Stock trend prediction method, system and device based on market emotion and hierarchical hypergraph convolutional neural network, and medium

• Main Authors: MIAO QIGUANG, KWON EUI-YOUNG, SUN PENGGANG, QUAN FENGLEI, WANG YUCHEN, LIN GUOQING
• Format: Patent
• Language: Chinese; English
• Published: 04.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

A market emotion and hierarchical hypergraph convolutional neural network-based stock trend prediction method, system, device and medium, the method comprises the steps of firstly performing named entity recognition on financial text data, hierarchically dividing text features into individual stock, industry and market levels, effectively distinguishing the influence degrees of different levels of market emotions on stocks, and predicting the stock trend according to the influence degrees of different levels of market emotions; named entity recognition also goes deep into word and sentence levels to solve the sentiment classification problem of multiple subjects and multiple sentiments; then, a layered hypergraph convolutional neural network is used for dynamically fusing quantized data of the stock and individual stock text data to serve as hypergraph points, emotional fusion of the industry level and the market level is completed through the hypergraph property, modeling of the real stock market is improved