Nonlinear Time Series Forecast Using Radial Basis Function Neural Networks

在线阅读 下载PDF 导出详情
摘要 IntheresearchofusingRadialBasisFunctionNeuralNetwork(RBFNN)forecastingnonlineartimeseries,weinvestigatehowthedifferentclusteringsaffecttheprocessoflearningandforecasting.Wefindthatk-meansclusteringisverysuitable.Inordertoincreasetheprecisionweintroduceanonlinearfeedbacktermtoescapefromthelocalminimaofenergy,thenweusethemodeltoforecastthenonlineartimeserieswhichareproducedbyMackey-Glassequationandstocks.Byselectingthek-meansclusteringandthesuitablefeedbackterm,muchbetterforecastingresultsareobtained.
机构地区 不详
出版日期 2003年02月12日(中国期刊网平台首次上网日期,不代表论文的发表时间)
  • 相关文献