Unveiling Superiority: Evaluating Bernoulli Matrix Factorization in Recommender Systems with Ciao Dataset Dominance

This paper examines the complex landscape of recommender systems, focusing in particularon the effectiveness of Bernoulli Matrix Factorization (BeMF). The performance of BeMF issystematically assessed against renowned state-of-the-art models, TrustEV, GCFA, SBRNE,RAWATD, and PMF, utilizing a diverse array of datasets, including the widely used Ciaodataset. evaluation, centered on the critical metric of Mean Absolute Error (MAE), consistentlyreveals the superior accuracy and proficiency of our BeMF model, notably excelling on theCiao dataset. This thorough examination encompasses various dimensions, encompassing userpreferences, social trust, behavior integration, and innovative trust synthesis. Contributing tothe ongoing discourse in recommender system research, this study illustrates Bernoulli MatrixFactorization's versatility and potency, highlighting its ability to improve recommendationaccuracy and adaptability in varied scenarios.

کلیدواژه ها:

Recommender systems ، matrixfactorization ، collaborativefiltering