I have data when clustered using Euclidian distance alone give bad accuracy in certain situation. However when clustered based on other dissimilarity measure provides bad accuracy in different situations. I wanna to increase the clustering accuracy by using two dissimilarity function simultaneously. So that data points that both dissimilarity function have the same decision about their clusters have higher confidence measure? Moreover , How to resolve the problem when the dissimilarity function about the cluster of specific data point?

Similar questions and discussions