Efficiently integrating kernel correlation and dissimilarity enhances clustering accuracy in Multiple Kernel k-Means.