Land Cover Mapping Using Ensemble Feature Selection Methods

TL;DR

This paper investigates ensemble feature selection for land cover mapping, finding that current diversity measures do not effectively improve ensemble accuracy.

Contribution

It introduces an exhaustive search-based ensemble feature selection method and evaluates the relationship between diversity measures and classification accuracy.

Findings

Diversity measures as currently formulated are inadequate for ensemble land cover mapping.

Ensemble classification accuracy does not strongly correlate with diversity measures.

Exhaustive search can identify feature subsets for ensemble classifiers.

Abstract

Ensemble classification is an emerging approach to land cover mapping whereby the final classification output is a result of a consensus of classifiers. Intuitively, an ensemble system should consist of base classifiers which are diverse i.e. classifiers whose decision boundaries err differently. In this paper ensemble feature selection is used to impose diversity in ensembles. The features of the constituent base classifiers for each ensemble were created through an exhaustive search algorithm using different separability indices. For each ensemble, the classification accuracy was derived as well as a diversity measure purported to give a measure of the inensemble diversity. The correlation between ensemble classification accuracy and diversity measure was determined to establish the interplay between the two variables. From the findings of this paper, diversity measures as currently formulated do not provide an adequate means upon which to constitute ensembles for land cover mapping.