A LDA Based Analysis for News Coverage Difference of Chinese Portals

Abstract

Portals can be important news resources for readers and researchers, they have different tastes of concern, which can affect readers or researches based on news corpus. Aiming to find the news coverage difference of three most popular portals in China, we conduct a research on portals news coverage analysis, which is based on the LDA model. We create a social news scale for news classification, and we also propose an effective approach in preprocessing Chinese news corpus, utilize Gibbs sampling to extract topics from the news corpus. Results show that Tencent, Netease and iFeng differ a lot in news reporting culture, economy, leisure and health, their preference of economy news reporting also reflect as weather wanes for different readers and researchers, and the LDA model is effective in Chinese news corpus analyzing.