Research And Implementation Of Securities Sector Of Correlation System Based On ChebNet Graph Convolutional Neural Network

The Chinese securities sector market originated in 1990 and has been in development for more than 30 years.As investors become more and more concerned about the continued robustness and safety of their investments,the study of the size of correlations between sectors has become more and more in-depth.When two sectors have a strong correlation,there is likely to be a linkage between them that goes up as well as down.Therefore,to explore whether there is correlation between sectors and to study how to use sector correlation to improve investment return,reduce investment risk and increase investment choice and accuracy is a problem worthy of in-depth study.In order to study the correlation between different securities boards in depth,a securities board correlation system is constructed,which mainly contains a data download module to realize the download,cleaning and storage functions of board and stock data;a user management module to realize the login information of different users for management;a board management module to manage the basic information of boards,board correspondence and board log information;a board ranking The module designs basic algorithms such as recommendation ranking,weighted ranking,cumulative ranking and up ranking and develops algorithms such as comprehensive ranking,dynamic ranking and up and down ranking based on the basic algorithms to observe the market trend;the board analysis module develops board correlation analysis,individual stock correlation analysis,board and individual stock correlation analysis,board and index correlation analysis and individual stock and index correlation analysis.The Pearson correlation coefficient and Euclidean algorithm are used to calculate the correlation size between different sectors and build a more effective portfolio according to the correlation size between different sectors.Based on deep learning techniques,an enhanced Cheb Net graph convolutional neural network model is constructed for predicting the plate correlation,and the plate adjacency matrix and the feature matrix are constructed and input to the model for plate correlation row prediction,which can capture the nonlinear relationship between historical plate data in time dimension,and in order to solve the problem that the dynamic change relationship in time series data is difficult to model directly,the matrix product is used The feature vectors of the previous time step and the current time step are fused so that the past information can be combined to the current time step for prediction,which improves the prediction accuracy of the model,and after comparison experiments,the enhanced Cheb Net model performs well in predicting plate correlations.The securities sector correlation system described in this study was developed based on a B/S architecture using the mainstream technology PHP and My SQL database.The system is based on board and stock data and combines board ranking algorithms and correlation analysis algorithms for historical data analysis.In addition,an enhanced Cheb Net graph convolutional neural network model is applied to predict the magnitude of future inter-sector correlations and assist investors in making investment decisions in the sector market.