In this task, we ask you to train a model that learns to rank properties.

Hotels.com travellers provide their trip information, like destination, holiday dates and number of people in the search. Our search engine then returns a list of properties, ranked by their relevance to that traveller. 
We provide you a data set with this travel information augmented by some features we have found to be predictive, including search-level features like booking window (length of time between booking the trip and taking it), and property-level features, like the country the property is in. 
Please feel free to model with all, some or none of these, or create your own features.
The label in this data set is booking_bool, indicating whether this property was the one that was booked by the customer.
Please provide justification for your metric for this ranking task.

In the leader board, you’ll see the popularity benchmark. This is a useful baseline. It’s the score you’d achieve on the test set if you simply ranked the properties in each search by the number of times it was booked in the training set.

Happy modelling, and good luck!