EzMetrics
The eaziest metrics library ever
Ez metrics is a library that calculates the fitness metrics of your ML model.
Features
- Support for fixed binary classification as well as probability based classification
- Support for regression models
- Up to 6 different metrics
And of course EzMetrics itself is open source with a public repository on GitHub.
Installation
EzMetrics requires no extra libraries to run.
pip install EzMetrics
from EzMetrics import Metrics as ezm
Usage
Create an object containing a list with the predictions of your model and a list containing the actual values por each prediction.
exmpl_obj = ezm( predicted_list, observed_list)
Then just choose a metric suited for your data and use it. In case of Mean Absolute Error it would be as follows.
exmpl_obj.mae()
Metrics
EzMetrics has 6 different metrics available, which one to use depends on your data type.
Discrete classification | |
---|---|
Accuracy | Metrics.accuracy() |
F1 score | Metrics.f1() |
Probability classification | |
---|---|
Area Under the Curve (AUC) | Metrics.roc_auc() |
Regression | |
---|---|
R squared | Metrics.r2() |
Mean Absolute Error | Metrics.mae() |
Mean Squared Error | Metrics.mse() |
License
MIT