Using Machine Learning to Build an Investor Sentiment Index in Vietnam Stock Market

Abstract

Nowadays, more and more investors make decisions in stock trading based on information posted on social networks. Therefore, whether it is official information or false rumours, it also has a spread effect that affects many people, thereby affecting the decision to buy and sell stocks. We use a quantitative method that relies on social network posts to determine investors’ sentiment index in forecasting VN-Index, HNX-Index, VN30, and HNX30. The research results show that the sentiment index of investors is statistically significant and predicts the movements of the VN-Index and VN30 within three days but has no statistical significance for the HNX-Index and HNX30. This index is a tool to support the decisions of investors, listed securities companies, brokers, securities companies, and regulators on the securities market.