Research And Development Of Stock Recommendation System Based On Deep Neural Network

China's reform and opening up and the rapid development of the Internet for many years,the Internet has become a channel for most people to release and obtain information.The financial industry is even more so.More and more institutions release news information and data,and more and more analysts analyze the economy.And with the diversification of investor investment needs,the demand for more diverse information is also increasing.For an investor,a lot of financial information on the Internet will not be able to generate investment assistance for it,but investors need hot stocks and other hot information that are worth investing in now.Therefore,how to effectively obtain current hot stocks and other hot information from a large amount of financial text information is a problem that the financial industry must solve in order to meet user needs.Traditional financial websites are more inclined to display data when recommending hot stocks and hot information.Users often need to spend a lot of energy and time to find hot stocks in dense numbers and texts.The display of hotspot information is basically a method based on statistical counting,which has relatively low referability.Aiming at the above problems,this article uses natural language processing technology to search the current hot stocks,hot sections,hot regions,etc.,and uses the Django framework to build a web system to provide hot stock recommendations and some specific information query functions.This system mainly uses the named entity recognition model to identify the entities in the text,establishes the relationship between the entities,uses the method of graph theory to find the key entities as hot spots,recommends the hot stocks.Use a graph database to efficiently process complex entity relationships,and respond to user queries for related content.The later part of this paper tests the stock recommendation system which mainly includes the core functions of the system.This article uses the current industry's widely used natural language processing technology and graph theory to search and determine hot spots.Use graph databases to process intricate financial industry information and present the information to users.Has credibility in the function implemented.Tests show that this paper uses a natural language processing algorithm to implement a reliable stock recommendation system,which will have a wide range of practical applications in the financial field.