| Name | NBCS JSON |
| Version |
1.4
JSON |
| download |
| home_page | None |
| Summary | Python package that implements an adaptive classification and regression system using simplicial complexes. The package provides a novel approach to handling both binary and multi-class classification problems, as well as regression tasks, by adaptively creating a simplicial decomposition of the feature space. |
| upload_time | 2024-12-11 08:04:42 |
| maintainer | None |
| docs_url | None |
| author | Amit Ronen |
| requires_python | >=3.6 |
| license | None |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# NBCS (Nested barycentric coordinate system)
**Version:** 1.2
**License:** MIT License
**Author:** Amit Ronen
**Email:** erankfmn@gmail.com
## Overview
NBCS (Non-Binary Classification System) is a Python package that implements an adaptive classification and regression system using simplicial complexes. The package provides a novel approach to handling both binary and multi-class classification problems, as well as regression tasks, by adaptively creating a simplicial decomposition of the feature space.
The package extends scikit-learn's base estimator and transformer interfaces, making it compatible with scikit-learn's ecosystem while providing unique capabilities for handling complex classification boundaries and non-linear regression problems.
## Features
- **Adaptive Classification:** Automatically adjusts the complexity of the decision boundary based on the data
- **Simplicial Complex Construction:** Creates an optimal decomposition of the feature space
- **Multiple Learning Modes:**
- Barry fitting mode for general purpose learning
- Adaptive fitting mode for classification tasks
- **Regression Support:** Handles complex regression problems with piece-wise linear approximations
- **Visualization Tools:** Built-in functions for plotting decision boundaries and regression results
- **Scikit-learn Compatible:** Implements scikit-learn's BaseEstimator and TransformerMixin interfaces
## Installation
You can install the package directly from PyPI using pip:
```bash
pip install nbcs
```
## Usage
### Classification Example
```python
import numpy as np
from nbcs import NBCS
from sklearn.svm import SVC
# Create and fit the NBCS model
model = NBCS(C=1000, k=4)
embedder = model.fit(X, y)
# Transform the data
points = embedder.transform(X)
# Use with any sklearn classifier
clf = SVC(kernel='linear')
clf.fit(points, y)
```
### Regression Example
```python
from nbcs import NBCS
from sklearn.linear_model import LinearRegression
# Create and fit the NBCS model
model = NBCS(k=3)
embedder = model.fit_barry(X)
# Transform the data
points = embedder.transform(X)
# Use with any sklearn regressor
reg = LinearRegression()
reg.fit(points, y)
```
## Parameters
- **C:** Regularization parameter (default=1)
- **k:** Number of refinement steps (default=1)
## Available Methods
- `fit(X, y)`: Fits the model using adaptive mode
- `fit_barry(X, y)`: Fits the model using barry mode
- `add_point(point)`: Adds a new point to the simplicial complex
## Visualization
The package includes several visualization tools:
```python
from nbcs.utils import make_meshgrid, plot_contours
# Create mesh grid for visualization
xx, yy = make_meshgrid(X[:, 0], X[:, 1])
# Plot decision boundaries
plot_contours(ax, clf, xx, yy, xy)
```
## Dependencies
- numpy
- scipy
- scikit-learn
- matplotlib
## Examples
### Classification with Regular Polytope
```python
import numpy as np
from nbcs import NBCS
from sklearn.svm import SVC
# Generate synthetic data
N = 1000
D = 2 # dimension
X = np.random.normal(0, 1, (N, D))
y = np.ones(N)
# Create and fit model
model = NBCS(C=1000, k=4)
embedder = model.fit(X, y)
```
### Regression with Multiple Lines
```python
import numpy as np
from nbcs import NBCS
from sklearn.linear_model import LinearRegression
# Generate synthetic data
x = np.linspace(0, 30, 60).reshape(-1, 1)
y = np.sin(x/10) + np.random.normal(0, 0.1, x.shape)
# Create and fit model
model = NBCS(k=3)
embedder = model.fit_barry(x)
```
## License
This project is licensed under the MIT License - see the LICENSE file for details.
