We have collected data samples that can be used for training. Samples are taken every 15 minutes starting from two weeks before the deadline and stopping at 0 minutes until the deadline. Training data has been split into a training set and a test set. They contains samples with the following fields:
Meal day of week
Minutes left until deadline (0 <= x <= 20160)
Current registrations on sample
Known registrations for meal c-1 at sample time
Known registrations for meal c-2 at sample time
Known registrations for meal c-3 at sample time
Known registrations for meal c-4 at sample time
Known registrations for meal c-5 at sample time
Known registrations for meal c-6 at sample time
Known registrations for meal c-7 at sample time
Registrations for meal p-1
Registrations for meal p-2
Registrations for meal p-3
Registrations for meal p-4
Registrations for meal p-5
Registrations for meal p-6
Registrations for meal p-7
Registrations for meal p+0
Registrations for meal p+1
Registrations for meal p+2
Registrations for meal p+3
Registrations for meal p+4
Registrations for meal p+5
Registrations for meal p+6
Registrations for meal p+7
Where c-1 means The meal the day before the current meal and p+0 means the meal the meal a year before the current meal. The date for the previous year is determined by taking the iso year, week number and day of week, subtracting one from the year and converting to a date again. This makes sure that c+0, c-7, p+7, p+0 and p-7 all fall on the same day of week.
Remarks
- During the first few weeks of usage of <a href="https://noms.debolk.nl>Bolknoms
no registration creation timestamps were stored. These meals only include the final sample (minutes = 0)
 - The meals in the first year don't have data for the (p-7 - p+7) fields, they have been filled with zeros. These meals are not included in the validation set.
Submission format
The goal is to read in the samples in challenge.csv, predict the corresponding result column and submit them in the format in submission-sample.csv.