Identifying and mining of outliers data based on statistical attribute's wavelet analysis
Jianzhou Wang · Journal of Northwest Normal University · 2004
In the case of time series data,a new technique of identifying and mining outlier data is proposed based on statistical attribute's wavelet analysis.Firstly the experience attribute function is obtained by setting queue data, and estimate the difference between experience attribute function and population attribute function,then use bootstrap method to reduce the dispersion between sample attribute function and population attribute function. At last using sample attribute function to replace population attribute function under some allow able dispersion,the identifying and mining of outliner data are done by wavelet analysis.