Analysis of bipartite network for new product recommendation

摘要

Availability of product consumption history of users make a challenge to extract users' preference towards products and recommend product to the potential consumers. A customer-product-attribute relation can be represented using network. When a new product arrives the attributes will be clubbed to identify the customers who prefer those attributes and recommend the new product to them. This paper is primarily concerned about network representation of the data and extracts meaningful information which helps in recommendation of a new product. The empirical study shows significant improvement over few previous works.