Project information
Summary
The evaluation metric used to measure the performance of the model is the mse (Mean Squared Error) metric, as my chosen case is a regression case. MSE basically measures the mean squared error of our predictions. For each point, it calculates the squared difference between the prediction and the target and then averages those values. The higher this value, the worse the model. The MSE value is never negative, as we square the individual prediction errors before summing them, but would be zero for a perfect model. Benefit: Useful if we have unexpected values that we have to care about. A very high or low value that we have to pay attention to. Weaknesses: If we make one very bad prediction, squaring will make the error worse and that might make the metric tend to overestimate the badness of the model. That is a very problematic behavior if we have noisy data (i.e., data that for whatever reason is not completely reliable) even a "perfect" model may have a high MSE in that situation, so it becomes difficult to judge how well the model is performing. On the other hand, if all errors are small, or more precisely, smaller than 1, than the opposite effect is felt: we can underestimate the badness of the model.