Overcoming key weaknesses of Distance-based Neighbourhood Methods using a Data Dependent Dissimilarity

This paper introduces the first generic version of data dependent dissimilarity and shows that it provides a better closest match than distance measures for three existing algorithms in clustering, ano