Description

Customers cancelling reservations severely impact the results of hotels and dining facilities. It not only damages revenue, but also wastes the food in stock and personnel expenses used to provide services.

The goal of this competition is to use data consisting of over 100,000 reservations to categorize reservations into those that will and will not be cancelled. The data uses actual reservations that were made in several cities across Portugal. This is a challenge that predicts cancellations. It’s also very meaningful for society because it can later be used for various purposes such as analyzing the causes behind the cancellations.

Since the data is in chronological order, you will be able to learn data processing methods that can be applied in various fields.

Evaluation

AUC
Submissions are evaluated on area under the ROC curve between the predicted probability and the observed target.
Use submit.csv to enter the predicted values and submit the file.
This leaderboard is calculated with approximately 30% of the test data. The final results will be based on the other 70%, so the final standings may be different.

Prizes

1st Place - 500USD
2nd Place - 300USD
3rd Place - 150USD

Data Description

The data set used in this challenge consists of 40,000 reservations from resort hotels and 80,000 reservations from city hotels. The resort hotel data comes from the Algarve region and the city hotel data comes from Lisbon, the capital city.

The data was obtained from the hotels’ property management system (PMS). The names of the hotels and booking agents have been redacted.

The task is to use data in which the reservation date is between July 1, 2015 to June 30, 2017 to predict whether reservations for July 1, 2017 to August 31, 2017 will be cancelled.

A downloadable baseline has been provided for people participating in the competition for the first time.

Columns

The image types and corresponding numbers are as follows.

hotel：Hotel (Resort Hotel or City Hotel)
lead_time：Number of days that elapsed between the entering date of the booking into the PMS and the arrival date
stays_in_weekend_nights：Number of weekend nights (Saturday or Sunday) the guest stayed or booked to stay at the hotel
stays_in_week_nights：Number of week nights (Monday to Friday) the guest stayed or booked to stay at the hotel
country：Country of origin. Categories are represented in the ISO 3155–3:2013 format
market_segment：Market segment designation. In categories, the term “TA” means “Travel Agents” and “TO” means “Tour Operators”
distribution_channel：Booking distribution channel. The term “TA” means “Travel Agents” and “TO” means “Tour Operators”
is_repeated_guest：Value indicating if the booking name was from a repeated guest (1) or not (0)
previous_cancellations：Number of previous bookings that were cancelled by the customer prior to the current booking
previous_bookings_not_canceled：Number of previous bookings not cancelled by the customer prior to the current booking
reserved_room_type：Code of room type reserved. Code is presented instead of designation for anonymity reasons
assigned_room_type：Code for the type of room assigned to the booking. Sometimes the assigned room type differs from the reserved room type due to hotel operation reasons (e.g. overbooking) or by customer request. Code is presented instead of designation for anonymity reasons.
booking_changes：Number of changes/amendments made to the booking from the moment the booking was entered on the PMS until the moment of check-in or cancellation
agent：masking
company：masking
days_in_waiting_list：Number of days the booking was in the waiting list before it was confirmed to the customer
customer_type：Type of booking, assuming one of four categories:Contract - when the booking has an allotment or other type of contract associated to it; Group – when the booking is associated to a group; Transient – when the booking is not part of a group or contract, and is not associated to other transient booking; Transient-party – when the booking is transient, but is associated to at least other transient booking
adr：Average Daily Rate as defined by dividing the sum of all lodging transactions by the total number of staying nights
required_car_parking_spaces：Number of car parking spaces required by the customer
total_of_special_requests：Number of special requests made by the customer (e.g. twin bed or high floor)

The annotation was done manually, and the expression variations are subject to the views of the annotator.

Distributed Data

data_all.zip
・train.csv -（learned data）
・test.csv -（predicted data）
・submit.csv -（file in the proper format）
・hotel_baseline.ipynb -（baseline for novices）

Rules

Use of outside data, pre-trained models

Training, inference, or prediction with models using data other than the distributed data is forbidden.
The use of APIs such as Google Cloud Vision API is forbidden. However, The use of open-source (source is clear, open access, and non-commercial) pre-trained models or libraries is allowed. When using pre-trained models, please specify the source (link to thesis, etc.).

Implementation Method

Only JupyterNotebook (Python or R) may be used to learn the model.
Divide the source code into preprocess, train, and predict as shown below so that the process progresses when each one is executed.

①preprocess
A module used to read the provided data, preprocess it, and output a file that can be entered into the model. Please divide the files for learning and evaluation like train.csv and test.csv and define the functions used for preprocessing.

②train
A module used to read the file created at ①. Please define the learned models, features, and functions used to output the cross-validation evaluations.

③predict
A module used to read the test data created in ① and the model created in ② to output a file with prediction results.

The proposed method must replicate results and be available for sustained use in ordinary environments without additional costs. The same logic must be used to make predictions with the same format and different data. For example, you are not allowed to make baseless revisions of predicted values.

Handling of Information

You may not share the data and source code relating to the task of this competition to other participants.

Receiving the Prize

Top finishers must submit their Notebook with implementation methods ①〜③ after the competition for source code inspections.

Cracking, cheating, impersonation, and any other actions that violate the rules will result in voiding the prize.

Disclosure Policy

License：CC BY-NC 4.0
https://creativecommons.org/licenses/by-nc/4.0/deed.ja
Source：Hotel booking demand, Data in Brief, 2020
https://www.kaggle.com/jessemostipak/hotel-booking-demand
Author：Jesse Mostipak

Metadata

Quevico Hotel booking demand: Mewcket.inc, Data in Brief, 2020
Dataset owner　Mewcket.inc
Last updated　2020-04-28
Data created　2020-04-28

System Usage

Individual participation
Each person may only use one account, the use of multiple accounts is forbidden.
Up to five submissions per day are allowed.

Competition Rules

Individual clauses are not specified.

Twitter Authority : Basic tasks of NLP

Predicting reservation cancellations: Predict whether resort/city hotel reservations will be cancelled

Cats disturbing : The expression estimation, Basic machine learning task.