In this paper the EOF methodology is performed jointly with the FDAapproach on a spatiotemporal multivariate data set with the aim to fill in missingvalues as accurately as possible when long gap sequences occur. Simulated datasets, containing ”artificial” gaps, are considered in order to test the performance oftwo proposed procedures; in the first one, observed data are reconstructed by EOFand then converted into functional ones; in the second one, observed data are transformedinto functional ones and then EOF reconstruction is applied. By comparingsome performance indicators computed for the two procedures, it is shown that apre-processing of data by FDA, followed by the EOF, may result in a better reconstruction.
|Numero di pagine||4|
|Stato di pubblicazione||Published - 2012|