Stocks price prediction based on optimized echo state network by sparrow search algorithm

The stock market reflects the level of economic development of a country to a certain extent. Therefore, stock price forecasting has become a popular research issue. More and more attention has been paid to the establishment of stock price forecasting models. However, the stock market is a nonlinear system filled with high noise, and the traditional fundamental analysis, technical analysis and ARIMA model cannot achieve the ideal prediction result. Although traditional neural networks have good adaptability and nonlinear approximation capability, but they also have drawbacks, such as slow training speed and easiness to fall into local optimization. The echo state network-sparrow search algorithm (SSA-ESN) model employed in this paper uses the sparrow search algorithm to optimize the parameters of echo state network (ESN). It has not only better prediction result but also inheriting the excellent properties of ESN. The SSA-ESN model is used to predict several stocks price, and the results are compared with those of BP neural network, Elman neural network, LSTM and echo state network. And the prediction accuracy of SSA-ESN model is significantly better than those of the other four prediction models.