Study of polynomial curve fitting algorithm for outlier elimination

An algorithm based on polynomial curve fitting algorithm was proposed to process outlier error of flight testing data. At first, the variance of data´s noise was obtained. Then based on pauta criterion static threshold, the paper raised a method using dynamic threshold to determine and remove outlier. At last these outlier datas were replaced according to statistical properties. After repeated and recycling processing using above algorithm, all outlier could be removed completely. When all outlier errors of various flight testing data were eliminated, available testing data without outlier can be obtained.