An Intelligent Machine Learning-Based System for Used Car Price Classification Using Ensemble Techniques

Abstract

The rapid growth of the automobile industry and online marketplaces has led to an increased demand for accurate pricing mechanisms for used cars. Determining the correct price of a used vehicle is a complex task influenced by multiple factors such as age, mileage, fuel type, transmission, ownership history, and engine specifications. Traditional pricing methods rely heavily on human judgment, which can be inconsistent and biased. To address this challenge, this project proposes an intelligent machine learning-based system for classifying used car prices into predefined categories. The system is developed using a combination of Django for the web framework and Scikit-learn for machine learning implementation. It leverages supervised learning techniques to analyze historical car data and predict the price range category of a given vehicle. The dataset is preprocessed and transformed into a structured format, where price values are categorized into three classes: below 5 lakhs, between 5 to 20 lakhs, and above 20 lakhs up to 100 lakhs. Multiple machine learning algorithms, including K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Logistic Regression, and Random Forest Classifier, are implemented and evaluated. To improve prediction accuracy and robustness, an ensemble approach using a Voting Classifier is employed. This approach combines predictions from multiple models to produce a final output, reducing the limitations of individual classifiers. The system also includes user authentication and role-based access, where users can input car details and receive predictions, while administrators can analyze trends, accuracy, and performance metrics. Visualization tools such as charts and reports are integrated to provide insights into prediction distributions and model effectiveness. Experimental results demonstrate that the ensemble model achieves higher accuracy compared to individual classifiers. The system provides a scalable and efficient solution for used car price classification, which can assist buyers, sellers, and dealers in making informed decisions. Overall, this project highlights the effectiveness of machine learning and ensemble techniques in solving real-world pricing problems and offers a practical implementation that can be extended for real-time applications.