soupagni / twittertwicker Goto Github PK
View Code? Open in Web Editor NEWThis project forked from tgore03/twittertwicker
A recommendation system for twitter that identifies new followers for a user based on the popularity of words within his tweets. The described method maps the tweets to TF-IDF vectors in a sparse zero matrix and clusters this data using k-means clustering. The method then computes a new TF-IDF vector array for the new tweet and classifies it to one of the clusters obtained from k-means.