Stock Price Predicting Using SVM Optimized by Particle Swarm Optimization Based on Uncertain Knowledge

Abstract

Stock prices have the characteristics of nonlinearity, randomicity and uncertainty, so It is difficult to accurately depict the change rules of stock prices using traditional linear forecasting methods, which lead to low stock price prediction accuracy. In order to improve the stock price prediction precision , this paper proposed a stock price predicting model using SVM optimized by particle swarm optimization based on uncertain knowledge(PSO-UK). We used the great optimization ability of PSO-UK to optimize the parameters of SVM, enhanced the learning ability of SVM , and used the SVM to predict stock price. We compared the accurancies of PSO-UK-SVM and PSO-SVM using SSE Composite Index. Experimental results showed that PSO-UK-SVM model performed better degree of fitting and accuracy. The model proposed in this paper has some guiding significance for investors .