Implementation of Friendbook: A Recommendation System for Social Networks

Abstract

A recommendation system is a software program which endeavors to limit determinations for clients in light of their communicated inclinations, past bahavior, or other information which can be mined about the client or different clients with similar interests. Existing social networking services recommend friends to users based on their social graphs, which may not be the most fitting to mirror a client's inclinations on companion choice, in real time. We present Friendbook, a novel semantic-based friend recommendation system for interpersonal organizations, which prescribes companions to clients in view of their life style rather than social graph. Friendbook finds life style of clients and measures the likeness of ways of life amongst clients, and prescribes companions to clients if their ways of life have high closeness. Inspired by achievements in the field of text mining, we model the daily life as life documents, from which his/her life styles are extracted by using the Latent Dirichlet Allocation algorithm. We utilize the probabilistic topic model to extract way of life data of clients. We propose a one of a kind similarity metric to portray the similarity of clients as far as life styles and after that prescribe companions to clients in light of their life styles. LDA algorithm is also useful to calculate the probability for each matched users. This system recommends the friends appropriately as compare to other state of art systems.