投资者情绪的识别与统计——基于非结构数据的分析

投资者的情绪特征主要来自于投资者的心理认知,基于行为金融学的前景理论、期望理论与后悔理论,深入细致地分析了影响投资者情绪的内涵,进而编制投资者情绪特征指数进行测度,以此研究投资者情绪特征指数在行为金融学中的影响机制。同时应用爬虫技术爬取相关投资者情绪的数据,经过合理清洗、筛选与朴素贝叶斯处理等过程,对投资者情绪特征进行统计实践。研究发现投资者情绪特征具有可甄别性、可预测性;基于对投资者情绪特征变量的统计测度方法的分析,论证这些统计测度方法不仅可以实现对投资者风险偏好特征的提取与测度,而且可行;积极情绪投资者对投资风险性较高的新兴行业板块更加青睐,且与消极情绪投资者呈现出显著的分布差异,但是积极情绪投资者与消极情绪投资者在投资收益率方面并无显著差异。

Identification and Statistics of Investor Emotions: Analysis Based on Unstructured Data

The emotional characteristics of investors mainly come from their psychological cognition. Based on the prospect theory, expectation theory, and regret theory of behavioral finance, this paper deeply and meticulously analyzes the connotations that affect investor emotions, and then prepares an investor emotional characteristic index for measurement, in order to study the impact mechanism of investor emotional characteristic index in behavioral finance. At the same time, crawler technology is applied to crawl relevant investor sentiment data, and through reasonable cleaning, screening, and naive Bayesian processing, statistical practice is carried out on investor sentiment characteristics. Research has found that the emotional characteristics of investors are distinguishable and predictable; Based on the analysis of statistical measurement methods for investor sentiment characteristic variables, it is demonstrated that these statistical measurement methods not only can extract and measure investor risk preference features, but also are feasible; Positive sentiment investors prefer emerging industry sectors with higher investment risks, and exhibit significant distribution differences compared to negative sentiment investors. However, there is no significant difference in investment returns between positive sentiment investors and negative sentiment investors.