Recommend users to follow based on paper 'OWA operator based link prediction ensemble for social network'

[iandioch]

This involves implementing 9 (simple) node similarity metrics, and normalising their answers to [0, 1].

Then, a programming problem must be solved to find the vector w. The paper goes into 3 different methods to choose a w; just do one, and choose the orness(w) to be maximal based on their results, as choosing a different one is probably computationally infeasible.

Then, use scipy to solve this nonlinear programming problem, based on the given method (MEM, MVM, or CSM) and the chosen orness, to get a vector w.

Finally, generate recommendations using this vector, and calculate the accuracy of the system.