Stock Price Forecasting in Real Estate Industry Based on Investor Sentiment

Abstract

Stock price fluctuations are unstable, and investors' decision-making is also subject to the influence of their feelings. This study suggests a real estate stock price prediction approach based on investors' moods in response to the present phenomena that investors' desire to participate in the real estate development business is falling and risk aversion is increasing. Firstly, we quantify the stock bar comment scores combined with the Baidu search index to construct a composite sentiment index. Next, we combine additional stock price influencing factors and perform functional principal component analysis and dimensionality reduction to address the issue of multiple covariances. Finally, we input the index into CNN's prediction model to determine when the stock price will rise or fall.