# TinyCP
TinyCP is an experimental Python library for conformal predictions, providing tools to generate valid prediction sets with a specified significance level (alpha). This project aims to facilitate the implementation of personal and future projects on the topic.
For more information on a previous project related to Out-of-Bag (OOB) solutions, visit [this link](https://github.com/HeyLucasLeao/cp-study).
## Changes about previous work
- `calibrate`: instead of `Balanced Accuracy Score`, it can be calibrated either `Matthews Correlation Coefficient` or `Bookmaker Informedness Score`, for better reliability
- `evaluate`: scores `bm` and `mcc` for more reliability
Currently, TinyCP supports Out-of-Bag (OOB) solutions for `RandomForestClassifier` in binary classification problems. For more options and advanced features, consider exploring [Crepes](https://github.com/henrikbostrom/crepes).
## Installation
Install TinyCP using pip:
```bash
pip install tinycp
```
> **Note:** If you want to enable plotting capabilities, you need to install the extras using Poetry:
```bash
poetry install --E plot
```
## Usage
### Importing Classes
Import the conformal classifiers from the `tinycp.classifier` module:
```python
from tinycp.classifier.class_conditional import BinaryClassConditionalConformalClassifier
from tinycp.classifier.marginal import BinaryMarginalConformalClassifier
```
### Example
Example usage of `BinaryClassConditionalConformalClassifier`:
```python
from sklearn.ensemble import RandomForestClassifier
from tinycp.classifier.class_conditional import OOBBinaryClassConditionalConformalClassifier
# Create and fit a RandomForestClassifier
learner = RandomForestClassifier(n_estimators=100, oob_score=True)
X_train, y_train = ... # your training data
learner.fit(X_train, y_train)
# Create and fit the conformal classifier
conformal_classifier = BinaryClassConditionalConformalClassifier(learner)
conformal_classifier.fit(y=y_train, oob=True)
# Make predictions
X_test = ... # your test data
predictions = conformal_classifier.predict(X_test)
```
### Evaluating the Classifier
Evaluate the performance of the conformal classifier using the `evaluate` method:
```python
results = conformal_classifier.evaluate(X_test, y_test)
print(results)
```
## Classes
### BinaryMarginalConformalClassifier
`BinaryMarginalConformalClassifier` A marginal coverage conformal classifier methodology utilizing a classifier as the underlying learner. This classifier supports the option to use Out-of-Bag (OOB) samples for calibration.
### BinaryClassConditionalConformalClassifier
`BinaryClassConditionalConformalClassifier` A class conditional conformal classifier methodology utilizing a classifier as the underlying learner. This classifier supports the option to use Out-of-Bag (OOB) samples for calibration.
## License
This project is licensed under the MIT License.
Raw data
{
"_id": null,
"home_page": null,
"name": "tinycp",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.10",
"maintainer_email": null,
"keywords": "machine-learning, conformal-prediction",
"author": "Lucas Le\u00e3o",
"author_email": "heylucasleao@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/f2/74/48c422d0c4bdbca89c1ef67d567363825d76d6bbaa3ec2c5265ac545d799/tinycp-0.0.5.tar.gz",
"platform": null,
"description": "# TinyCP\nTinyCP is an experimental Python library for conformal predictions, providing tools to generate valid prediction sets with a specified significance level (alpha). This project aims to facilitate the implementation of personal and future projects on the topic.\n\nFor more information on a previous project related to Out-of-Bag (OOB) solutions, visit [this link](https://github.com/HeyLucasLeao/cp-study).\n\n## Changes about previous work\n- `calibrate`: instead of `Balanced Accuracy Score`, it can be calibrated either `Matthews Correlation Coefficient` or `Bookmaker Informedness Score`, for better reliability\n- `evaluate`: scores `bm` and `mcc` for more reliability\n\nCurrently, TinyCP supports Out-of-Bag (OOB) solutions for `RandomForestClassifier` in binary classification problems. For more options and advanced features, consider exploring [Crepes](https://github.com/henrikbostrom/crepes).