A Comparative Study of Machine Learning and Linear Regression in Prediction of Pricing: A Case Study of Used Cars Price Prediction in Turkey

Used cars market is very buoyant in the World as well as in the Turkey and has a significant size. While sellers are looking for the maximum price for their vehicles, buyers are willing to buy a good quality vehicle with a minimum price. However, during the trading process various problems might occurred and even dealer might not have standard pricing policy for a specific vehicle. In this study, it is focused on whether a pricing policy for a used car can be created or not, collected real data, and applied methods based on statistics and artificial intelligence. As a subbranch of artificial intelligence, In addition to Machine learning’s success in complex problems such as natural language processing, text mining, image processing, it can also outperforms classical methods such as linear regression to solve regression problems. From this point, during this study Artificial Neural Networks, Support Vector Machines and Linear Regression methods are used and compared with the aim of standardizing used cars pricing. Based on the findings, it is possible to standardize used cars pricing with machine learning methods and chosen methods gives better results than classic linear regression method.