PREDICTIVE MODELS OF SEASONAL TIME SERIES FOR HEMORRHAGICFEVER WITH RENAL SYNDROME BASED ON WAVELET ANALYSIS

摘要

The application of wavelet analysis in statistical field is still in its first stage and is the focus of international interest. In our study, the incidences of hemorrhagic fever with renal syndrome from two surveillance spots, Heihe with seasonal horizontal time series and Linyi with seasonal tendency time series, from 1995 to 1998 were used to predict the 12 monthly incidences in 1999 by wavelet analysis. The relative errors between actual value and predictive value were -0.030111 and -0.008307 respectively.