Stock Index Prediction Based on CNN-LSTM Probability Prediction Model with Attention Mechanism
Given the high volatility of the securities market and the high difficulty of predicting it,this paper integrates the Attention Mechanism into the CNN-LSTM model based on the encoder-decoder structure.The Attention Mechanism is used to capture data dependency patterns between different time points,long series information is extracted,and based on this,a probability density function is provided for sampling prediction,point prediction and interval prediction of stock prices are obtained ultimately.The experimental results show that the CNN-LSTM probability prediction model incorporating the Attention Mechanism outperforms other benchmark models in terms of comprehensive performance,and can make high-precision multi-step predictions of the closing price of the Shanghai Composite Index.
Attention Mechanismprobability density functionShanghai Composite Index
国家科技期刊平台
NSTL
万方数据
