Performance Analysis of CatBoost Algorithm and XGBoost Algorithm for Prediction of CO2Emission Rating

- The objective of this project is to improve the accuracy of CO2 Emission Rating predictions through the development of a model that incorporates various features by comparing CatBoost algorithm over XGBoost algorithm. Materials and Methods: The clincalc tool was used to determine the total number of iterations N equals 20 (10 iterations for each group) and to predict novel
$\text{CO}_2$
Emission Rating with the improved accuracy. The significance value is obtained with the help of an independent sample T-test. Here the G-power analysis was carried out with 80%, alpha rate of 0.05. The novel CO
2
Emission Rating dataset with a size of 7530 was collected from Kaggle. Results: According to the findings, the difference in accuracy between the CatBoost algorithm is 98.01% and the XGBoost algorithm is 94.11% is statistically significant. For accuracy, the significant values are calculated
$\text{asp} =0.000(\mathrm{p} < 0.05)$
which concludes that they are statistically significant. Conclusion: The accuracy value of the CatBoost algorithm is 98.01% whereas the accuracy value of XGBoost algorithm is 94.11%. CatBoost performs better when compared to XGBoost.