1

I am trying to predict future forecasting of COVID-19 data using Polynomial Regression model and SVM model.

The plot of Test Data versus Polynomial Regression Predictions come as: enter image description here

MAE: 2073239.576368933

MSE: 6776173669404.772

R-Squared Score:0.8736469328226459

The plot of Test Data versus SVM Predictions come as:

enter image description here

MAE: 27215388.500802647

MSE: 1027135705300895.9

R-Squared Score: -1.0060054294340914

From above errors values & plots of both the models, how can i say that which model is performing better than the other? and what else can i study from all the results?

Dave
  • 62,186
Soumya Shree
  • 13
  • What is on the horizontal axes of your plots? Have you tried scatterplots, with the actual on the horizontal and the predicted value on the vertical axis? – Stephan Kolassa Oct 22 '20 at 05:44

1 Answers1

0

Your polynomial regression has a lower mean absolute ERROR, lower mean squared ERROR, and higher $R^2$, which is a function of mean squared error, thus redundant. ($R^2$ is useful in linear regressions, which your support vector regression is not, as it has an interpretation about the "amount of variance explained” that does not apply to nonlinear regression models.)

Whether you should care about the MAE or MSE is a decision for you to make, and if you were in a position where one model had superior MSE and the other superior MAE, you would have some decisions to make. However, your polynomial regression model has lower MAE and lower MSE than the support vector regression.

It sounds better to me to have lower error than higher error. Thus, conclude that the polynomial regression has superior performance.

Dave
  • 62,186