Applying learning to filter text

• Main Author: Sainin, Mohd Shamrie
• Format: Conference or Workshop Item
• Language: English
• Published: 2005
• Subjects: QA76 Computer software
• Online Access: http://repo.uum.edu.my/12431/1/rie%20seit3_ver5.pdf; http://repo.uum.edu.my/12431/

Text filtering has been a successful application especially in e-mail filtering. The use of probabilistic approaches such as naïve Bayes algorithm is the effective algorithms currently known for learning to filter or classify text document.Naïve Bayes algorithm is one of the algorithms in Machine Learning that manipulates probability estimation or reasoning about the observed data.The growing of bulk e-mail or known as spam e-mail becomes a threat to users’ privacy and network load and in the case of e -mail filtering,naïve Bayes classifier can be trained to automatically detect spam messages.Similar to the e-mail, forum application may be misused by the user to send bad messages and in some extent may offence other readers.Forum filtering may be less important compared to e-mail spam filtering; however there is a possibility of using naïve Bayes to learn the messages and automatically detect bad messages.Most of the forum application found in the web is applying keyword based text filtering which scan the words and change the detected words into certain representation.Instead of defining a set of keywords to filter the forum messages, this paper will explains the experiment in applying a learning to filter text especially in the educational and anonymous forum message, where there is no user registration required to submit messages.