基于投资者情绪和规划识别的融资行为预测

Financing Behavior Prediction Based on Investor Sentiment and Planning Recognition

对于投资者投资行为的规划中,会受到投资者自身的情绪的影响。规划识别能从观察到的某一智能体的动作出发,结合智能体心理,推导出该智能体的规划。论文将该方法运用到股市中去,通过分析数据并且结合投资者的情绪,对投资者做出合理的融资行为。首先通过PLS构建出投资者情绪。接着构建三层规划识别层,在第一层中引入马尔科夫链去预测智能体的行为概率,在第二层中构建规划识别图,进一步预测行为概率,在第三层中规划出智能体的最佳选择。最后进行实例验证和得出结果。

The planning of investors'investment behavior will be affected by investors'own emotions.Planning recognition can deduce the planning of an agent based on the observed action of the agent and its psychology.In this paper,the method is applied to the stock market,through the analysis of data and combined with the mood of investors,investors make a reasonable financing behavior.Firstly,investor sentiment is constructed through PLS.Then three layers of planning recognition layer are constructed.Markov chain is introduced in the first layer to predict the agent's behavior probability,and the planning recognition map is constructed in the second layer to further predict the behavior probability,and the optimal selection of agents is planned in the third layer.Finally,the example is verified and the result is obtained.