Experimenting with various regression models, we try to arrive at the most accurate prediction of the CO2 emission of a new car model. Following are the regression models we experiment with:-
• Single Variable Linear Regression - To begin with, we build a ML model considering only 1 independent variable- Engine Size and try to establish a best fitting line to determine the CO2 emissions.
• Multiple Variable Linear Regression - Moving ahead, we use all paramteres which linearly affect the CO2 emissions- Engine Size, No. of cylinders and Fuel Consumption build a ML model to determine the CO2 emissions.
• Polynomial Regression - We try to fit polynomials of different degrees to determine the best fitting curve to determine the CO2 emissions.
• Non-linear Regression - We move to a different dataset, since CO2 emissions are mostly linearly dependent on all its independent parameters. We explore the different kinds of non-linear curves viz. exponential, sigmoidal, logarithmic, parabolic etc. and try to find out the best fitting curve to determine the CO2 emissions.
Analysis is done using Python Scikit-learn library on Jupyter notebooks. Accuracy of each model is verified using Residual MSE, Mean abosolute error, R-squared value and Explained-variance.