Self-Switching Classification Framework for Titled Documents

Abstract

Ambiguous words refer to words that have different meanings such as apple, window, etc. In text classification they are usually removed by feature reduction methods like information gain. Sometimes there are too many ambiguous words in the corpus that we cannot simply throw them away, especially when classifying documents from the Web. In this paper we look for a method to classify titled documents with the help of ambiguous words. Titled documents are a kind of documents that have a simple structure containing a title and an excerpt. News, messages, and paper abstracts with titles are such examples. Instead of introducing another feature reduction method, we describe a framework to make the best of ambiguous words in the titled documents. The framework improves the performance of traditional bag-of-words classifier with the help of a bag-of-word-pairs classifier. We implement the framework using one of the most popular classifiers, multinomial naive Bayes (MNB), as a case in point. The experiments with three real life datasets show that in our framework the MNB model performs much better than traditional MNB classifier and the naive weighting algorithm, which simply puts more weight on the title words.