Bayesian distance metric learning for discriminative fuzzy c-means clustering.

•A probabilistic model is proposed for discriminative fuzzy clustering which integrates both the fuzzy c-means clustering and distance metric learning in a joint formulation.•This proposed method tries to maximize the separability among different clusters.•This method is able to find the number of dimensions of concern in the projected space in an automatic manner, and in most cases, it has the best clustering accuracy in these dimensions.•This method outperforms its counterparts which perform both clustering and dimensionality reduction simultaneously.