# milwrap
[](https://github.com/inoueakimitsu/milwrap/actions/workflows/python-tests.yml)
<a href="https://github.com/inoueakimitsu/milwrap/issues"><img alt="GitHub issues" src="https://img.shields.io/github/issues/inoueakimitsu/milwrap"></a>
[](https://colab.research.google.com/github/inoueakimitsu/milwrap/blob/master/introduction.ipynb)
Python package for multiple instance learning (MIL).
This wraps single instance learning algorithms so that they can be fitted to data for MIL.
## Features
- support count-based multiple instance assumptions (see [wikipedia](https://en.wikipedia.org/wiki/Multiple_instance_learning#:~:text=Presence-%2C%20threshold-%2C%20and%20count-based%20assumptions%5Bedit%5D))
- support multi-class setting
- support scikit-learn algorithms (such as `RandomForestClassifier`, `SVC`, `LogisticRegression`)
## Installation
```bash
pip install milwrap
```
## Usage
For more information, see [Use scikit-learn models in multiple instance learning based on the count-based assumption](https://github.com/inoueakimitsu/milwrap/blob/main/intro-en.md).
```python
# Prepare single-instance supervised-learning algorithm
# Note: only supports models with predict_proba() method.
from sklearn.linear_model import LogisticRegression
clf = LogisticRegression()
# Wrap it with MilCountBasedMultiClassLearner
from milwrap import MilCountBasedMultiClassLearner
mil_learner = MilCountBasedMultiClassLearner(clf)
# Prepare follwing dataset
#
# - bags ... list of np.ndarray
# (num_instance_in_the_bag * num_features)
# - lower_threshold ... np.ndarray (num_bags * num_classes)
# - upper_threshold ... np.ndarray (num_bags * num_classes)
#
# bags[i_bag] contains not less than lower_thrshold[i_bag, i_class]
# i_class instances.
# run multiple instance learning
clf_mil, y_mil = learner.fit(
bags,
lower_threshold,
upper_threshold,
n_classes,
max_iter=10)
# after multiple instance learning,
# you can predict instance class
clf_mil.predict([instance_feature])
```
See `tests/test_countbased.py` for an example of a fully working test data generation process.
## License
milwrap is available under the MIT License.
Raw data
{
"_id": null,
"home_page": "https://github.com/inoueakimitsu/milwrap",
"name": "milwrap",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": null,
"keywords": "machine learning",
"author": "Akimitsu Inoue",
"author_email": "akimitsu.inoue@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/4c/fb/3dd6a131b3efd470e46775e9861dff2fada5db3430ebb091bafcac00dce0/milwrap-0.2.0.tar.gz",
"platform": null,
"description": "# milwrap\n\n[](https://github.com/inoueakimitsu/milwrap/actions/workflows/python-tests.yml)\n<a href=\"https://github.com/inoueakimitsu/milwrap/issues\"><img alt=\"GitHub issues\" src=\"https://img.shields.io/github/issues/inoueakimitsu/milwrap\"></a>\n[](https://colab.research.google.com/github/inoueakimitsu/milwrap/blob/master/introduction.ipynb)\n\nPython package for multiple instance learning (MIL).\nThis wraps single instance learning algorithms so that they can be fitted to data for MIL.\n\n## Features\n\n- support count-based multiple instance assumptions (see [wikipedia](https://en.wikipedia.org/wiki/Multiple_instance_learning#:~:text=Presence-%2C%20threshold-%2C%20and%20count-based%20assumptions%5Bedit%5D))\n- support multi-class setting\n- support scikit-learn algorithms (such as `RandomForestClassifier`, `SVC`, `LogisticRegression`)\n\n## Installation\n\n```bash\npip install milwrap\n```\n\n## Usage\n\nFor more information, see [Use scikit-learn models in multiple instance learning based on the count-based assumption](https://github.com/inoueakimitsu/milwrap/blob/main/intro-en.md).