I have a dataset which contains vectors that generated from subtitles and have been L1 normalised, I want to calculate cosine similarity & Euclidean distance, I thought it is better if I use Z-score or Min-Max normalization, is it necessary and why? The L-1 normalization only guarantee the scale of each feature is 0-1 right?

(The results before and after Z-score or Min-Max normalization are different, seems Min-max has better result, not sure why) Many thanks.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.