It is very important to predict car price before launching it in the market. In the research, regression models are developed to predict the price of the car. Three models have been developed in the research paper: Backward Elimination, Backward Elimination with VIF, and forward selection. The data is taken from Kaggle. The most important factors are decided by correlating other variables with the car price. A linear regression model is finally developed, with engine size as the most influencing factor, the type of driver as the second influencing factor, and the type of the car body as the third influencing factor. Linear regression model predicts the car price with good model accuracy. The exploratory data analysis is done to know about the data set. The variables having variance influence factor r more than ten are omitted to avoid the problem of multicollinearity. The first model developed is forward selection, in which engine size is used to build the first regression model having a single variable. The value of adjusted R2 is 0.764, and the aim is to increase the value of this factor, and all the coefficients in this model are statistically significant. The second variable included is the type of carburetor (2bbl) that is incorporated in the model, and a regression model is developed. The adjusted R2 is 0.778 and all the coefficients are statistically significant. The third regression model is developed by incorporating types of the drive (Reverse drive), and the value of adjusted R2 is 0.802, and all the coefficients are statistically significant. Further, it was tried by the hit and trial method incorporated the other variables in the model to increase the model accuracy and adjusted R2, but there was no benefit. The types of variables are selected based on correlation and VIF. The second approach adopted to build the regression model is backward elimination, in which all the variables are included and eliminated one by one based on VIF and statistical significance. The adjusted R2 is 0.919, but some variables are statistically insignificant as the p-value is more than 0.05 with a 95% confidence interval. After eliminating all the variables having VIF of more than ten and statistically significant, the final regression model has adjusted R2 is 0.863. The third approach adopted is backward elimination, where only statistically significant factors are considered. The final regression model with backward elimination has adjusted R2 is 0.863. Finally, we recommend the forward selection method of regression to predict the price of the car as it has less omitted variables bias.

Linear regression, correlation, forward selection, backward elimination, data analysis, Backward elimination

[1]. forecasting for used cars using adaptive” Expert Systems with Applications”, Vol. 36, Issue 4, pp. 7809-7817, 2009. https://doi.org/10.1016/j.eswa.2008.11.019
[2]. https://www.mckinsey.com/industries/automotive-and-assembly/our-insights/five-trends-shaping-tomorrows-luxury-car-market accesed on 29-04-2023.
[3]. J.J. Vilchez, R. Pasqualino, Y. Hernandez, “The new electric SUV market under batt of Cleaner Production” Journal of Cleaner Production, Vol. 383, pp.135294, 2023.
[4]. K. Noor, S. Jan, “Vehicle Price Prediction System using Machine Learning Techniques”. International Journal of Computer Applications, Vol. 7, Issue 9, pp. 27-31, 2017.
[5]. S. Gongqi, W., Yansong, Z. Qiang, “New Model for Residual Value Prediction of the Used Car Based on BP Neural Network and Nonlinear Curve Fit. In the proceeding of the In Measuring Technology and Mechatronics Automation (ICMTMA), IEEE ,Vol. 2, pp. 682-685, 2011.
[6]. A. Gokhale, A. Mishra, D.R. Veluchamy, “Factors Influencing Purchase Decision and Consumer Behavior in Luxury Cars” Dogo Rangsang Research Journal, Vol. 11, Issue 7, pp. 176-188, 2021.
[7]. Y. Rü?tü. “Determinants of Used Car Demand Price: Empirical Evidence from Turkey”, The Empirical Economics Letters, 2018).
[8]. Ufuk Demiro?lu, Ça?lar Yüncüler, “Estimating light-vehicle sales in Turkey, Central Bank Review, Vol. 16, Issue 3, pp. 93-108, 2016. https://doi.org/10.1016/j.cbrev.2016.08.003.
[9]. P. Venkatasubbu, M. Ganesh, “Used Cars Price Prediction using Supervised Learning Techniques” International Journal of Engineering and Advanced Technology Vol. 9, Issue 3, pp. 216-223, 2019. DOI:10.35940/ijeat.A1042.1291S319
[10]. B. Lavanya, S. Reshma, N. Nikitha, M. Namitha, “Vehicle resale price prediction using machine learning” UGC Care Group I. (2021).
[11]. M. T., Pratola, H. A. Chipman, E. I., George, R. E. McCulloch , “Heteroscedastic BART via multiplicative regression trees. Journal of Computational and Graphical Statistics,” Vol. 29, Issue 2, pp. 405-417, 2020.
[12]. A. Alhowaity, A. Alatawi, H. Alsaadi, “Are Used Cars More Sustainable? Price Prediction Based on Linear Regression”. Sustainability, Vol.15, Issue 2, pp. 911, 2023.
[13]. L. Bottmer, C. Croux, I. Wilms, “Sparse regression for large data sets with outliers” European Journal of Operational Research, Vol. 297, Issue 2, pp.782-794, 2022.
[14]. C. K., Puteri, L. N. Safitri, Analysis of linear regression on used car sales in Indonesia. In Journal of Physics: Conference Series Vol. 1469, No. 1, pp. 12143, IOP Publishing. 2020.
[15]. K. S. A. Kumar,. “Used Car Price Prediction using K-Nearest Neighbor Based Model”, International Journal of Innovative Research in Applied Sciences and Engineering, Vol. 4, pp 686, 2020.
[16]. R. Sutaria, R. Jain, “Auto-Price Forecast: An Analysis of Car Value Trends”. In 2023 4th International Conference for Emerging Technology (INCET) (pp. 1-6). IEEE, 2023.
[17]. E. Gegic, B. Isakovic, D.Keco, Z.Masetic, J. Kevric, (2019). “Car price prediction using machine learning techniques”, TEM Journal, Vol.8, Issue 1, pp. 113, 2019.