Research on Investor Sentiment and Trading Behavior in the New Energy Sector Based on the PCA-HMM Model

Renewable energy stocks are affected by investor sentiment and trading behavior, particularly during changes in policy or market trends. This study applies a Principal Component Analysis–Hidden Markov Model (PCA–HMM) to examine how sentiment, trading volume, and momentum relate to stock returns in the U.S. renewable energy market. The data include 36 listed companies from 2016 to 2025 across the solar, energy storage, and electric vehicle sectors. Principal Component Analysis was used to extract main factors from market and macroeconomic variables, and the Hidden Markov Model was used to identify hidden market states. The results show that stock returns react more to sentiment than to trading volume or short-term momentum. In high-sentiment periods, both volatility and return fluctuation increase, while low-sentiment periods show partial momentum reversal. Compared with a single-state model, the PCA–HMM achieves a better fit and describes time-varying relationships more accurately. These findings suggest that a state-based approach to sentiment can help monitor market risk and support investment analysis in renewable energy finance. Further research should include higher-frequency data and cross-market sentiment links to enhance model reliability.