Can We Estimate a Useful Range of Prices for Popular Models of Used Cars Using Past Data?

There are some models of cars that remain popular with the consumers, even if pre-used. However, they sell at a wide variety of prices. We built a dataset of past prices for two popular Toyota SUVs, Harrier and RAV4 using the final selling price obtained from a large Japanese auction house, USS. We then tried to predict a useful range of prices for each of these two models of pre-used Toyota SUVs, based on the parameters of that particular car like how many kilometers has it already travelled, whether the powertrain is hybrid or runs on gasoline only, which year was it originally built in, 2019 or in 2020 or later, etc. We also tried to find out whether options like having Sunroof/moonroof had any effect on the final bid of auction prices. Feature Selection and Feature Transformation were done using the Classification Learner App of the MATLAB Statistics and Machine Learning Toolbox version R2020b. Our analysis shows that we can establish a useful range of prices.