I have a sample signal, and I computed MFCC using this code.Now I have an array of 1225x12 size, where there are 12 coefficients for 1225 frames. Now I want to test another signal against this sample, and use vector quantization for this.I have this code, but cannot quite understand what to do. It takes data vectors as input,but I have a matrics. Do I convert it into a vector in some way? And how do I determine if two arrays match?
From your reference : Inputs: D contains data vectors (one per row) ... Outputs J indicates which centre each data vector belongs to
You need to have 12 coefficients in each row of your input matrix. In J you will have a number from 1 to K indicating the cluster number. The number of rows in J will be as in D. You can use cell2mat and reshape if you need. I just not sure that clustering in such high dimensions make sense. It may worth making PCA first.