Mitigate the cold start problem in Recommendation System based on Matrix Factorization and Similarity Technique
Chour Singh Rajpoot, Varun Tiwari, Santosh Kumar Vishwakarma · 2024
Recommendation System (RS) is a powerful information filtering artificial intelligence tool, which recommend relevant and authentic information for e-commerce user’s onecommerce platform like amazon, flip cart etc. but lack of availability in the rating or reviews for the new customer and items, complicate the prediction of user’s interest and result is less accurate recommendation that is cause of cold start issue in RS. Therefore, to overcome cold start issue in RS, we proposed sequence set similarity technique along with k-NN, Matrix Factorization(MF) and Singular value decomposition(SVD++) and Non negative Matrix factorization(NMF).The Python platform is used for its implementation, accuracy obtain on both Movielens dataset, ML-100k is $0.9454 \%$ and ML-1M is $0.9711 \%$.In this paper, proposed solution verify and validate performance of RS and addressing cold start issue.