Investment portfolio recommendation method based on stock inference network

The invention belongs to the technical field of investment recommendation, and particularly relates to an investment portfolio recommendation method based on a stock inference network. The method gets rid of dependence on financial indexes of listed companies and other off-site factors, automatically captures and constructs the incidence relation between the stocks only by means of market transaction data, and carries out investment portfolio recommendation according to the constructed stock network. Therefore, convenience is provided for overall analysis of the market, and meanwhile support is provided for optimizing investment portfolio composition and helping investors to make investment decisions better. Verification is carried out on historical transaction data of Shanghai deep 300 index constituent stocks in the A-stock market, and effectiveness and accuracy of the method are proved. 本发明属于投资推荐技术领域，具体涉及一种基于股票推断网络的投资组合推荐方法。本发明摆脱对上市公司财务指标和其他场外因素的依赖，仅凭借市场交易数据自动捕捉和构建股票之间的关联关系，同时根据构建出的股票网络进行投资组合推荐。这为市场整体分析提供了