A Comprehensive Research on Hybrid Recommender Systems for Cardiac Patients' Dietary Preferences: Integrating Multiple Algorithms for Enhanced Accuracy

Abstract

In this paper, we introduce a new recommender system called NBRS, specifically designed for making food recommendations to people with heart conditions. The NBRS system uses a combination of three different methods: a model that predicts based on what similar users like, a method that considers the most popular choices, and a personalized approach based on individual preferences. We tested the system using nine different techniques— BaselineOnly, CoClustering, KNNBasic, KNNWithMeans, KNNWithZScore, SlopeOne, SVD, SVD++, and Random Forest—to see how well they work in terms of accuracy, using measures like Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared (R²).
We gathered a lot of data from both online and in-person sources, focusing on the food preferences and nutritional needs of people with heart conditions. By analyzing this data, we found important patterns that help us understand what kinds of foods are best. The NBRS system significantly improves how accurately it can make recommendations by blending the strengths of these different techniques. This approach is better at suggesting suitable food options for new users, too. Our research shows the value of using several methods together to create a powerful tool for personalized nutrition advice, which could help improve the eating habits and health of people with heart issues