The boom of Big Data has made the development of prediction algorithms more intelligent, so the studies have gradually shifted from the traditional linear prediction algorithm (a typical representative of time-series prediction algorithm) to the popular deep learning prediction algorithm. The nonlinear deep learning algorithm can better reflect the changeable internal laws and external relations of data, especially for complex stock price data. Long Short Term Memory network (LSTM) is a special algorithm for processing time-series problem. In this work, we conducted a stationary analysis of the stock\'s time-series data and then used the LSTM neural network algorithm to predict stock data under different stationary conditions, and performed statistical analysis on multiple experimental data. In addition, an ARIMA algorithm was introduced to compare with the LSTM. A large number of experimental results show that the LSTM neural network prediction algorithm has higher prediction accuracy and is not sensitive to the stability response.
Neural networks
• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask,hadoop Frontend :-python Backend:- MYSQL
₹10000 (INR)
2019