Comparative Analysis of Ensemble Machine Learning Models for Personalized In-Vehicle Infotainment Recommendation Systems

Despite significant advancements of recommendation systems across various applications, recommendation systems within the automotive domain face a unique set of challenges and requirements. This paper underscores the need for a specialized approach in developing recommender systems for in-vehicle infotainment (IVI) systems, aimed at enhancing the driving experience through personalized, routine-based feature suggestions under appropriate contexts. Through the evaluation and comparison of various ensemble machine learning models, including Random Forest, LightGBM, XGBoost and CatBoost, our research identifies effective strategies for delivering context-aware recommendations. Moreover, a broad spectrum of IVI and contextual features from real-world driving data have been applied to the proposed algorithms, justifying the efficacy of the models. Consequently, our study demonstrates the potential for enhanced driving experiences through machine learning-based feature recommendations and lays the groundwork for future innovations in personalization in the automotive domain.