论文信息 - Anti-spam filtering: a centroid-based classification approach

Anti-spam filtering: a centroid-based classification approach

Nowadays, electronic mail is the most popular and convenient way for communication in daily life. Spam or junk E-mails are also increasingly appearing in the mail box from commercial Web sites. Therefore, we investigated the way to filter these junk e-mails through a variety of techniques, i.e. naive Bayesian, k-nearest neighbor and centroid based approach. We found that the centroid-based approach is the most suitable for the mail filtering application with 83.00 % of correctness. The outcome of our research has been successfully implemented as an intelligent Web mail service with the anti-spam mail filtering feature plug-in.

N. Soonthornphisaj | K. Chaikulseriwat | Piyanan Tang-On

[1] Thorsten Joachims,et al. A Probabilistic Analysis of the Rocchio Algorithm with TFIDF for Text Categorization , 1997, ICML.

[2] Yiming Yang,et al. An Evaluation of Statistical Approaches to Text Categorization , 1999, Information Retrieval.

[3] C. J. van Rijsbergen,et al. Information Retrieval , 1979, Encyclopedia of GIS.

[4] Martin F. Porter,et al. An algorithm for suffix stripping , 1997, Program.

[5] Mehran Sahami,et al. Learning Limited Dependence Bayesian Classifiers , 1996, KDD.

[6] David D. Lewis,et al. Naive (Bayes) at Forty: The Independence Assumption in Information Retrieval , 1998, ECML.

[7] Prasad Tadepalli,et al. Active Learning with Committees for Text Categorization , 1997, AAAI/IAAI.

[8] William W. Cohen. Learning Rules that Classify E-Mail , 1996 .