...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>International Journal of Innovative Computing Information and Control >THE DATA-DRIVEN FUZZY COGNITIVE MAP MODEL AND ITS APPLICATION TO PREDICTION OF TIME SERIES
【24h】

THE DATA-DRIVEN FUZZY COGNITIVE MAP MODEL AND ITS APPLICATION TO PREDICTION OF TIME SERIES

机译：数据驱动的模糊认知图模型及其在时间序列预测中的应用

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

As a soft computing method, fuzzy cognitive map (FCM) has become a hotresearch topic in recent years. However, the traditional method of constructing FCMdepends on expert knowledge which has obvious drawbacks of subjectivity and limitationespecially for long-term or complex time series. In this paper, a new method for constructingFCM is proposed which extracts knowledge from data by exploiting clustering,and FCAd’s weights can be automatically obtained by particle swarm optimization algorithm(PSO) according to historical data of time series. Further the proposed method isapplied to predicting, in which FCM is used to represent, storing fuzzy logical relationshipsof time series and realizing prediction by iterations. Three benchmark time series:the enrollments time series, the TAIEX time series and the W olf’s sunspot time seriesare applied to verifying prediction performance of the proposed method, whose resultsshow that the proposed numerical prediction method of time series is effective and canobtain better prediction accuracy than traditional methods, and the potential advantageof the proposed method is capable of processing the prediction problem of long-term orcomplex time series. In addition the influence of parameters of the method is analyzedindividually.
机译：作为一种软计算方法，模糊认知图（FCM）已成为近年来的研究热点。然而，传统的FCM构建方法依赖于专家知识，这具有明显的主观性和局限性，特别是对于长期或复杂的时间序列。本文提出了一种构造FCM的新方法，该方法通过利用聚类从数据中提取知识，并且可以使用粒子群优化算法（PSO）根据时间序列的历史数据自动获得FCAd的权重。将该方法进一步应用于预测，以FCM表示，存储时间序列的模糊逻辑关系，并通过迭代实现预测。将三个基准时间序列：入学时间序列，TAIEX时间序列和沃尔夫太阳黑子时间序列用于验证该方法的预测性能，其结果表明，所提出的时间序列数值预测方法是有效的，并且可以获得比预测更好的预测精度。传统方法，该方法的潜在优势是能够处理长期或复杂时间序列的预测问题。另外，分别分析了该方法参数的影响。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号