$*'-*.0%"$ P1$08>L4
46. O-%"+46#-%6
47. %B6=;:".B6" !(302-4&(#6(/-#%<(0(<*$0&(e-D(-'(?-1$*#%<(3$%/$+$B(!4<<$6& ! @0*<$(6"0+$(/0&0(6$& ! >--/(g40+#&B(*$"-..$%/0&#-%6 !(A4%#%<(2$+?6 ! C$+?(C0/--?(1#&2(&068(0++-"0&#-%(#'(%$"$660*B ! :0*&#&#-%(B-4*(/0&0(0??*-?*#0&$+B ! T$(60E$(fHh(*$6-4*"$6 !(N6$(0%(0??*-?*#0&$(0+<-*#&2. ! P&$.Z(E6(46$*ZD06$/([3)CQNAZFLL^\ ! T$(#%"*$06$(?*$"#6#-%(DB(__G_h
48. V-*.0+#6$/().0a-%(C-4*6 V-G(>--/(,$"-..$%/0&#-%6bFL L F5 K5 ]5 ^5 H5 _5 `5 L LGH F FGH K KGH O-6&+B(c(S0/ O-6&+B(c(>--/ O2$0?(c(S0/ O2$0?(c(>--/ _GH59(FGH Q*#<G(#&$.ZD06$/ O46&G(#&$.ZD06$/ !KG^59(FGH Q*#<G(46$*ZD06$/ !F59(KGH ] O46&G(46$*ZD06$/ !LG]59(KGH 28*DJ 0QIL4 302-4&W6 :$*'-*.0%"$ P1$08>L4 2&&?Mbb111G.$%/$+$BG"-.b?*-'#+$6b8*#6Ze0"8b