Conference Paper

Practical Application of Associative Classifier for Document Classification.

DOI: 10.1007/11562382_36 Conference: Information Retrieval Technology, Second Asia Information Retrieval Symposium, AIRS 2005, Jeju Island, Korea, October 13-15, 2005, Proceedings
Source: DBLP

ABSTRACT In practical text classiflcation tasks, the ability to interpret the classiflcation result is as important as the ability to classify exactly. The associative classifler has favorable characteristics, rapid training, good classiflcation accuracy, and excellent interpretation. However, the associative classifler has some obstacles to overcome when it is applied in the area of text classiflcation. First of all, the training process of the as- sociative classifler produces a huge amount of classiflcation rules, which makes the prediction for a new document inefiective. We resolve this by pruning the rules according to their contribution to correct classiflca- tions. In addition, since the target text collection generally has a high dimension, the training process might take a very long time. We propose mutual information between the word and class variables as a feature selection measure to reduce the space dimension. Experimental classi- flcation results using the 20-newsgroups dataset show many beneflts of the associative classiflcation in both training and predicting.


Full-text (2 Sources)

Available from
Jun 4, 2014