https://github.com/lucidfrontier45/anybayes
A Bayesian Classifier with Any Distributions
https://github.com/lucidfrontier45/anybayes
bayes machine-learning naive-bayes-classifier python
Last synced: 4 months ago
JSON representation
A Bayesian Classifier with Any Distributions
- Host: GitHub
- URL: https://github.com/lucidfrontier45/anybayes
- Owner: lucidfrontier45
- License: mit
- Created: 2023-10-03T23:29:16.000Z (over 1 year ago)
- Default Branch: develop
- Last Pushed: 2023-10-04T13:28:33.000Z (over 1 year ago)
- Last Synced: 2024-12-31T23:12:04.615Z (6 months ago)
- Topics: bayes, machine-learning, naive-bayes-classifier, python
- Language: Python
- Homepage:
- Size: 37.1 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# AnyBayes
A Bayesian Classifier with Any Distribution
# Install
Please first install PDM >= 2.0 with pip/pipx.
```bash
pdm install --prod
```
# Develop
```bash
pdm install
```
# VSCode Settings
```bash
cp vscode_templates .vscode
```
Then install/activate all extensions listed in `.vscode/extensions.json`
# Usage
## API
```py
class AnyBayesClassifier(BaseEstimator, ClassifierMixin):
"""
A Bayesian classifier that can use any distribution for the class-conditional densities.
Parameters
----------
distribution_factory : Callable[[], Distribution]
A callable that returns an instance of a `Distribution` class. This is used to create the density estimators
for each class.
Attributes
----------
distribution_factory : Callable[[], Distribution]
The callable that returns an instance of a `Distribution` class.
dists_ : list[Distribution]
A list of density estimators, one for each class.
n_classes_ : int
The number of classes in the training data.
Methods
-------
fit(X, y)
Fit the classifier to the training data.
predict_proba(X, class_weight=1.0)
Predict the class probabilities for the given test data.
predict(X, class_weight=1.0)
Predict the class labels for the given test data.
"""
def __init__(self, distribution_factory: Callable[[], Distribution]):
...
def fit(
self, X: NDArray[Shape["N, D"], Number], y: NDArray[Shape["N"], Int]
) -> Self:
"""
Fit the classifier to the training data.
Parameters
----------
X : numpy.ndarray, shape (N, D)
The training data, where N is the number of samples and D is the number of features.
y : numpy.ndarray, shape (N,)
The class labels for each sample. The labels are assumed to start from 0 without gaps
Returns
-------
self : AnyBayesClassifier
The fitted `AnyBayesClassifier` instance.
"""
...
def predict_proba(
self,
X: NDArray[Shape["N, D"], Number],
class_weight: list[float] | float = 1.0,
) -> NDArray[Shape["N, C"], Float]:
"""
Predict the class probabilities for the given test data.
Parameters
----------
X : numpy.ndarray, shape (N, D)
The test data, where N is the number of samples and D is the number of features.
class_weight : float or list[float], optional
The weight of each class. If a float is given, it is used as the weight for all classes.
If a list is given, it must have the same length as the number of classes.
Returns
-------
probs : numpy.ndarray, shape (N, C)
The predicted class probabilities for each sample, where C is the number of classes.
"""
...
def predict(
self,
X: NDArray[Shape["N, D"], Number],
class_weight: list[float] | float = 1.0,
) -> NDArray[Shape["N"], Int]:
"""
Predict the class labels for the given test data.
Parameters
----------
X : numpy.ndarray, shape (N, D)
The test data, where N is the number of samples and D is the number of features.
class_weight : float or list[float], optional
The weight of each class. If a float is given, it is used as the weight for all classes.
If a list is given, it must have the same length as the number of classes.
Returns
-------
labels : numpy.ndarray, shape (N,)
The predicted class labels for each sample.
"""
...
```
## Example
Check notebooks in the `examples` directory.
## Implement Custom Backend Distribution
This package currently only includes empirical distribution backed by scikit-learn's KDE. If you want to use other distributions you need to add custom wrapper class that implements `Distribution` abstract class. For more detail, please check `src/anybayes/backends/kde.py` to understand how it is implemented.