Machine Learning model performance metrics & charts with confidence intervals, optimized with numba to be fast
Project description
fronni
A Python library for quickly calculating & displaying machine learning model performance metrics with confidence intervals.
How fronni works?
Requirements
-
Python >= 3.6
-
numba
-
numpy
-
scikit-learn
-
plotly
Installing fronni
pip install fronni
Full documentation
Functions from the classification module:
classification_report
Generates confidence intervals for precision, recall, & F1 metrics for a binary or multi-class classification model, given arrays of predicted & label values.
| Parameter | Type | Default |
|---|---|---|
| label | Numpy array or Pandas series | None |
| predicted | Numpy array or Pandas series | None |
| n | integer, number of bootstrap iterations | 1,000 |
| confidence_level | integer value between 1 & 100 | 95 |
| as_dict | Boolean, return nested dictionary if True otherwise Pandas dataframe | False |
| confidence_level | value between 1 & 100 | 95 |
| sort_by_sample_size | Boolean, return the Pandas dataframe, sorted in descending order of class sample size | False |
plot_classification_report
Plots precision, recall, & confidence intervals for F1 metrics for a binary or multi-class classification model, given a classification report input.
| Parameter | Type | Default |
|---|---|---|
| report | output from classification_report | None |
| save_to_filename | string, path of filename image to save like "image.png" | None |
From the regression module:
regression_report
Generates confidence intervals for RMSE, MAE, and R^2 metrics for a regression model, given arrays of predicted & label values.
| Parameter | Type | Default |
|---|---|---|
| label | Numpy array or Pandas series | None |
| predicted | Numpy array or Pandas series | None |
| n | integer, number of bootstrap iterations | 1,000 |
| as_dict | Boolean, return nested dictionary if True otherwise Pandas dataframe | False |
See the CONTRIBUTING file for how to help out.
License
fronni is Apache 2.0 licensed, as found in the LICENSE file.
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
File details
Details for the file fronni-0.0.6.tar.gz.
File metadata
- Download URL: fronni-0.0.6.tar.gz
- Upload date:
- Size: 9.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/1.15.0 pkginfo/1.7.1 requests/2.26.0 setuptools/41.0.1 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/2.7.18
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 02be7872f5b97c5962c001e87210bc63ac3d8abd42f4e63c55a593a53909bf8d |
|
| MD5 | c2a95217e9cb64e504ff516bb759f49c |
|
| BLAKE2b-256 | 1285177db228e5b587a9a4bb5ecb9cd2dfe4f7a6b360a7c0063bb1287e891cdc |
