基于布谷鸟算法优化 BP 神经网络模型的股价预测

摘要

针对当前智能算法对股票市场预测精度不高的问题，提出使用布谷鸟算法优化神经网络（CS-BP）的方法，对股票市场进行预测。并与粒子群算法优化神经网络模型（PSO-BP）和遗传算法优化神经网络模型（GA-BP）的测试结果进行比较。通过对SZ300091（金通灵）日线的收盘价数据回测分析看出，布谷鸟算法优化神经网络模型明显优于这两种算法，能有效对股票市场进行预测，对于30天的预测精度约为98．633％。%This paper puts forward the method of predicting the stock market by using the cuckoo search algorithm to optimise BP-neural network(CS-BP)aimed at the problem of current intelligent algorithms in poor prediction accuracy on the market.Besides,it compares its test result with the results of PSO-BP model (optimising BP-neural network with particle swarm optimisation)and GA-BP model (optimising BP-neural network with genetic algorithm).After analysing the data backtesting result of the closing price of daily candlesticks of SZ300091 (JTL),we can conclude that the CS-BP model is obviously superior to these two algorithms,it can effectively predict the stock market with about 98.633% of accuracy for thirty days prediction.