Sentiment analysis of stock market behavior from Twitter using the R Tool

Abstract

Sentiment analysis is a hot research topic with widespread applications in organizations. Recently, it has been applied to create investor sentiment indicators that can be utilized in the prediction of stock market variables such as returns, volatility and trading volume. This chapter presents a use case about sentiment analysis in tweets related to stock mar- ket using the open-source R tool. We provide and explain code for the execution of diverse tasks, such as tweets collection, preprocessing tasks, utilization of opinion lexicons, application of machine learning (ML) methods, and respective evaluation. A data set composed by stock market tweets is also provided. This use case shows that the utilization of ML techniques using manually classified training data outperforms some standard sentiment word lists when applied to test data.