Enhancing Recommendation Systems: A Comparative and Optimization Study of KNN-Based Algorithms
Avani Sharma, Amritanshu · 2024
Recommendation systems emerged as a vital tool in this digital era for navigating the immense ocean of available content by offering personalized suggestions to users. As the accuracy of these systems relies on their capacity to learn from data, optimizing their predictive models, especially through hyperparameter tuning, becomes a crucial undertaking. This study investigates one of the algorithms used in predictive modeling, knearest neighbours (KNN), by exploring SURPRISE library. Our work aims to provide a complete comparative analysis of KNN base models and perform hyperparameter tuning for each of them so that the practitioners can make an informed decision. To validate our analysis, we have used MovieLens and BookCrossing datasets and measured the performance of model using different performance matrices.