A New Method to Compute the Word Relevance in News Corpus

In this paper we propose a new method to compute the relevance of term in news corpus. According to the characteristics of news corpus, we first propose that the news corpus should be divided into different channels, second we make use of the feature of news document, we divide the co-occurrence of terms into two cases, on the one hand the co-occurrence in the title of the news, On the other hand the co-occurrence in the news text, we use different methods to compute the co-occurrence. In the end, we introduce the web corpus Wikipedia to overcome some shortcomings of the news corpus.