A Novel Recommendation Algorithm Based on Diffusion of Innovation Theory

Abstract

Determining whether an item is "novel" for target users is a crucial problem and understanding their preference and awareness degree of the item has been widely investigated in novelty recommendations. A recommendation algorithm based on "diffusion of innovation" theory was proposed in this study to verify the novelty of recommendation results on the precondition of ensuring their accuracy. Items were clustered using the K-means clustering algorithm. Users' positivity of innovation adoption for each item was calculated on the basis of the items adopted by the users, and then the novelty of the items for target users was calculated by combining their popularity change, user preference, and user difference. Results of traditional recommendation algorithms were integrated for recommendation on the basis of fusion strategy results. The effectiveness of the proposed algorithm was verified through an offline experiment. Results indicate that the novelty of the recommendation list of the proposed algorithm is remarkably higher than that of traditional algorithms. The novelty is high when the quantity of alternative sets reaches 400 or more, where the average popularity of the recommendation list declines by 40%, and the coverage is elevated by 100%, thereby improving the ability of the proposed system to extract all kinds of items. This study serves as reference for the improvement of user satisfaction with recommendation systems.