Predicting Second-Hand Car Sales Price Using Decision Trees and Genetic Algorithms

It is important to predict the sales price of second-hand car for both persons and institutions who are operating in second-hand market. The sales price of cars are affected by many factors which makes predicting difficult. Especially there is no readily available method to determine which factors are affecting the sales price most. The purpose of this study is to predict the sales price of second-hand cars with decision trees. Genetic Algortihm is used to select the most relevant features. For this purpose, 252645 advertisements are scanned fort his study. For each advertisement there are 139 features available. Different models are examined using genetic algorithms with selecting 5, 10, 15 and 20 features. The best predicting performance in the out-of-sample experiment is 65.67%. Proposed model can be used as a decision support system for those operating in second-hand car market