A revised approach to detect time series correlation

摘要：

This paper proposes the concept of function correlation, which takes the concept of linear correlation and concept of non-linear correlation together with a unified. On this basis, a BP neural network model was constructed to approximate the correlation function. Experimental results show that the correlation function determined by BP neural network can be a higher correlation coefficient.

关键词： time series analysis correlation coefficient neural network

作者: WANG Lei LI Shushan

作者单位: College of Information Science and Engineering Shandong University of Science and Technology Qingdao,China

会议类型: 国际会议

会议名称: 2010 Second International Conference on Intelligent Human-Machine Systems and Cybernetics(第二届智能人机系统与控制论国际学术会议 IHMSC 2010)

会议地点: 南京

会议语种:英文

页码: 321-324

在线出版日期: 2010-08-26（万方平台首次上网日期，不代表论文的发表时间）