I am doing K means on multidimensional data, how should my plot of Minimum Distortion vs K look like?
1 ビュー (過去 30 日間)
I'm using 5 x 10000 or 5 x N to represent 5D data. Each 5 x 1 sub-matrice represents one five dimensional point, so 5 x N would represent N 5D points.
I did it and now want to check the result. How should a plot of the minimum distortion (Error which converges for that value of K) vs various values of K look like for 5D data?
I did the same for 1D data and the results look something like this.
PS: I'm not using the inbuilt kmeans() to do this, but rather i'm coding it up manually hence I need to check if my logic is correct.
If this is how it looks like for 1D data, how should it look like for 5D data?