\n\n```python\n# Prepare single-instance supervised-learning algorithm\n# Note: only supports models with predict_proba() method.\nfrom sklearn.linear_model import LogisticRegression\nclf = LogisticRegression()\n\n# Wrap it with MilCountBasedMultiClassLearner\nfrom milwrap import MilCountBasedMultiClassLearner\nmil_learner = MilCountBasedMultiClassLearner(clf)\n\n# Prepare follwing dataset\n#\n# - bags ... list of np.ndarray\n# (num_instance_in_the_bag * num_features)\n# - lower_threshold ... np.ndarray (num_bags * num_classes)\n# - upper_threshold ... np.ndarray (num_bags * num_classes)\n#\n# bags[i_bag] contains not less than lower_thrshold[i_bag, i_class]\n# i_class instances.\n\n# run multiple instance learning\nclf_mil, y_mil = learner.fit(\n bags,\n lower_threshold,\n upper_threshold,\n n_classes,\n max_iter=10)\n\n# after multiple instance learning,\n# you can predict instance class\nclf_mil.predict([instance_feature])\n```\n\nSee `tests/test_countbased.py` for an example of a fully working test data generation process.\n\n## License\n\nmilwrap is available under the MIT License.\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "milwrap - multiple instane meta-learner that can use any supervised-learning algorithms.",
"version": "0.2.0",
"project_urls": {
"Homepage": "https://github.com/inoueakimitsu/milwrap",
"Repository": "https://github.com/inoueakimitsu/milwrap"
},
"split_keywords": [
"machine",
"learning"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "648685a943b1b599b72aaf4cefa28a861d9c559e61584e6c52b7f1a53fecbcd8",
"md5": "bcda1a556ba36160f07e95152494c061",
"sha256": "e31bb8edeaff7ad1a3f8f781a17f378f9643ba96738e01ea61057c6c93ee3b24"
},
"downloads": -1,
"filename": "milwrap-0.2.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "bcda1a556ba36160f07e95152494c061",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 6063,
"upload_time": "2025-01-24T06:34:51",
"upload_time_iso_8601": "2025-01-24T06:34:51.984244Z",
"url": "https://files.pythonhosted.org/packages/64/86/85a943b1b599b72aaf4cefa28a861d9c559e61584e6c52b7f1a53fecbcd8/milwrap-0.2.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "4cfb3dd6a131b3efd470e46775e9861dff2fada5db3430ebb091bafcac00dce0",
"md5": "6b19213212faa353591fb1b0e0477c19",
"sha256": "6a5eade323f4f705ad6d61d1f9c688581a5d4bb54b82af1f9ce745a69b823757"
},
"downloads": -1,
"filename": "milwrap-0.2.0.tar.gz",
"has_sig": false,
"md5_digest": "6b19213212faa353591fb1b0e0477c19",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 5085,
"upload_time": "2025-01-24T06:34:53",
"upload_time_iso_8601": "2025-01-24T06:34:53.200328Z",
"url": "https://files.pythonhosted.org/packages/4c/fb/3dd6a131b3efd470e46775e9861dff2fada5db3430ebb091bafcac00dce0/milwrap-0.2.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-01-24 06:34:53",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "inoueakimitsu",
"github_project": "milwrap",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"requirements": [
{
"name": "joblib",
"specs": [
[
"==",
"1.4.2"
]
]
},
{
"name": "numpy",
"specs": [
[
"==",
"2.2.2"
]
]
},
{
"name": "pandas",
"specs": [
[
"==",
"2.2.3"
]
]
},
{
"name": "python-dateutil",
"specs": [
[
"==",
"2.9.0.post0"
]
]
},
{
"name": "pytz",
"specs": [
[
"==",
"2024.2"
]
]
},
{
"name": "scikit-learn",
"specs": [
[
"==",
"1.6.1"
]
]
},
{
"name": "scipy",
"specs": [
[
"==",
"1.15.1"
]
]
},
{
"name": "six",
"specs": [
[
"==",
"1.17.0"
]
]
},
{
"name": "threadpoolctl",
"specs": [
[
"==",
"3.5.0"
]
]
},
{
"name": "tzdata",
"specs": [
[
"==",
"2025.1"
]
]
}
],
"lcname": "milwrap"
}
JSON
