• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Data Mining Data Mining
 

Data Mining Data Mining

on

  • 720 views

 

Statistics

Views

Total Views
720
Views on SlideShare
720
Embed Views
0

Actions

Likes
1
Downloads
8
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Data Mining Data Mining Data Mining Data Mining Document Transcript

    • ¡  © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ Model Model Data Support Mining Support Mining Decision Data Decision Data DM, then DS DM & DS in Data Pre-Processing Sequential Application: –Meta-learning and multi-strategy learning Model Model –ROC methodology Improving models by data analysis • MS Analysis Services • e.g.: MS OLE DB for DM • Supporting decisions in the DM process, Incorporating DM methods into DSS, e.g.: Support Mining Decision Data Expertise Data Support Mining Support Mining Decision Data Decision Data Models "Decision Support for Data Mining" "Data Mining for Decision Support" Integrating DM and DS through $0 0$ $0( ! 0$ !† …  0$ t ' s " …  ! 0$ !t ' s " © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ U3`RcbR dR1VX €€€ u Decision Support 89@6I 56BQPv8PE9 FPD u A5B89@6I 9QvpBFQQPoBQ8vI 89@6I 56BQPv8PE9 FPD u ? C98vF 56BQPBD6CCP u dR1VX knfm elik gjihf fgfe @9CG8P5P CF9I6QCvD u CFPD H6 C9yGQ u Data Mining @9F9QCv8D CF9I6QCvD u x„€ƒƒ‚ {€z z ~}v |{z{zy w xwv PQP@ I6FH … … “’‘‰ˆ‡† ’†•“” ’”–‰ d™˜“— trshriq GF9E6DCB@ 9A@987654  ! ! u  ! 0$ !t ' s "   &0  „$ƒ) ‚  ƒ  © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ U3Ued2W2 WVTUTSR1 aU3`RcbR aU`RY2X WVTUTSR1 dR1VX €€€ u Decision Support 56BQP5P8yx9 u 56BQPwB8PvCBE u A5B889@6I CBCG8P5P u 56BQPv8PE9 u !0))(' &%$ #     # "   !   A5BF9QCv8D u 2321 dR1VX 56BQPDBHBCCP8D u     tgshriq pi hgf Data Mining PQP@ I6FH GF9E6DCB@ 9A@987654
    • ‡  © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ false positive rate false positive rate 100% 80% 60% 40% 20% 0% 100% 80% 60% 40% 20% 0% 0% 0% slope = 4 2 = 2 20% 20% Pos =4 40% 40% Neg 60% 60% true positive rate true positive rate FNcost 80% 80% 2 = FPcost 1 100% 100% 0$ ¾ && % 0$ ¾&& ƒ%  & &ø $  % $ ƒƒ(ø ‚$ ù% ÂÁ $ø ÷ © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ False positive rate 100% 80% 60% 40% 20% 0% 0% 20% false positive rate ðïîí ñ ëê ÕÔÓ Ò 100% 80% 60% 40% 20% 0% 0% 40% ðïîí ì ëê ÕÔã Ò CN2 classifier 2 W RAcc classifier 1 §š” Çé ÎÈÊ èçæ à 20% 60% Confirmation