Abstract: Social media provides an opener platform for individual investors to express opinions and gather information In order to study the relationship between investor sentiment and IPO underpricing deeply, web crawler (Python) technology is used to mine social media big data, and a stock market emotional lexicon is manually constructed to analyze investor sentiment and IPO underpricing included in social media speeches The results show that the positive sentiment of investors of the social media big data has a significantly positive correlation with IPO underpricing, the negative sentiment of investors has a significantly negative correlation with IPO underpricing, and negative sentiment has a more significant effect on IPO underpricing rate Further research shows that social media big data can be used as a reliable index to measure investor sentiment, which provides new methods for measuring investor sentiment and lays the theoretical foundation for government′s regulation on social media information

Key words: investor sentiment, IPO underpricing, data mining