Prediction of Average Fuel Consumption in Automobiles using Ensemble Stacking in Python
Abstract
In this paper, we predicted the mpg of the automobile through ensemble stacking. The ensemble method is used to estimate the mpg and provides improved predictive performance, since it uses multiple machine learning algorithms. The stacked regressors used are LASSO Regression, Elastic Net Regression, Kernel Ridge Regression, Gradient Boosting Regression and the meta models are XGBoost, LightGBM. We have used Root Mean Squared Logarithmic Error (RMSLE) function to calculate the relationship between the values predicted by the machine learning model introduced in this model and the actual target value i.e., mpg of the automobile.