摘要
应用相空间重构技术对时间序列进行分割,将原序列映射到多维的数据空间中。将期望最大化(EM)聚类算法和神经网络相结合,提出了一种基于相空间重构技术的EM聚类模糊神经网络预测模型。在股票市场上进行了应用,结果表明该预测模型降低了预测误差,提高了系统的性能。
Applying phase space reconstruction method to divide time series into segments, we have mapped original series into multidimensional data space. We present a new forecasting model of fuzzy neural network combined with Expectation Maximization method. And use it to make forecasts on stock market. The results show that this model could reduce the error of forecasts effectively and improve the system＇s performance.
出处
《中国软科学》
CSSCI
北大核心
2006年第8期147-153,共7页
China Soft Science
基金
国家社会科学基金(03BJY099)
教育部博士点专项科研基金(20020532005)
教育部高校青年教师教学科研奖励基金项目
关键词
EM聚类
模糊神经网络
相空间重构
时间序列
EM - Cluster
fuzzy neural network
phase space reconstruction
time series
