Considering N users, with indices i = 1, 2, 3, ..., N. Suppose that we have collected the data reflecting whether user i knows j and assume everyone knows
themselves (of course).
(1) Use the concepts in linear algebra to represent the data of who knows whom? (5%)
(2) How to measure the similarity between two users from the people they know, say they
are more similar if they know more users in common?