WRMSSE score for the M5 dataset
Project description
M5 WRMSSE
Calculate the WRMSSE of a 28-day forecast for the M5 competition hosted by Kaggle. Instead of uploading submission files to Kaggle for an accuracy score, install the m5-wrmsse package and calculate it locally.
For more information on the derivation, visit
https://www.pmorgan.com.au/tutorials/wrmsse-for-the-m5-dataset/
Installation
Clone the repo
git clone git@github.com:pmrgn/m5-wrmsse.git
Or download and install the package using pip
pip install m5-wrmsse
Usage
The wrmsse
function returns the WRMSSE of a 28-day forecast, equivalent to what Kaggle calculates for it's public leaderboard. First, import the function
from m5_wrmsse import wrmsse
Pass your forecast as a numpy array to the function, which must be of shape (30490, 28).
my_forecast = np.ones((30490,28)) # Forecast example containing all ones
score = wrmsse(my_forecast)
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for m5_wrmsse-1.0.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2c925862d8558feafa9990b29c9ba5b1ccc323119792c6215eb69811a20a9f53 |
|
MD5 | 8609ef8b2f5734e2928160453a71f95e |
|
BLAKE2b-256 | d7608f96d58666d7493feaab031da34527777124c39584803f099c8016c35829 |