\n\n## Installation\n\nInstall TinyCP using pip:\n\n```bash\npip install tinycp\n```\n\n> **Note:** If you want to enable plotting capabilities, you need to install the extras using Poetry:\n\n```bash\npoetry install --E plot\n```\n\n## Usage\n\n### Importing Classes\n\nImport the conformal classifiers from the `tinycp.classifier` module:\n\n```python\nfrom tinycp.classifier.class_conditional import BinaryClassConditionalConformalClassifier\nfrom tinycp.classifier.marginal import BinaryMarginalConformalClassifier\n```\n\n### Example\n\nExample usage of `BinaryClassConditionalConformalClassifier`:\n\n```python\nfrom sklearn.ensemble import RandomForestClassifier\nfrom tinycp.classifier.class_conditional import OOBBinaryClassConditionalConformalClassifier\n\n# Create and fit a RandomForestClassifier\nlearner = RandomForestClassifier(n_estimators=100, oob_score=True)\nX_train, y_train = ... # your training data\nlearner.fit(X_train, y_train)\n\n# Create and fit the conformal classifier\nconformal_classifier = BinaryClassConditionalConformalClassifier(learner)\nconformal_classifier.fit(y=y_train, oob=True)\n\n# Make predictions\nX_test = ... # your test data\npredictions = conformal_classifier.predict(X_test)\n```\n\n### Evaluating the Classifier\n\nEvaluate the performance of the conformal classifier using the `evaluate` method:\n\n```python\nresults = conformal_classifier.evaluate(X_test, y_test)\nprint(results)\n```\n\n## Classes\n\n### BinaryMarginalConformalClassifier\n\n`BinaryMarginalConformalClassifier` A marginal coverage conformal classifier methodology utilizing a classifier as the underlying learner. This classifier supports the option to use Out-of-Bag (OOB) samples for calibration.\n\n\n### BinaryClassConditionalConformalClassifier\n\n`BinaryClassConditionalConformalClassifier` A class conditional conformal classifier methodology utilizing a classifier as the underlying learner. This classifier supports the option to use Out-of-Bag (OOB) samples for calibration.\n\n## License\n\nThis project is licensed under the MIT License.\n\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "A small toolbox for conformal prediction",
"version": "0.0.5",
"project_urls": null,
"split_keywords": [
"machine-learning",
" conformal-prediction"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "694a4341b3279843198cdfae517add816b457684bde6abe5ff5ed485a22045a1",
"md5": "0e7328bc6de6594c9d349af3d958e4eb",
"sha256": "6ce1a8f5d1a4913c1dc303243af687db26d537ed74257c7acba4adc70fc3170a"
},
"downloads": -1,
"filename": "tinycp-0.0.5-py3-none-any.whl",
"has_sig": false,
"md5_digest": "0e7328bc6de6594c9d349af3d958e4eb",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.10",
"size": 14813,
"upload_time": "2025-01-19T03:01:37",
"upload_time_iso_8601": "2025-01-19T03:01:37.829691Z",
"url": "https://files.pythonhosted.org/packages/69/4a/4341b3279843198cdfae517add816b457684bde6abe5ff5ed485a22045a1/tinycp-0.0.5-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "f27448c422d0c4bdbca89c1ef67d567363825d76d6bbaa3ec2c5265ac545d799",
"md5": "d8c5f2461a4a774a2fb2d1c394808991",
"sha256": "7bc8aec5410ca921035af674a3efa4e9c12556297af32b40d16fab0624d2fac5"
},
"downloads": -1,
"filename": "tinycp-0.0.5.tar.gz",
"has_sig": false,
"md5_digest": "d8c5f2461a4a774a2fb2d1c394808991",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.10",
"size": 11447,
"upload_time": "2025-01-19T03:01:40",
"upload_time_iso_8601": "2025-01-19T03:01:40.047918Z",
"url": "https://files.pythonhosted.org/packages/f2/74/48c422d0c4bdbca89c1ef67d567363825d76d6bbaa3ec2c5265ac545d799/tinycp-0.0.5.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-01-19 03:01:40",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "tinycp"
}
JSON