Research on Used Car Valuation Based on Gradient Descent Multiple Regression Model

Abstract:
Due to the particularity of "one car, one situation" of used cars, there is always the problem of non-standard pricing in transactions. We comprehensively consider various factors affecting the price of used cars and propose a gradient descent multiple regression algorithm to solve the problem of used car valuation. By preprocessing the 30000 sample data collected from C market of used car trading, analyzing and extracting the influencing factors affecting the sales price of used cars, and then gradient reducing each factor to obtain a reasonable used valuation model. Finally, using the trained model, we can complete the price estimation of used cars in the new sample.