Package For Numerai users to check their models performance
Project description
Numetrics
Numetrics is a small package that has some functions regarding checking the performance of a model's predictions.
Version
- Github Release
v0.0.1
- PyPi Version
v0.0.3
Why did I create Numetrics?
- The first reason is: I want to contribute to the Numerai community. I have learned a lot from Arbitrage, JRB, Wigglemuse, Michael Oliver, the team, and everybody in the community.
- The second reason is: I was tired of copying and pasting my code throughout my notebooks, so I decided to create a package to make it easier for me to just install it through pip
How to install Numetrics?
You can install it with pip
pip install numetrics
import numetrics
How to use Numetrics?
You have one function called diagnose
this is the main function that does all the magic.
It takes a dataframe of your prediction. Prediction dataframe must have the following
- The id
- target column
- era
- predicton column
you can just pass only the predicton dataframe
numetrics.diagnose(preds_df)
or you can spicify the target name and the predicton name
numetrics.diagnose(preds_df,TARGET_NAME='target',PREDICTION_NAME='prediction_kazutsugi')
See the example file to learn how to use this package.
This is the first version/draft/attempt of the package. The code does not look pretty and it still missing lots of functionalities. However, I promise I will do my best to make it better ASAP.
Feel free to contribute to this package.
Change Log
0.0.1 (Nov/09/2020)
- First Release
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
numetrics-0.0.3.tar.gz
(3.4 kB
view hashes)
Built Distribution
Close
Hashes for numetrics-0.0.3-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 5843a2c9d94faa7d970c2cb2b7146f3510d6261cebacbeccd117cc87e67f6252 |
|
MD5 | 043cf2af4c973f548a60780acefb0f72 |
|
BLAKE2b-256 | d679a11081b1583002717cde4db7b0f30d5b0bbb2f8dfe1798d7758b635cac66 |