Sentiment Analysis Based on Financial Tweets and Market Information

Abstract

With the rapid development of Internet technology, people are increasingly accustomed to obtaining information and expressing their sentiments on the Internet. In the financial market, the sentiments in the financial news are supposed to have some correlation with the movement of the stock market. This paper aims to explore such correlation by conducting an empirical and data analysis. It first introduces the procedure for financial news sentiment analysis and the methods such as data mining and text processing. Then with the stock market data analysis and the calculation of the sentiment information based on Textblob, it proposes the manageable way of improving the performance of the sentiment lexicon, especially the way of text mining and choosing the feature words artificially. Finally the paper builds the model for the sentiment analysis. By doing several experiments, this paper deeply analyzes the relationship between the news sentiment values and the stock market price, and proves the effectiveness of the methods in this research. The findings demonstrate that most of the time the fluctuation of the sentiment can match with the stock market price.