Best metrics to evaluate the performance of a regression model?

Question

I've just started with machine learning and I have a lot to learn but one of the recent problems I'm facing is evaluating the performance of a regression model. I know about MSE, RMSE, MAE theoretically but cannot fully apply them while trying to evaluate the model performance. I'd really appreciate all the answers. So this is one of the problems that I need advice on. Here first 6 columns are X features and the Last column that is '2011 Revenue' is my y label.

Based on the X features, I created various regression models and using grid search I trained these models according to the respective best_params. Here are the result of all the metrices for each of the model.

I can see various metrics and I feel overwhelmed by that, So can anyone please guide me and tell me out of these various metrices, which one I should give more preference and how I can evaluate my model performance. Also how well my models are performing?

Answer 1

Mostly depends on the business problem statement & how you want to convey your results.

Business folks would like MAPE:

Mean absolute percentage error, this atleast gives you how good your model is doing on a percentage basis. Its sensitive to outliers so keep that in mind. This is good to talk about as on percent basis you can tell how close the predictions are.

Rsquared and Adjusted Rsquared i use it mostly for feature reduction.

I also look at p-values (also VIFs) of coefficients when explanation/contribution of features is important.

And MSE (square of RMSE) is used during training as the loss function.