Machine Learning Approach to Sentiment Recognition from Periodic Reports

Abstract

We propose a novel indicator to measure fund managers’ sentiment, a topic of significant interest to both academia and the financial industry, as it relates to investor sentiment and stock volatility. As mutual funds continue to gain traction, fund managers have emerged as crucial players in the Chinese stock markets. Drawing upon a dataset comprising 4,142 mutual funds over a five-year period, we construct a fund manager sentiment index by analyzing periodic reports. Additionally, we examine the mediating effect of the investor sentiment index on stock volatility. This study contributes to the understanding of fund managers’ sentiment and its potential implications for stock market fluctuations.