ON SMOOTHING TIME SERIES DATA USING A CLASSICAL MOVING AVERAGE FORMULA
MetadataShow full item record
In time series realizations, assuming that the trend component to be approximated by a polynomial in time, smoothing filters based on a moving-average formula are proposed which link the degree of this polynomial to the number of terms operating on the moving-average formula. Their properties are examined. Illustrations via data collected from three East Africa Regional meteorological stations are reported.