Research And Application Of Recurrent Neural Network In Stock Index Forecasting Model

In today's learning life,artificial intelligence,machine learning,and deep learning are in full swing.Artificial intelligence has become increasingly mature in the field of automatic driving.Machine learning and deep learning are in the fields of speech recognition,face recognition,and natural language processing.The application was very successful.The stock market is a barometer of the national economy,and the stock index is an intuitive reflection of the value of the stock.The price of the stock index is easily affected by various economic factors,with complex nonlinearity,instability and other factors.The ability of recurrent neural networks on time series data is better than that of time series models in econometrics and statistics.The ability of traditional neural networks to predict stock prices has been fully exploited.Therefore,it is theoretically and practically worthwhile to propose a forecasting price of stock index based on a recurrent neural network.The paper data is selected from the daily history data of Hong Kong Hang Seng Index from March 27,2000 to March 20,2018,and the data comes from Yahoo Finance Channel.Three data attributes,namely the opening price,the lowest price,and the highest price,were taken as the input variables of the recurrent neural network model.After the original data was collected,4390 sets of data remained after the missing records were removed.The original data is divided into two parts.The first 70% of the original data is used as a training data set,and the last 30% is used as a test data set.The training data set contains 3083 sets of data,and the test data set contains 1037 sets of data.Three kinds of neural network models were used: cyclic neural network model,long-term and short-term memory neural network model,and Clockwork-RNN model to predict the historical data of Hong Kong Hang Seng Index,and the results were compared and evaluated.The forecast results show that: Clockwork-RNN has the best prediction effect,and long-term and short-term memory neural network model has the best prediction effect.The average absolute error and the root-mean-square error between the prediction results of the three stock index price models and the actual prices are very small,and the two errors of the long-term and short-term memory neural network prediction models and the CW-RNN prediction model are smaller.