rules True positive rate ÝÙ × ÙÛÚÙ × Þ ÕÔã Ò 40% 80% ÝÙå × ÙÛÚÙä × Ö ÕÔã Ò 60% 100% ¸ ˜’Ð Ñ ¸ §•³ ’§—š¥Ð true positive rate 80% Ï ÇÌ ÎÅ ÇÍËÌËÊÉ È ÇÊâ Îá à 100 70 30 50 50 0 Negative examples 50 20 30 Positive examples ÝÙà × ÙÛÚÙß × Þ ÕÔÓ Ò 100% Predicted negative Predicted positive 100 50 50 ÝÙÜ × ÙÛÚÙØ × Ö ÕÔÓ Ò Classifier 2 50 40 10 Negative examples 50 10 40 Positive examples Predicted negative Predicted positive ¸ §•³Ð Ñ ¸ §•³ ’´œ “Ð Classifier 1 Ï ÇÌ ÎÅ ÇÍËÌËÊÉ È ÇÆÅÄ Ã ƒƒ(ø ‚$ ù% ÂÁ $ø ÷ ƒƒ ø  ÂÁ ø ÷ $% )& ÂÁ ÂÁ öõíôã ÕòóòÔ © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ §•ž§’‘ž– ‘’—¶•œ ³ “’œ ’¥¥ž– ˜žœ ’–ž§•©  ¼»º² ž ž¹¢£ ·¸ ˜¸ ’ · §—’–•‘ ”˜´•œ ”“ ˜žž¶‘•©  §’¨œ ´© Ž ·¸ ˜¸ ’ Œ±{€z xw ~z°ƒƒw °}°°w„wƒ® ‹v µ{w yv‰ ˆ §§’©•œ ³ Ÿ£ ’”“ ž”“žÀ “œ •³³´§ •ž§ž©’£ Ÿ£ œ •¥ ˜ž§§’©•œ ³›’œ ³ š“š– ž §—’–•‘ ˜ž§´  ¬ Œ±{€z xw ~z°ƒƒw °wz x{}‚¯} ® ‹yv {}­ x v‰ ˆ ”©š•œ ³³š –š² ›“¦›§•ž§ž©’£  ¬ Œ±{€z xw ~z°ƒƒw °wz x{}‚¯} ® ‹v {}­ x yv‰ ˆ «š“š– ‘•œ ¥ª •ž§ž¨’œ —’–•‘  §’©ž¨œ ’  §ž§™—š¦  Ÿ  Ÿ£ œ •¥ ¤£ ¢¡   Ÿ  Œ ‹v „€ Š yv‰ ˆ ˜žœ š’—›š“’‘  ™˜•—•–•”“’‘ Ž  x„€ƒƒ‚ {€z z ~}v |{z{zy w xwv Œ ‹yv „€ Š v‰ ˆ '  ' # "    "  ¿" 0 ¾ '½  ½
    • ú  © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ Model (Zupan & Bohanec) —’–•‘ ’“´¶žœ ““š›ž“—´‘ —š©ž”©œ šœ ’ž” ’¨ž“š“ž—š´8 §™šÀ “’œ ’¥¥ž– ž –’³•—’¨’– §—’–•‘›¶´§ ˜žž¶‘•© û vqXt…„ Wƒ S‚ ¼»º² –š ¹¢£ ™¶ «§ª—’–•‘ ¥• “’‘³•—’¨’– —’——šœ š³ ûY qYY xr x S€ 65 43 7 21 “œ ’³a’ ’”“ ™¶ –’ž¥’œ ™—“’´y’§¶´§ —’–•‘ –’³•—’¨’–›¼»º² ûY xtr qh Sw •ž§ž¨œ ’³´§ “œ ’³a’ œ ’–´ š“š– ‘•œ ¥ û vqut sr qpi h Sg š“š– ‘•œ ¥ û `YX W fedc Sb x„€ƒƒ‚ {€z z ~}v |{z{zy w xwv Expertise Data ’§ž“œ ’³a’ ‘•œ ¥ û `YX W VUT SR Q@PIHG BFE D@ CBA@9 ©§©© ©¨§¦ ¦ „ '  „ ¿' # " "½ © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ !  x„€ƒƒ‚ {€z z ~}v |{z{zy w xwv 0) $( ' &% #$#" ©§©© ©¨§¦ & ù$Á ƒ$#" „ ¿' 0 ¾ "½   Á ƒ # ¿   ½ „ ¿' # "½ © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ ëêïðï òÕ ôò¥ê¤ Ò false positive rate ðòõï Õòý ðïð£ôíë¢ ¡ Õò Õòý ÿþý ýü Ò 100% 80% 60% 40% 20% 0% 0% û §§’©•œ ³  £ ’”“ •“ž §–•”“’‘ Ÿ£ ˜ž©´–•œ “º slope = 4 8 = .5 20% Pos =4 40% Neg 60% true positive rate FNcost 80% 8 = FPcost 1 100% x„€ƒƒ‚ {€z z ~}v |{z{zy w xwv 0$ ¾ && % 0$ ¾&& ƒ%  & &ø ¿' 0 ¾ "½  ½
