摘要
该文用非线性时间序列分析方法，对一段股市行情序列进行了拟合，指出可用逐段线性回归拟合趋势，用门限自回归模型拟合消除趋势后的平稳序列，通过对１９９７年４月２２日至５月１２日期间深圳股市行情预测值与实际值的对比，说明在正常状态（即无违规操作及无特殊政策出台）下，所建立的模型有较好的拟合效果，从而提供了一个行情预测的有效方法。
In this paper, a method of nonlinear time series analysis is used to fit the series of a part of current prices of stock. It is pointed out that a piecewise linear regression method can be used to fit the trend of series where applying thresheld autoregressive model to fit the stationary error series in which the trend of original time series has been eliminated. It is also proved that under normal conditions, our models have nice fitting results, through comparing Shengzhen stotk index with its prediction values. So an effective method for current prediction is provided.
出处
《南京理工大学学报》
EI
CAS
CSCD
1998年第1期82-85,共4页
Journal of Nanjing University of Science and Technology
基金
南京理工大学科研发展基金
关键词
预测
阈限
股市行情
非线性
时间序列分析
forecasting, threshold limit, autoregressive models
current prices of stock
