Using Multiple Linear Regression (MLR) to predict the real cost of a car model

Abstract

In this paper, an analysis of basic car characteristics is taken into account to predict the real price of different automobile models. Multiple linear regression (MLR) analysis was performed on the data using structural equation modeling with JMP 17. My methodology is divided into three main steps: the first uses various statistical analysis techniques to evaluate and preprocess the data and collected variables; the second involves choosing the most significant variables using multiple methods. The final phase uses RMSE, AICc, BIC, Mallow’s Cp, and Adjusted 𝑅2 to compare the outcome of many MLR models built using the chosen variables. The collected findings indicate that the model produced with variables chosen using the Stepwise Selection approach performs better than the models utilizing other approaches, having the lowest AICc, RMSE, and highest Adjusted R2. In the results, a reasonable regression model acquired a remarkable ability to predict the price of car models.