Abstract: High-speed rail has become a beautiful business card of China. However, some quality problems in high-speed rail catering, such as "mildewed box lunches", are widely reviled. With the launch of one after another improving measures, such as online reservation service and forbidding non-frozen box lunches supply, more attention are paid to the quality of high-speed rail catering. This paper focuses on the Internet public opinions related to the high-speed rail catering service. In order to explore the central demand of the public, data collected from Weibo are analyzed from three perspectives, namely, the trend of Internet public opinion dissemination, changes in public sentiment and the contents of identified topics. Methods applied in the present paper include social network analysis, sentiment calculation and topic clustering. Finally, based on the results, some suggestions on improving the catering service of China's high-speed railway are proposed from the perspective of Internet public opinion.

Key words: Internet public opinion, high-speed rail catering, sentiment calculation, social network, topic clustering