
Preprocessing Data
The smoothed d ata are shifted from the previous plot. convn with the 'same'
parameter returns the central part of the convolution, the same length as the
data.
filter returns the initial part of the convolution, the same length as
the data. Otherwise, the a lg orithms are identical.
Smoothing estimates the center of the distribution of response values at each
value of the predictor. It invalida tes a basic assumption of many fitting
algorithms, namely, that the errors at each value of the predictor a re normally
distributed. Accordingly, smoothed data should not be used to fit a model. Use
smoothed data to identify a model.
See “Filtering Data” in the M ATLAB Data Analysis documentation for more
information on smoothing and filtering.
5-9
Commentaires sur ces manuels