Evaluation Criteria

Submissions are evaluated using Accuracy Score.

Leaderboard:

This challenge does partial evaluation.
The initial/public leadeaboard is based on 50% of test data.
The final results (private leaderboad) will be displayed after the challenge ends.
So, please note that your results can vary in public and private leaderboards.

How do we evaluate?

Once we release the data, anyone can download it, build a model, and make a submission. We give competitors a set of data (training data) with both the independent and dependent variables.

We also release another set of data (test dataset) with just the independent variables, and we hide the dependent variable that corresponds with this set. You submit the predicted values of the dependent variable for this set and we compare it against the actual values.

The predictions are evaluated based on the evaluation metric defined in the datathon.

The public leaderboard that is calculated during the competition evaluates only 50% of the test dataset, whereas the private leaderboard will be calculated based on the remaining 50% dataset. Hence, you might see some variation in ranks during and after the Datathon.

Participants are also requested to submit their Notebook with proper documentation. We believe this helps to learn from each other in the community. Further, we also select the best notebooks and share them with the community on social media and other communication channels.

Deep Learning Bootcamp - Graded Challenge

Challenge Starts

Registration Ends

Challenge Ends