Hi Michael,
I have a question for data preprocessing before sending them out for clustering: I think we directly compute the distance on original data without doing PCA to reduce its dimensions, will this affect the results of clustering, especially for distance measurements like Euclidean distance and cosine similarity?
Hi Michael,
I have a question for data preprocessing before sending them out for clustering: I think we directly compute the distance on original data without doing PCA to reduce its dimensions, will this affect the results of clustering, especially for distance measurements like Euclidean distance and cosine similarity?