## Acknowledgements
- scikit-learn for providing the foundational tools for machine learning in Python
- numpy and scipy for numerical computations
- matplotlib for data visualization
## Contact
If you have any questions or suggestions, feel free to reach out at erankfmn@gmail.com
Raw data
{
"_id": null,
"home_page": null,
"name": "NBCS",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.6",
"maintainer_email": null,
"keywords": null,
"author": "Amit Ronen",
"author_email": "erankfmn@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/e2/85/4b75059597b9ad55bbed1d21238a06b69fca8c0713d4e0305e7a1ffe8bfe/NBCS-1.4.tar.gz",
"platform": null,
"description": "# NBCS (Nested barycentric coordinate system)\n\n**Version:** 1.2 \n**License:** MIT License \n**Author:** Amit Ronen\n**Email:** erankfmn@gmail.com \n\n## Overview\n\nNBCS (Non-Binary Classification System) is a Python package that implements an adaptive classification and regression system using simplicial complexes. The package provides a novel approach to handling both binary and multi-class classification problems, as well as regression tasks, by adaptively creating a simplicial decomposition of the feature space.\n\nThe package extends scikit-learn's base estimator and transformer interfaces, making it compatible with scikit-learn's ecosystem while providing unique capabilities for handling complex classification boundaries and non-linear regression problems.\n\n## Features\n\n- **Adaptive Classification:** Automatically adjusts the complexity of the decision boundary based on the data\n- **Simplicial Complex Construction:** Creates an optimal decomposition of the feature space\n- **Multiple Learning Modes:**\n - Barry fitting mode for general purpose learning\n - Adaptive fitting mode for classification tasks\n- **Regression Support:** Handles complex regression problems with piece-wise linear approximations\n- **Visualization Tools:** Built-in functions for plotting decision boundaries and regression results\n- **Scikit-learn Compatible:** Implements scikit-learn's BaseEstimator and TransformerMixin interfaces\n\n## Installation\n\nYou can install the package directly from PyPI using pip:\n\n```bash\npip install nbcs\n```\n\n## Usage\n\n### Classification Example\n\n```python\nimport numpy as np\nfrom nbcs import NBCS\nfrom sklearn.svm import SVC\n\n# Create and fit the NBCS model\nmodel = NBCS(C=1000, k=4)\nembedder = model.fit(X, y)\n\n# Transform the data\npoints = embedder.transform(X)\n\n# Use with any sklearn classifier\nclf = SVC(kernel='linear')\nclf.fit(points, y)\n```\n\n### Regression Example\n\n```python\nfrom nbcs import NBCS\nfrom sklearn.linear_model import LinearRegression\n\n# Create and fit the NBCS model\nmodel = NBCS(k=3)\nembedder = model.fit_barry(X)\n\n# Transform the data\npoints = embedder.transform(X)\n\n# Use with any sklearn regressor\nreg = LinearRegression()\nreg.fit(points, y)\n```\n\n## Parameters\n\n- **C:** Regularization parameter (default=1)\n- **k:** Number of refinement steps (default=1)\n\n## Available Methods\n\n- `fit(X, y)`: Fits the model using adaptive mode\n- `fit_barry(X, y)`: Fits the model using barry mode\n- `add_point(point)`: Adds a new point to the simplicial complex\n\n\n## Visualization\n\nThe package includes several visualization tools:\n\n```python\nfrom nbcs.utils import make_meshgrid, plot_contours\n\n# Create mesh grid for visualization\nxx, yy = make_meshgrid(X[:, 0], X[:, 1])\n\n# Plot decision boundaries\nplot_contours(ax, clf, xx, yy, xy)\n```\n\n## Dependencies\n\n- numpy\n- scipy\n- scikit-learn\n- matplotlib\n\n## Examples\n\n### Classification with Regular Polytope\n\n```python\nimport numpy as np\nfrom nbcs import NBCS\nfrom sklearn.svm import SVC\n\n# Generate synthetic data\nN = 1000\nD = 2 # dimension\nX = np.random.normal(0, 1, (N, D))\ny = np.ones(N)\n\n# Create and fit model\nmodel = NBCS(C=1000, k=4)\nembedder = model.fit(X, y)\n```\n\n### Regression with Multiple Lines\n\n```python\nimport numpy as np\nfrom nbcs import NBCS\nfrom sklearn.linear_model import LinearRegression\n\n# Generate synthetic data\nx = np.linspace(0, 30, 60).reshape(-1, 1)\ny = np.sin(x/10) + np.random.normal(0, 0.1, x.shape)\n\n# Create and fit model\nmodel = NBCS(k=3)\nembedder = model.fit_barry(x)\n```\n\n## License\n\nThis project is licensed under the MIT License - see the LICENSE file for details.\n\n## Acknowledgements\n\n- scikit-learn for providing the foundational tools for machine learning in Python\n- numpy and scipy for numerical computations\n- matplotlib for data visualization\n\n## Contact\n\nIf you have any questions or suggestions, feel free to reach out at erankfmn@gmail.com\n",
"bugtrack_url": null,
"license": null,
"summary": "Python package that implements an adaptive classification and regression system using simplicial complexes. The package provides a novel approach to handling both binary and multi-class classification problems, as well as regression tasks, by adaptively creating a simplicial decomposition of the feature space.",
"version": "1.4",
"project_urls": null,
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "4126e75d397c3ac60f2420e31a98aed811b129bec118b7a67230eab760db900a",
"md5": "3bb94ab38e19caeae32978650916df86",
"sha256": "7e8a69c4a310ea766152ba22a554edcf02300c31af73831c58590dbc504c7ffe"
},
"downloads": -1,
"filename": "NBCS-1.4-py3-none-any.whl",
"has_sig": false,
"md5_digest": "3bb94ab38e19caeae32978650916df86",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.6",
"size": 4897,
"upload_time": "2024-12-11T08:04:39",
"upload_time_iso_8601": "2024-12-11T08:04:39.801314Z",
"url": "https://files.pythonhosted.org/packages/41/26/e75d397c3ac60f2420e31a98aed811b129bec118b7a67230eab760db900a/NBCS-1.4-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "e2854b75059597b9ad55bbed1d21238a06b69fca8c0713d4e0305e7a1ffe8bfe",
"md5": "b691df58e65457efd6c4bb74af32e6d1",
"sha256": "2312d6fa8ec03a2c20e5d9ad4a040254a066e35be263dd5e3cf406a0cc2da7c7"
},
"downloads": -1,
"filename": "NBCS-1.4.tar.gz",
"has_sig": false,
"md5_digest": "b691df58e65457efd6c4bb74af32e6d1",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.6",
"size": 4768,
"upload_time": "2024-12-11T08:04:42",
"upload_time_iso_8601": "2024-12-11T08:04:42.474827Z",
"url": "https://files.pythonhosted.org/packages/e2/85/4b75059597b9ad55bbed1d21238a06b69fca8c0713d4e0305e7a1ffe8bfe/NBCS-1.4.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-11 08:04:42",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "nbcs"
}
