Good day Guys..

I have a data set represents spoken digits, where the features of each digit are extracted by means of MFCC (mel frequenct cepstral co-effiecints).
Each digit is having 13 features extracted in raw formt, but there are multiple of raws per digit (even for the same digit, size of block varies).

I need this dataset to be input to a network analyzer that accepts liner inputs only.

Is there a way to normalize the variable length and size of these blocks of dataset, while maintaing the same features for each digit? Is there any MATLAB function that coud be helpful?

Your feedback is highly appreciated.

Kind regards..
Ali