Help: Wizard ‐ Challenge ‐ Baseline

You must supply a so-called "Starting kit" that will allow the participants to create sample code and result submissions. DOWNLOAD AN EXAMPLE.

We recommend to provide in the starting kit:

1. A README file (preferably in the form of a iPython notebook.
2. Some sample data resembling the task of the challenge, but SMALL for easy download and practice.
3. Some examples of data visualisation functions.
4. The metric function used to score the challenge.
5. A sample result submission ready-to-upload, providing predictions in the correct format.
6. An example of predictor class (a classifier or regressor). If scikit-learn is used, this class will have a "fit" and a "predict" method.
7. An example of "unit test" for the predictor class, for example a cross-validation loop calling "fit" to train and "predict" to test, then calling the scoring metric to compute performance.
8. A call to a function that zips the code or prediction results to create ready-to-upload submissions.

As a further refinement, you can make it that the starting kit itself can just be uploaded as a sample submission too!