[R] Histogram equalization of dataset

I read about histogram equalization of dataset. It is commonly used for image data. It is written: Each variable in dataset is assigned to one bin. Then, each variable in the dataset is assigned to one bin, incrementing the value of that bin by one. Then, a cumulative histogram is created by adding

[R] Feature selection for kmeans

I know that within sum of squares, DB, sillhouette and cophenetic are indicators of clustering quality, but what indicators I need to observe when I choose attributes for kmeans? -- View this message in context: http://r.789695.n4.nabble.com/Feature-selection-for-kmeans-tp4657830.html Sent fro

[R] What methods you use for choosing attributes for kmeans and other clustering (unsupervised) methods?

What characteristics makes attribute great for kmeans clustering and how do you measure these characteristics? Do you look at the distribution shape, correlation matrix between attributes, pairs graphs, statistical tests...? -- View this message in context: http://r.789695.n4.nabble.com/What-m