A Hybrid Approach to Vehicle Price Prediction: Combining PCA and Supervised Learning

Abstract

Accurately predicting car prices is a complex task that involves analyzing multiple interacting factors. This study proposes a comprehensive methodology that integrates dimensionality reduction, clustering, and supervised learning to enhance the predictive accuracy of car price models. Using a dataset of 399 vehicles described by ten numerical and categorical features, we first apply Principal Component Analysis (PCA) to reduce dimensionality while preserving essential information. Next, K-means clustering is used to identify natural groupings within the data, revealing meaningful market segments. Finally, we compare the predictive performance of two supervised learning algorithms: Multiple Linear Regression (MLR) and Support Vector Machines (SVM). The results demonstrate that MLR achieves superior performance, with a coefficient of determination (R²) of 0.92 and a Root Mean Square Error (RMSE) of 2,310.12, compared to SVM's R² of 0.85 and RMSE of 3,212.14. These findings underscore the value of combining exploratory analysis with predictive modeling and suggest that MLR remains highly effective when relationships among variables are largely linear. This integrated approach provides valuable insights for stakeholders in the automotive and insurance sectors seeking to assess vehicle value accurately.