模糊认知图在股票市场预测中的应用研究  

Application research in stock market forecast with fuzzy cognitive maps

作　　者：林春梅[1] 何跃[2] 汤兵勇[1] 刘兴华[1] 

机构地区：[1]东华大学信息学院,上海200051 [2]绍兴文理学院计算机系,浙江绍兴312000

出　　处：《计算机应用》2006年第1期195-197,201,共4页journal of Computer Applications

摘　　要：复杂系统中存在着大量的过程依赖、自组织,并且一直是进化的,用传统的方法对其建模十分困难。模糊认知图作为一种模糊逻辑和神经网络相结合的产物,为复杂系统建模提供了一种有效工具。文中根据模糊认知图的特点,提出了用遗传学习算法建立系统的模糊认知图方法,为复杂系统分析及预测提供了一种解决方案。最后,以股票市场的数据为例进行了分析和预测模拟,结果表明,该方法是有效的。There is a number of process relying, self organization and evolution in the complex system, so the traditional modeling method is difficult. Fuzzy cognitive maps as a soft computing method, which combines the synergistic fuzzy logic and theories of neural networks provides a powerful tool for the complex system modeling. According to the characteristics of fuzzy cognitive maps, a genetic learning algorithm was proposed for the construction of the system fuzzy cognitive maps. It provided a method for the analysis and the forecast of complex system. Finally, the stock market data were used as the test data for the simulated test of analysis and prediction. The result shows that the forecast technology has short-term forecast ability.

关 键 词：模糊认知图 系统建模 遗传算法 预测 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]