Machine Learning-Powered Car Recommendation System: A Content-Based and Collaborative Approach

This research introduces an advanced car recommendation system designed to provide tailored guidance to consumers in the car purchasing process. By incorporating user input and preferences expressed in natural language, the system offers personalized recommendations based on vehicle body types. Comparative analysis reveals the multiclass random forest algorithms’ superior performance in content-based filtering and item-item collaborative filtering, surpassing neural networks for multiclass classification, logistic regression, and SVM algorithms. The hybrid approach, combining these techniques, demonstrates its effectiveness in delivering high-quality, personalized suggestions. This research not only contributes to the field of recommendation systems but also offers a practical tool to assist individuals in navigating the process of car selection.