Application Research of Collaborative Filtering Algorithm in Catering Recommendation System

Collaborative filtering algorithm is commonly used in recommendation systems. Based on the traditional collaborative filtering algorithm, this paper proposes an improvement scheme. When calculating similarity, an improved cosine similarity formula is first used, followed by adding time threshold parameters for weighting. Furthermore, the similarity results are re-weighted based on the number of common scoring items among users, and finally normalized.