Performance Evaluation of Linear Regression Models Based on Specific Features in the Turkish Second-Hand Car Market
Nonlinear Dynamical Control, Computer Simulation and Optimization Systems: Theory and Applications: Volume 2, World Scientific Publishing , ss.251-263, 2025
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2025
- Doi Numarası: 10.1142/9789819815432_0023
- Yayınevi: World Scientific Publishing
- Sayfa Sayıları: ss.251-263
- Dokuz Eylül Üniversitesi Adresli: Evet
Özet
We evaluate the effectiveness of various linear regression models for predicting second-hand car prices in Turkey by analyzing a dataset from the country's leading second-hand car platform, comprising over 17,000 records and 25 attributes, including brand, model, age, mileage, and fuel type. We developed and compared Linear Regression, Ridge, Lasso, and Elastic Net models, finding that while traditional Linear Regression provides a solid baseline, Ridge, Lasso, and Elastic Net significantly enhance prediction accuracy by addressing multicollinearity and improving feature selection. Our results demonstrate the superiority of these advanced models for accurate price estimation in the secondhand car market.