<![CDATA[Determining a fair market price for a used vehicle is a significant challenge for both sellers and buyers due to a lack of data transparency and the variety of influencing factors such as brand, production year, and mileage. This research aims to address this issue by developing a market data-based used vehicle price prediction system using a machine learning approach. The methodology adapts the CRISP-DM framework. Data is collected through web scraping from leading online marketplaces and processed through cleaning, normalization, and encoding before being used for modeling. Various regression algorithms were implemented, and the Linear Regression model was chosen for its optimal performance. The model was evaluated using the R Squared metric, yielding a score of 74% on the training data and 76% on the test data, demonstrating good accuracy and adaptability to new data. The best model was then implemented into a simple user interface based on Streamlit, allowing users to get a more objective recommendation for buying and selling prices. Overall, this system has great potential to facilitate more efficient and transparent transactions in the used automotive market, helping users make smarter and more profitable decisions.]]>
Copyrights © 2026
