基于遗传算法优化投影寻踪技术的神经网络集成模型及其应用

NEURAL NETWORKS ENSEMBLE MODEL WITH PROJECTION PURSUIT TECHNOLOGY BASED ON GENETIC ALGORITHMS OPTIMIZATION AND ITS APPLICATION

首先利用遗传算法优化的投影寻踪技术对神经网络学习矩阵降维,再利用Bagging技术和不同的神经网络学习算法生成集成个体,并再次用遗传算法进化的投影寻踪技术对神经网络个体集成。建立基于遗传算法优化的投影寻踪技术神经网络集成模型,通过上证指数开盘价、收盘价进行实例分析,计算结果表明该方法具有较好的学习能力和泛化能力,在股市预测中预测精度高、稳定性好。

First of all,we use the projection pursuit technology optimized by genetic algorithm to make the dimension reduction for learning matrix of neural networks ,and then use Bagging techniques and different neural network learning algorithms to generate the individuals of the ensemble. And projection pursuit technology evolved by genetic algorithm is used again to assemble these individual neural networks, the neu- ral network ensemble model with projection pursuit technology based on genetic algorithm optimization is established. Through the analysis of the opening price and closing price instances of the Shanghai Stock Exchange index, the computation results showed that the ensemble network has reinforcement learning capacity and generalization ability, and is of precise and fairly stable in its utilization for the stock market prediction.