Improvement on ARC-BC Algorithm in Text Classification Method

Abstract

With the rapid development of automatic text clustering and classification, many techniques and algorithms studying have been made focused in the field of text categorization. However, there is still much work to be done for improving the effectiveness of these classifiers, and new models need to be examined. This paper introduce an ARC-BC algorithms by using of association rule mining in text categorization systems, and proposes some new concepts and improvements on ARC-BC. The experimental results show that the training time of association-rule-based classifier is comparable to other well-known text classifiers, but classification quality is slightly lower than KNN algorithm. Moreover, the improvement proposed here can well improve the classification quality and can shorten training time. In all, our investigation leads to conclude that association rule mining is a good and promising strategy for efficient automatic text categorization, and it has a large room to enhance.