Exploring Association Rules in EDOC Server Using WEKA: Analysis and Insights
250 likes | 358 Vues
This document delves into the application of Association Rules generated from EDOC server data using WEKA, conducted by Altuğ KUL. It outlines a comprehensive taxonomy capturing various definitions and attributes, including session data, and discusses the challenges faced, such as low support and the generation of rules with false statements. With a total of 12,336 sessions analyzed, the study highlights effective attributes and association rules derived from high support levels, emphasizing the practical implications for data mining and analysis.
Exploring Association Rules in EDOC Server Using WEKA: Analysis and Insights
E N D
Presentation Transcript
Web Mining 03/04 (Undirected) Association Rules in EDOC Server generated using WEKA by Altuğ KUL
Agenda • The Taxomony • The ARFF File • WEKA • Association Rules
The Taxonomy • The first regex file had more than 100 definitions • The second had 51 definitions • The third had 21 definitions • The fourth had 13 definitions • The fifth had 11 definitions • The sixth had 8 definitions • The final one (!) had 7 definitions
The Taxonomy • HOME • AUTHOR • SEARCH • DSS-LOOKUP • DSS-DISS • DSS-ABSTR • OTHER
The ARFF File @relation log @attribute HOME {1,0} @attribute AUTHOR {1,0} @attribute SEARCH {1,0} @attribute DSS-LOOKUP {1,0} @attribute DSS-DISS {1,0} @attribute DSS-ABSTR {1,0} @attribute OTHER {1,0} @data 1,0,0,0,0,0,1 1,0,0,0,0,0,1 0,0,0,1,0,0,0 …
WEKA • Some of the pages are deleted due to low support, like HOME-E • The problem with the WEKA is that it creates rules containing only FALSE statements. • Total of 12336 sessions. • 7 Attributes • Min support = 0.5 & min conf. = 0.9 • Max 1119 rules created.
WEKA • HOME: 1 4399, 0 7937 ------------- 35 % • AUTHOR: 1 1738, 0 10598 ------- 14 % • SEARCH: 1 311, 0 12025 --------- 3 % • DSS-LOOKUP: 1 5531, 0 6805 -- 45 % • DSS-DISS: 1 585, 0 11751 ------- 5 % • DSS-ABSTR: 1 706, 0 11630 ---- 6 % • OTHER: 1 3408, 0 8928 ----------- 28 %
Association Rules HOME=0 AUTHOR=0 DSS-LOOKUP=1 4374 ==> SEARCH=0 4374 conf:(1)
Association Rules HOME=0 OTHER=1 2244 ==> SEARCH=0 2244 conf:(1)
Association Rules HOME=0 AUTHOR=0 SEARCH=0 DSS-LOOKUP=0 DSS-DISS=0 DSS-ABSTR=0 2069 ==> OTHER=1 2069conf:(1)
Association Rules HOME=0 DSS-LOOKUP=1 4389 ==> SEARCH=0 4388 conf:(1)
Association Rules HOME=0 DSS-LOOKUP=1 DSS-ABSTR=0 4309 ==> SEARCH=0 DSS-DISS=0 4306 conf:(1)
Association Rules HOME=1 DSS-DISS=0 4007 ==> DSS-ABSTR=0 3863 conf:(0.96)
Association Rules DSS-DISS=0 11751 ==> DSS-ABSTR=0 11322 conf:(0.96)
Association Rules DSS-LOOKUP=1 5531 ==> SEARCH=0 DSS-DISS=0 OTHER=0 5318 conf:(0.96)
Association Rules DSS-LOOKUP=1 5531 ==> AUTHOR=0 DSS-DISS=0 OTHER=0 5303 conf:(0.96)
Association Rules OTHER=0 8928 ==> DSS-DISS=0 8535 conf:(0.96)
Association Rules DSS-LOOKUP=1 5531 ==> DSS-ABSTR=0 5278 conf:(0.95)
Association Rules HOME=1 DSS-ABSTR=0 4044 ==> SEARCH=0 3852 conf:(0.95)
Association Rules DSS-LOOKUP=1 5531 ==> SEARCH=0 DSS-ABSTR=0 5259 conf:(0.95)
Association Rules DSS-LOOKUP=1 5531 ==> SEARCH=0 DSS-DISS=0 DSS-ABSTR=0 5247 conf:(0.95)
Association Rules HOME=1 4399 ==> SEARCH=0 4169 conf:(0.95)
Association Rules HOME=0 7937 ==> SEARCH=0 DSS-ABSTR=0 7505 conf:(0.95)
Association Rules OTHER=1 3408 ==> DSS-DISS=0 DSS-ABSTR=0 3187 conf:(0.94)
Association Rules HOME=1 4399 ==> DSS-ABSTR=0 4044 conf:(0.92)