Stock market data analysis method of recurrent neural network based on dimension reduction technology optimization

Abstract

The invention belongs to the technical field of financial information data processing, and discloses a stock market data analysis method of a recurrent neural network based on dimension reduction technology optimization, which comprises the following steps: respectively carrying out factor analysis on a data set, and respectively taking out a first class and a second class of quality factors of the data set; finding out a parameter, and analyzing the relationship of the parameter to the rise and fall of the stock along with the time change; and substituting the parameters into an LSTM model for prediction, adding a forgetting gate to each excitation source of the optimized LSTM, and screening previous information. According to the invention, the LSTM neural network is expanded in the financial field, the concept of forgetting the door is successfully applied to stock market analysis, and the accuracy is improved; a dimension reduction algorithm is introduced into data and processing and is compared. According to the method, the advantages of the dimension reduction technology and the accuracy of the LSTM network are highlighted, so that stock market prediction is more credible thana traditional analysis method; prediction may be applied to reality.

Description

technical field [0001] The invention belongs to the technical field of financial information data processing, and in particular relates to a stock market data analysis method based on a recurrent neural network optimized by dimensionality reduction technology. Background technique [0002] At present, the closest existing technology: principal component analysis, in the research and application of many fields, usually needs to observe the data containing multiple variables, collect a large amount of data and then analyze and find the law. Multivariate large data sets will undoubtedly provide rich information for research and application, but also increase the workload of data collection to a certain extent. More importantly, in many cases, there may be correlations between many variables, which increases the complexity of problem analysis. If each indicator is analyzed separately, the analysis is often isolated and the information in the data cannot be fully utilized, so bl...

AI Technical Summary

Problems solved by technology

[0012] (1) The existing principal component analysis method and the traditional cyclic neural network only use a small number of two or three of them to make predictions, while ignoring the rest of the indicators, resulting in inaccurate data results

[0013] (2) If the traditional RNN algorithm is optimized using the gradient descent method, there will be serious problems of "gradient disappearance" or "gradient explosion", and the accuracy of the data is low

Images