摘要
反向传播算法有容易陷入局部最小点、收敛速度慢的问题,为了克服这些缺点,在粒子群优化算法中,引入遗传算法中的克隆算子和变异算子,得到一种改进的粒子群遗传优化算法(PSGO),建立一种PSGO优化BP神经网络模型.利用该模型通过matlab编程仿真对证券市场指数和股票收盘价进行预测研究,试验结果证明了该方法的有效性和可行性.
The back-propagation algorithm suffers from the problem of slow learning speed and a tendency to get stuck in local minima. To overcome these disadvantages, the clone operator and mutation operator of genetic algorithm were put forward in particle swarm gengtic optimization （PSGO） based on the predictability of stock market. By the model matlab programming simulation to predict the stock＇s market indices and closing prices, the effectiveness and the feasibility of the introduced method were demonstrated by experimental results of the prediction mode.
出处
《湖南文理学院学报（自然科学版）》
CAS
2009年第2期15-18,共4页
Journal of Hunan University of Arts and Science(Science and Technology)
基金
福建省教育厅科技项目(JB05181)
