17. Analyzing and forecasting used car selling prices using the linear regression method

Abstract

The used car market has been rapidly expanding both globally and in Vietnam, creating a demand for more accurate price prediction models to support sales and market transparency. This study applies the multiple linear regression method to analyze and predict used car prices using a dataset of more than 4,000 records from Kaggle. The research process includes data preprocessing, exploratory data analysis, regression modeling, and performance evaluation. Results indicate that model year, mileage, brand, and transmission type significantly influence car prices. The regression model achieved an R² value of 0.82, demonstrating its effectiveness in explaining price variation. This study contributes not only academically but also practically by supporting dealerships, showrooms, and consumers in decision-making. Future work will focus on extending the dataset to the Vietnamese market and comparing linear regression with advanced machine learning models.