Theory of Complex Networks
Tutorial 4
17.Pearson correlation similarity derivationsolution
Show that the definition of the Pearson correlation similarity
can be derived from the definition of the Pearson correlation of two random variables (u, v), upon choosing .18.Equivalence of cosine similarity formssolution
Show that the two expressions given for the cosine similarity
of two nodes i and j in a nondirected graph are identical. In other words, show that .19.Bounds and equality conditions for cosine similaritysolution
Show that
, and that if and only if .20.Book recommendation using cosine similarity on bipartite graphsolution
The graph is a bipartite graph of books (squares) and customers (circles). Each edge links a book to a customer if that customer purchased the corresponding book. Customer 5 would like to purchase a new book. Please provide a recommendation to this customer by using the cosine similarity.
21.Equivalence of Pearson correlation similarity formssolution
Show that the two expressions given for the Pearson correlation similarity
of two nodes i and j in a nondirected graph are identical.22.Bound on Pearson correlation similaritysolution
Show that
. Hint: define for each node i the vector with entries , and write in terms of and . Next, use again the Cauchy-Schwartz inequality.
Solutions
17.Pearson correlation similarity derivation
No solution yet.
18.Equivalence of cosine similarity forms
No solution yet.
19.Bounds and equality conditions for cosine similarity
No solution yet.
20.Book recommendation using cosine similarity on bipartite graph
No solution yet.
21.Equivalence of Pearson correlation similarity forms
No solution yet.
22.Bound on Pearson correlation similarity
No solution yet.