Design And Implementation Of Stock Information Platform Based On Artificial Intelligence Algorithm

With the rapid development of the Chinese economy and the continuous expansion of investment scale,the stock market has produced more and more transaction data and market public opinion information,which makes it more difficult for investors to distinguish effective investment information.With the development of artificial intelligence,more and more artificial intelligence technologies are applied in various fields.The application of artificial intelligence technology has made financial-related practitioners more and more diversified in their analysis of industry data.Relevant practitioners have become increasingly dependent on artificial intelligence technology.However,the design of artificial intelligence algorithms applied in the financial field has high level requirements for relevant financial practitioners.Traditional financial practitioners are poor at emerging algorithms.Algorithms such as time series data algorithms and natural language processing can effectively help financial workers learn more information behind the data from the perspective of computational science.On the other hand,for most natural person investors,they often lack effective market analysis tools,and it is easy to fall into the state of blind investment,which leads to the loss of investor interest.Based on the above reasons,this paper designs and implements a comprehensive stock information platform based on artificial intelligence algorithms.The platform is mainly divided into an online service system and an offline algorithm system.The offline system part mainly includes three parts: data acquisition,stock index data analysis(time series data analysis)and this article information analysis(natural language processing).In the data acquisition part,this article uses a variety of ways to obtain different types of data to provide data set guarantee for the implementation of related analysis algorithm design.In the stock index data analysis part,this article uses three time series prediction methods to predict stock data,which include Facebook Prophet,ARMIA,Holt-Winters.And this part designs a model evaluation index system to inform users of the performance of different models in historical datasets.At the level of text information analysis,this article uses Text Rank,TF-IDF,Bert,Bi LSTM,CRF and other algorithms for natural language processing model design.We designed keyword extraction for news text information,event extraction for company announcement information and analysis for market investor public opinion information.The three functions can help users quickly understand news and event announcement information,and understand market sentiment factors through comment information.We design an evaluation model of sentiment factors and actual trading trends to inform users of the market the correlation between emotion and actual trend.In the online service system part,this article uses Java and Java Script language as the development technology of online system platform to provide the interaction and other function for users.Based on the aforementioned methods,this article starts from five aspects: platform demand analysis and overall design,data source module,stock index module,text data processing module,platform implementation module,and is used to design this platform system.The platform is intended to provide platform users with auxiliary decisionmaking information from multiple dimensions.It helps users judge the current market form and make auxiliary decisions to prevent users from blindly investing.