Research On Investment Decision Support Model Of Listed Companies Based On Deep Learning Algorithm

In recent years,while the Internet has changed the way of mass consumption,it has also promoted the rapid expansion of the e-commerce industry.Many companies in the Internet and e-commerce sectors have started to go public in order to obtain more lasting development,and e-commerce has developed due to its development.Uncertainty and greater competition also bear greater risks.Therefore,research on the stocks of listed companies in the Internet and e-commerce fields has important practical significance for both companies and investors.As a unique product of the market economy,listed company stocks are not only an important means for corporate investment and financing,but also an important way for investors to make profits.They also play an important role in macroeconomic control and resource allocation.The article takes the listed company's stock related to internet and e-commerce business as the research object,and studies the theories and methods of stock quality evaluation at home and abroad,and proposes to use deep learning algorithm to establish a stock quality assessment and forecasting model with the purpose of providing investors with decision-making stand by.The research mainly includes the following contents:Firstly,it is background and ideological elaboration.The article analyzes the research status quo of e-commerce,stock evaluation theory and neural network development,introduces the influencing factors of stock quality and commonly used evaluation indicators in detail,and elaborates the neural network theory,several common structural frameworks,and the algorithms used in this paper.Secondly,an indicator evaluation system was created based on the factors affecting the quality of stocks.This paper selected 502 stocks of listed companies involved in Internet and e-commerce business as the experimental sample,and analyzed the information disclosed in the company's financial statements for the past three years and related websites.Initially set 30 evaluation indicators to form a stock quality evaluation system and prepare relevant data.Finally,a neural network stock quality evaluation model was built,including the model structure,parameters and algorithm design;the experimental data was input into training,and the data processing process and prediction results were displayed.Through the repeated training of the model,various parameters are modified to obtain the experimental model.Two methods are used to model the data after processing,and the optimal classification evaluation and prediction model is selected.Using the Python development environment for model training,combined with the evaluation and prediction model,provides a reference for investor decision-making.The stock quality assessment forecasting model established in this paper is compared with the traditional technical index analysis method.It does not need to calculate the original data in strict accordance with the traditional mathematical formula.Instead,it directly sets a series of influencing factors into indicators to form the index system.Facilitate the understanding of ordinary investors.This evaluation and prediction model is mainly used to judge the quality of Internet or e-commerce companies' stocks.Based on this model,the company can analyze the various indicators of the company's data,which can help investors to make decisions.