Evaluation metrics and objective functions
In a Kaggle competition, you can find the evaluation metric in the left menu on the Overview page of the competition. By selecting the Evaluation tab, you will get details about the evaluation metric. Sometimes you will find the metric formula, the code to reproduce it, and some discussion of the metric. On the same page, you will also get an explanation about the submission file format, providing you with the header of the file and a few example rows.
The association between the evaluation metric and the submission file is important, because you have to consider that the metric works essentially after you have trained your model and produced some predictions. Consequently, as a first step, you will have to think about the difference between an evaluation metric and an objective function.
Boiling everything down to the basics, an objective function serves your model during training because it is involved in the process of error minimization...
