Rough clustering for web transactions
Grouping web transactions into clusters is important in order to obtain better understanding of user's behavior. Currently, the rough approximation-based clustering technique has been used to group web transactions into clusters. It is based on the similarity of upper approximations of tran...
| 第一著者: | |
|---|---|
| フォーマット: | Dissertation |
| 言語: | 英語 英語 英語 |
| 出版事項: |
2011
|
| 主題: | |
| オンライン･アクセス: | http://eprints.uthm.edu.my/2629/ |
| Abstract | Abstract here |
| 要約: | Grouping web transactions into clusters is important in order to obtain better
understanding of user's behavior. Currently, the rough approximation-based
clustering technique has been used to group web transactions into clusters. It is based
on the similarity of upper approximations of transactions by given threshold.
However, the processing time is still an issue due to the high complexity for finding
the similarity of upper approximations of a transaction which used to merge between
two or more clusters. In this study, an alternative technique for grouping web
transactions using rough set theory is proposed. It is based on the two similarity
classes which is nonvoid intersection. The technique is implemented in MATLAB
®
version 7.6.0.324 (R2008a). The two UCI benchmark datasets taken from:
http:/kdd.ics.uci.edu/ databases/msnbc/msnbc.html and
http:/kdd.ics.uci.edu/databases/ Microsoft / microsoft.html are opted in the
simulation processes. The simulation reveals that the proposed technique
significantly requires lower response time up to 62.69 % and 66.82 % as compared to
the rough approximation-based clustering, severally. Meanwhile, for cluster purity it
performs better until 2.5 % and 14.47%, respectively. |
|---|