Forecasting Behavior of Economic Multivariate Time Series with Neural Networks

摘要：

In this paper,the neural network approach to nonlinear multivariate time series forecasting is investigated.(1)The BDS test is employed to test nonlinear dependence within time series.(2)A statistical method based on fractal dimension is designed for testing nonlinear dependence between time series.(3)A learning algorithm for neural networks based on modified genetic algorithm is proposed.The methods of this paper have been used in forecasting the Shanghai Composite Index of Chinese Restate and work very well.

关键词： Neural network Modified genetic algorithms BDS test Fractal dimension Nonlinear Time series Economic forecasting Statistical inference Real estate price index

作者: Chongjun Fan

作者单位: School of Management,University of Shanghai for Science and Technology,Shanghai 200093,China

会议类型: 国际会议

会议名称: 2008 International Conference on System Management(2008年系统管理学术研讨会)(2008 CSM)

会议地点: 上海

会议语种:英文

页码: 37-41

在线出版日期: 2008-05-30（万方平台首次上网日期，不代表论文的发表时间）