Maintaining User Security in Consumer Electronics-Based Online Recommender Systems Using Federated Learning

The recommender system is an important component of applications that are used in the real world. In this era of information overload, it has evolved into a necessity and a lucrative market for many online businesses, like consumer electronics and others. The service can't make personalized recommendations without having access to private data like a user's likes and dislikes, location, and purchasing history. Users are worry of giving out too much personal data, though, because it can be mistreated so readily. Thus, there are inescapable security issues that will be exposed through attempts at unauthorized entry while offering the recommendation solutions, especially in consumer electronics. While providing the recommendation, the important details must be hidden from the network to avoid privacy violations. This is accomplished by employing a federated learning method that is both successful and cost-effective. The users in the Federated Learning approach utilize their privately saved data and algorithms for prediction and computing model changes after receiving a primary machine learning model from the server. This work suggests a federated learning-based paradigm to perform matrix factorization through collaborative filtering. Our suggested method is easier to use and more reliable in practical settings when contrasted to a bench-mark federated recommender system. Our algorithm is tested on benchmark data, and its performance is measured in terms of the root-mean-squared error (RMSE) of user evaluation predictions. The suggested methodology outperforms other current solutions and displays considerably higher accuracy (96.32%) in terms of computational and communication overhead.