The Application Of Temporal Nerual Network Model In Stock Classification Forecasting

Twenty-first Century is an era of rapid development of AI technology.The rapid development and application of big data technology and machine learning have penetrated into many practical fields of the society.Stock market is called the barometer of macro-economy,and forecasting stock trend has always been an important practical work.Stock price is a kind of non-stationary,volatile and irregular time series data,and stock price prediction is affected by various complex factors,which leads to the prediction of stock price trend has been a very difficult research topic.Temporal data is the data which is transformed from irregular time series data according to a certain temporal pattern.The transformed temporal data may have some regularity in the temporal pattern.Then the temporal data is mined to discover the temporal knowledge which can not be found in the source data mining.Temporal data mining model provides a way to solve complex and irregular time series data mining problems.Another method of mining irregular data knowledge is neural network model.Neural network has the advantages of parallel,fault tolerance,hardware implementation and self-learning.It can also be used as an effective method for stock classification and prediction.By combining temporal data mining model with neural network model,this paper proposes a classification model and algorithm of temporal neural network,which converts stock data into temporal data,and uses the model to forecast the trend of temporal stock data.By analyzing the data collected from several listed companies in the past ten years,a neural network classifier based on temporal data is constructed,and the trend of eight stocks is classified and predicted.The experimental results show that compared with logical regression classification,naive Bayesian classification,random forest classification,support vector machine method and general neural network method,neural network classifier based on temporal data has higher classification accuracy.After the discretization of temporal stock data,theaccuracy of classification is greatly improved.The results show that the classifier based on temporal neural network is very effective for classification prediction of multi-branch stocks and can be well applied to classification prediction of stock market.