Car price prediction using machine learning

Abstract

The objective of this study was to assess how machine learning (ML) techniques can be used to predict vehicle prices. Vehicle price prediction is a complex process, as there are many variables that can influence a vehicle's market price. The automotive industry has continued to grow, and there are an ever-increasing number of variables that can influence an automotive vehicle's price, such as manufacturer and model, fuel economy, additional features, etc. Automotive vehicle pricing prediction is important to many different parties who have an interest in the automotive market. This study will demonstrate the importance of collecting all relevant data about the automotive vehicle and pre-processing this data properly using a large dataset with as many characteristics of vehicles as possible and used in combination with various ML algorithms, including Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN).
Initially, the use of these algorithms in a single classifier method to predict automotive vehicle pricing was found to have some drawbacks. This initial study formed the basis of developing an ensemble (barrier method of combining the best features of the algorithms) method of predicting automotive vehicle pricing. The use of an ensemble method greatly increased the overall accuracy of the prediction's accuracy rate to 92.38%. The research concluded that there are many parameters to consider with respect to trade-offs between computational power and accuracy; the authors believe that the use of an ensemble ML method is a good option to help improve automotive vehicle pricing predictions.

Keywords