Which method do you think is more suitable for computing user similarities between two users in this context? and why?

SMC(Simple Matching Coefficient) or Jaccard Coefficient.

Hint: The answer depends on whether the similarity should account for all movies (regardless of whether any of the two users watched them or not), or whether it should only rely on the set of movies watched by either users and disregard movies that were never watched by any of them.

Another Hint: In real user-movie datasets, there are tens of thousands of movies. Each user usually watches a few 100's of movies. That is, his row will contain a few 100's of 1's and the remaining 10's of thousands of cells will have 0's.

Similarities Between Binary Data Points This dataset represents user-movie watching records