摘要
关于股票价格准确预测问题,借助股票价格指数,投资者可以掌握股市整体的发展动态。为了增加收益,降低风险,制定正确的投资决策,合理预测股指是必要的。然而传统预测方法存在方法单一、缺乏定性分析等不足,难以适应国内复杂的股票市场。为解决上述问题,在股市可以预测的前提下,从股市自身特点出发,提出了一种定性与定量相结合的新式组合算法。将粒子群优化(PSO)、非线性独立成分分析(NLICA)、BP神经网络三种算法相结合,建立上证综指预测模型,并通过计算机仿真进行模型验证。结果表明新式组合预测模型比传统方法的适应性和智能性更强,预测精度更高,在股市短期预测中具有一定实用价值。
Stock investment is an important investment style for modern people. With the help of stock price index, investors can know the developments of the whole stock market. In order to increase benefits, reduce risks and make correct investment decision, it is necessary to predict stock price index reasonably. However, traditional fore-cast methods are simplex and lack qualitative analysis. Base on this, under the precondition that stock market can be forecast, in consideration of the boundedness of traditional forecast methods and stock market characteristics, a new combinational algorithm which is combined with qualitative and quantitative analysis was proposed. It was combined with Particle Swarm Optimization (PSO), Nonlinear Independent Component Analysis (NLICA) and BP Neural Net- work, A forecast model of Shanghai composite index was established, and the model was checked by computer simulation. The results show that this new combinational forecast model has stronger adaptability, more intelligence and higher forecast accuracy than traditional methods. It has a certain practical value in short - term forecast of stock market.
出处
《计算机仿真》
CSCD
北大核心
2013年第12期203-207,共5页
Computer Simulation
关键词
股市预测
上证综合指数
粒子群优化
非线性独立成分分析
Stock market forecast
Shanghai composite index
Particle swarm optimization ( PSO )
Nonlinear inde- pendent component analysis