Extraction and Association Co-Occurrence Analysis of Hotwords in Tibetan Web

Hotwords mining is an important part of network public opinion analysis research, and has been applied successfully to Chinese and English texts, but the research on minority languages is still in its infancy. This paper proposes a hotwords extraction algorithm for Tibetan web texts and mines the association among hotwords in order to provide a reference for the subsequent research on Tibetan opinion analysis. Firstly, a large number of Tibetan articles were collected from the Internet to construct the corpus. Then the corpus was divided into four parts according to the release time. Based on the word frequency, we introduced the position and the length weight to calculate the words weight, and extract the hotwords of the four time periods. Finally, the frequent itemset extraction method in association rule mining was used to obtain the frequent co-occurrence word sets from Tibetan text and it is analyzed to track relevant hot events. The analyses show that the hotwords extracted by the algorithm in this paper can effectively reflect the propaganda focus of the websites and the focus of the Tibetan people in different time periods.