The Empirical Study of Investor Sentiment on Stock Return Prediction

Abstract

In Taiwan stock market, most participants are individual investors. Thus, the objective of this empirical study is to explore whether the investor sentiment and investor behavior have considerably influence on the stock return. The study tries to search for predictable indicators and measure them based on two approaches: one is the investor behavior indicator measured by using proxy variables (such as short-term rate of return, long-term average rate of return, turnover rate, and earning-to-price ratio) and the other is the investor sentiment measured by using proxy variables (investor sentiment index, the consumer confidence index, and the market volatility index). In addition, this study creates a stock prediction using the neural networks technique and examines whether the predicted returns reflect the actual returns. Finally, this study expects that the empirical results not only providethe important academic value in financial field, but also provide efficiently an investment strategy for investors and financial institutions.Keywords: Artificial Neural Networks, Investor Sentiment, Behavioral Finance, Stock Return Prediction.JEL Classifications: F37, F39DOI: https://doi.org/10.32479/ijefi.7528