<p>In the era of abundant digital information, recommender systems serve as critical bridges for data value exchange, enhancing personalized services and content discovery by effectively managing vast data volumes. Existing methods have illustrated the enhancement of recommendation precision and personalization through the aggregation of items’ latent feature matrix shared by others. However, these approaches often neglect the principles of fairness and security in data sharing, and lack mechanisms to oversee participant behavior during the collaborative training phase. To address these limitations, we propose a novel framework that integrates blockchain technology with recommender systems, leveraging smart contracts and attribute-based encryption to facilitate a trusted and equitable environment for data exchange. This amalgamation ensures that data sharing occurs without reliance on central authorities, thereby preserving user privacy and data integrity. Moreover, to encourage positive participation and penalize detrimental actions within the training process, we introduce an innovative appeal algorithm. This mechanism autonomously monitors participant conduct and dynamically adjusts token incentives from the blockchain, fostering a cooperative ecosystem. Our empirical evaluations demonstrate that the proposed blockchain-assisted matrix factorization (BAMF) scheme achieves commendable recommendation accuracy and robustness in decentralized networks. By mitigating the deficiencies of current methodologies, BAMF paves the way for more secure and efficient collaborative filtering in the realm of big data.</p>