Classification of dissimilarity data with a new flexible Mahalanobis-like metric
Résumé
Statistical pattern recognition traditionally relies on features-based representation. For many applications, such vector representation is not available and we only possess proximity data (distance, dissimilarity, similarity, ranks ...). In this paper, we consider a particular point of view on discriminant analysis from dissimilarity data. Our approach is inspired by the Gaussian classifier and we defined decision rules to mimic the behavior of a linear or a quadratic classifier. The number of parameters is limited (two per class). Numerical experiments on artificial and real data show interesting behavior compared to Support Vector Machines and to kNN classifier (i) lower or equivalent error rate, (ii) equivalent CPU time, (iii) more robustness with sparse dissimilarity data.