Data Intelligence, Business Analytics
For the below dataset, what would be the best similarity metric in recommending movies to user1?
Currently I'm using TFIDF to calculate weights for movie attributes and Cosine similarity to calculate the similarity values.If any attributes occurs more then its weight is coming down. For example: If the attribute ACTOR1 is present in 10 movies and ACTOR2 is present in 5 movies the weight for ACTOR2 is more than ACTOR1. How do I fix this condition?