基于改进遗传算法优化的BP神经网络上证.rar

摘要:股票市场在国家经济中占据着极其重要的地位，对股票市场的分析与预测具有重要理论意义和实际应用价值。但由于股票市场受经济等多方面因素的影响，导致传统方法效果不甚理想。人工神经网络作为一门非线性科学，由于其自身很强的容错性、自适应性和非线性映射能力，为股票市场的建模和预测提供了新的技术与方法，其中应用最为广泛的是按误差逆传播算法训练的多层前馈网络BP神经网络。但BP神经网络存在学习速度慢、易陷入局部极小值、权值、阈值及网络结构的选择具有很大的随机性、预测结果精度不高等缺点。

　　　本文首先编程实现了基于粒子群算法的改进遗传算法，并通过六种典型函数测试，验证了编程的准确性及改进遗传算法在提高算法稳定性和收敛精度方面的有效性。以上证综合指数为研究对象，在通过仿真实验确定了BP神经网络隐层单元数目的基础上，利用改进遗传算法优化BP神经网络权值与阈值，建立了BP神经网络基本数据预测模型和技术指标预测模型。仿真实验结果表明，BP神经网络应用于上证综合指数预测是可行的，而基于改进遗传算法优化的BP神经网络能有效提高预测精度。

关键词 预测；BP神经网络；粒子群算法；遗传算法；上证综合指数

Abstract:The stock market plays an important role in chinese economy, and the analysis and prediction of the stock market has the important theoretical significance and practical application value. As the stock market is influenced by various factors such as economy, the traditional forecasting methods effect is not very ideal. Artificial neural network as a nonlinear science, because of its strong fault tolerance, adaptability and nonlinear mapping ability, provides a new technology and method for the modeling and forecasting of the stock market. One of the most widely used mehod is the BP neural network, which according to the error back propagation algorithm training of the multilayer feedforward network. But the BP neural network has many shortpionts,such as it’s learning speed is slow, and it is easy to fall into local minimum value, weights, thresholds, and it has great randomness about the choice of network structure, the prediction precision of faults and so on. 

　　　Firstly the programme realized the improved genetic algorithm based on particle swarm optimization (pso) algorithm, and through six kinds of typical function test, verify the accuracy of the programming and improved genetic algorithm to improve the effectiveness of the stability and precision of convergence. With the Shanghai composite index as the research object, through the simulation experiments confirmed the BP neural network hidden layer unit number, on the basis of improved genetic algorithm was used to optimize the BP neural network weights and threshold value, the basic data of BP neural network prediction model and the technical index prediction model are established. Simulation experimental results show that the BP neural network is applied in the Shanghai composite index prediction is feasible, and based on improved genetic algorithm to optimize the BP neural network can effectively improve the prediction precision.

Keywords  Prediction   BP neural network  PSO optimization  Genetic algorithm   Shanghai composite index