Abstract—Because of the popularity of the Internet, customer care has been transferred to the Internet-based or Web-based systems. Online discussion forums are common methods used in electronic Customer Relationship Management. However, one emerging problem is that normally companies offer a series of categories of forums to be discussed. Posting messages to incorrect categories may delay the response time and it may take several trials to redirect the messages to the right person in charge or right category for discussion. In this study, we propose the use of text categorization approach to automatically select a target forum category. The empirical evaluations demonstrate the utility of text categorization approach. We also found that decision tree outperformed other machine classifiers.

Index Terms—Text mining, text classification, machine learning, forum selector

C. H. Chou is from College of Charleston, SC, USA (e-mail:chouc@cofc.edu).

Cite: Chen-Huei Chou, "Empirical Evaluations of Automatic Forum Selector," International Journal of Computer and Communication Engineering vol. 1, no. 1, pp. 55-58 , 2012.