A collaborative filtering recommendation algorithm based on improved similarity measure method

Abstract

Collaborative filtering recommendation algorithm is one of the most successful technologies in the e-commerce recommendation system. With the development of e-commerce, the magnitudes of users and commodities grow rapidly; the performance of traditional recommendation algorithm is getting worse. So propose a new similarity measure method, automatically generate weighting factor to combine dynamically item attribute similarity and score similarity, form a reasonable item similarity, which bring the nearest neighbors of item, and predict the item's rating to recommend. The experimental results show the algorithm enhance the steady and precision of recommendation, solve cold start issue.