Research On The Stock Price Rise And Fall Forecast Based On LSTM Neural Network Model

Deep neural network algorithm is a hot topic in the field of machine learning in recent years.The earliest artificial neural network originated in the 1940 s and focused on engineering applications.Until the emergence of the concept of deep network,deep neural network has become a new darling,and there are many applications in the fields of picture recognition and speech recognition.At present,the processing of time series data is preferred as a cyclic neural network(RNN).When application data involves certain sequential machine learning tasks,RNN can achieve high precision due to its limited short-term memory advantages.In principle,in 1997,after introducing the LSTM-based architecture into the first-generation RNN network,the LSTM neural network model greatly improved the network accuracy.This paper attempts to apply it to financial market forecasting to study a more effective stock market forecasting model.This paper designs and models the stocks based on the LSTM neural network model.In the aspect of model design,different combinations are considered to improve the accuracy.Moreover,different reference samples are provided for the training examples of the model,and the empirical research is done with the industry as the dividing line,and the prediction effect of the model training is compared.According to the input characteristics,this paper extracts individual stock market indicators,market market indicators,financial valuation indicators,and compares them as input variables of the network,and also introduces the comparison of the bull market cycle.In terms of model network structure,this paper uses a layer of implicit layer by default.In the process,the optimal model structure is selected by continuously adjusting the number of hidden layer nodes.In terms of parameter setting,the learning rate and the number of iterations are modified,and the optimal adaptation parameters are selected.At the same time,the LSTM neural network model established in this paper is optimized based on the Adam algorithm model.Finally,we use the model to perform quantitative backtesting.The empirical results show that the model has achieved certain empirical results in stock market forecasting.This also confirms the validity of the model.