    • †  © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ • !  ‡s† nq ƒ zrjqt‚ wq nqm kj€mnqwyn~}p Ž !  znqm kyxw kvnqu tsrqlp |l{ nm onmnml j kji x„€ƒƒ‚ tsrql nqmvmysi ” !  {€z z ~}v |{z{zy …tj„ nq ƒ w xwv ƒ vtts} kwq‚‚x} nqmvmysi  ' $ &$  " $ ) (" ' $ø ! &ƒ$# " $ƒ) !ƒ(" ©h–g ¨  § –©  f©h–g ¨fe™d™˜§—˜—–© „& % )) $ 0 „& ! %ƒ)) ˆ ƒ$ƒƒ 0 ’ ’ ôí óò ë팋Š ‰íôˆ © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ Model • !  ” !  x„€ƒƒ‚ |{z{zy {€z z ~}v w xwv x„€ƒƒ‚ {€z z ~}v |{z{zy w xwv     ¿' $ø ! " !&0‰½ ¿  ø!  !  ½ ¿' $ø ! " !&0‰½ ¿  ø!  !  ½ „ ! %ƒ)) ˆ ƒ  !$(‡$' ! ƒ ˆ ƒ ! „ ! %ƒ)) ˆ ƒ  !$(‡$' ! ƒ ˆ ƒ ! © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ Model §’“´¶žœ ““š –’“šœ ’’‘ §’“´¶žœ ““š “´³º of pulse sensation changes Deformities Symptoms Amputations Ulcers Absence Loss of prot. Other Tests Present status History |{z{zy w xwv x„€ƒƒ‚ {€z z ~}v RISK  " $ ' &0 ¿" $ø ! ' !&0‰½  &&$%0’“$0’  !   ¿" $ø ! '½ ½ „ % )) $(‡$' „ ! %ƒ)) ˆ ƒ  !$(‡$'
    • ” “B’‘@G % $ % „ %$# % ˆ „ ! 0$ !t ' s " 0$ t ' s " IQB“B˜BP—• – •P “BAH• – D@ Q@PI•PABG °€€­ ~ ž„wz„œ °€€­ ~ ­|zŸ #$#" w¯} ® Œyv Œyv ­¬«Ÿ › š ™ › ¢ ¡   £ final achievement GA 1st grade c5 Û Õê Ø ¤ðòó틥íÕ¥ ¤ <=1 >1 c1 for lang 8th grade LEGEND: 2 Slovene gen ach 7th grade GA 1st grade - general achievement of the first high c2 school grade <= 3 >3 SSS Slovene - mark of subject Slovene language regular enrol History - mark of subject History GA 1st grade 5 for lang ß Õê å ¤ðòó틥íÕ¥ ¤¦ Physics - mark of subject Physcis age enrol - age at enrolment (in months) unex ab 3rd sem - unexcused absence in the third <= 2 >2 c7 c3 semester (hours) History 4 citizenshi p z ±°v Œv birth state ¥ò¥ëêôêՌ ¤§ <= 2 >2 c6 Physics 4 gen ach prim sch c4 <= 1 >1 math 8th grade age enrol unex abs 3rd sem phys 8th grade <= 180 > 180 <= 0 >0 4 1 4 2 òóíô ðôïí© ¤ª ëêêð ðôïí© ¤¨ {€z x~zµ}„œ © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ ’   ! ƒ ˆ ƒ ƒƒ  „ ! %ƒ)) ˆ ƒ$ƒƒ 0 ² ²  &00  ƒ  &(ƒ% ˆ Œ}„w }­ ~w€„ƒƒw v ¸ yv Models for Building Construction Models for Client Value Project Attributes Building Feasibility £Õíóëò¤òôŒ¤êõ Ò £Õíóëò¤òôŒŒ‹ð Ò Decision Support Data Mining ˆ w}„w ­ ~„w} }„ |{zƒ€°} º}µ µ{w ¹}¬ ˆ Œ {€z xw{z»€ ~ °w ~zƒ ž­ Á¼ À¿¾ ½¼ Ò ½ À¿¾ Á Building Feasible ¼ ¼ Ò Designs to ½ ÆÅÄà ÂÁ Building ¼ ¼ Ò maximise Designs Á¼ ÆÅÄà ½¼ Ò Client Value ½ ÈÆÇ Á ¼ ¼ Ò ˆ Œ {€z x}‚¯ {}ƒÉ Value Feasible ëêïóíÕ¥òóëï ýÍÎüÍ ©ê Ì´Õê·ò¤íÕ©Ë ëêïóíÊïôí¤Õê© Ò Shape Zone Zone ðòöõíêՌŒí ¥ëí ðòï¥êôê¥êöóò¤ ëꤤêõ Ò ëêïóíÊï¥Õí¥ëíóð Ò lity Size ua Q ï´ð·êÕóð¶ õïÕµ ³õòëíöêˆ ê´Õíü ³òô£êü ò òóý © £¨ ¦§ ¦¥¤ £¢ © £¨ ¦§ ¦¥¤ £¢ 