An approach for discovering Multilingual news events and term association from the Web

摘要

We have investigated an approach for automatically discovering news events from Web online news downloaded from different sites of different languages. The story content is analyzed. Unsupervised learning is conducted to discover events. From the comparable news stories in the events, statistical analysis of term co-occurrence is developed for mining bilingual term associations. We have conducted some experiments to evaluate our approach on discovering events and term associations. According to the result of the experiment, the approach is a effective way for the discovering Multilingual news events and term association from the web.