Analysis of the impact of investor sentiment on stock price using the latent dirichlet allocation topic model
1 Introduction
In finance research, studies on investor sentiment are carried out by selecting proxy indicators such as volatility index (VIX), initial public offering (IPO) number, etc., but fail to make use of a large number of text data generated in the interaction of investors in the stock bulletin board system (BBS), which reflects investor sentiment and affect the final investment decision to some extent (Zhang and Zhang, 2021, pp.143-150). However, there is a large amount of noise in text data and a massive workload of labelling investor sentiment.
In recent years, significant achievements have been made in text mining, machine learning and sentiment analysis, and considerable progress has been achieved in natural language…
