# BANSHEE
Bayesian Networks (BNs) are probabilistic, graphical models for representing complex dependency structures. They have many applications in science and engineering. Their particularly powerful variant – Non-Parametric BNs – are implemented as a Matlab toolbox and an open-access scriptable code, in the form of a Python-based package.
This repository contains:
* PyBanshee a Python-based open source of the MATLAB toolbox.
The supporting SoftwareX paper for PyBanhsee, _PyBanshee version (1.0): A Python implementation of the MATLAB toolbox BANSHEE for Non-Parametric Bayesian Networks with updated features_, can be found at https://doi.org/10.1016/j.softx.2022.101279
These codes are an update of the original version supporting SoftwareX paper: https://doi.org/10.1016/j.softx.2020.100588. The latest version of MATLAB BANSHEE (v1.3) supprting SoftwareX papper can be found at https://doi.org/10.1016/j.softx.2023.101479
The packages allows for quantifying the BN, validating the underlying assumptions of the model, visualizing the network and its corresponding rank correlation matrix, sampling and finally making inference with a BN based on existing or new evidence. MATLAB BANSHEE v1.3 and Py_BANSHEE have the same features.
# PyBanshee
PyBanshee is a Python-based open source of the MATLAB toolbox [BANSHEE](https://doi.org/10.1016/j.softx.2020.100588).
## Installation and updating
Use the package manager [pip](https://pip.pypa.io/en/stable/) to install last stable version of the package.
```bash
pip install py-banshee
```
However, if you are looking for the latest updates, consider installation directly from sources.
```
git clone https://github.com/mike-mendoza/py_banshee.git
cd py_banshee
python setup.py install
```
## Usage
Features:
* py_banshee.rankcorr.bn_rankcorr --> Create a Bayesian Network rank correlation matrix
* py_banshee.bn_plot.bn_visualize --> Visualize the structure of a defined Bayesian Network
* py_banshee.copula_test.cvm_statistic --> Goodness-of-fit test for copulas
* py_banshee.d_cal.gaussian_distance --> Measure the distance between Gaussian densities
* py_banshee.sample_bn.generate_samples --> Make samples using the defined Bayesian Network
* py_banshee.sample_bn.sample_base_conditioning --> Make samples based conditioning on intervals
* py_banshee.prediction.inference --> Make predictions using a non-parametric Bayesian Network
* py_banshee.prediction.conditional_margins_hist --> Visualize the un-conditional and conditional marginal histograms
#### Demo of some of the features:
```python
from py_banshee.rankcorr import bn_rankcorr
from py_banshee.bn_plot import bn_visualize
from py_banshee.prediction import inference,conditional_margins_hist
#Defining the variables of the BN
names = ['V1','V2','V3'] #names of the variables (nodes)
N = len(names) #number of nodes
#parametric distributions of the nodes
distributions = ['norm','genextreme','norm']
parameters = [[100,23],[-0.15,130,50],[500,100]]
#Defining the structure of the BN
ParentCell = [None]*N
ParentCell[0] = []
ParentCell[1] = [0]
ParentCell[2] = [0,1]
#Defining the rank correlation matrix
RankCorr = [None]*N
RankCorr[0] = []
RankCorr[1] = [.1]
RankCorr[2] = [.41,-.25]
#Conditional rank correlation matrix
R = bn_rankcorr(ParentCell,RankCorr,var_names=names,is_data=False, plot=True)
#Plot of the Bayesian Network
bn_visualize(ParentCell,R,names,fig_name='BN_TEST')
# Inference
condition_nodes = [0] #conditionalized variables (node V1)
condition_values = [181] #conditionalized value (node V1)
F = inference(Nodes = condition_nodes,
Values = condition_values,
R=R,
DATA=[],
SampleSize=100000,
empirical_data=False,
distributions=distributions,
parameters=parameters,
Output='full')
#Conditional and un-conditional histograms
conditional_margins_hist(F,[],names,condition_nodes,
empirical_data = False,
distributions=distributions,
parameters=parameters)
```
## Contributing
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
## License
[GNU](https://choosealicense.com/licenses/gpl-3.0/)
Raw data
{
"_id": null,
"home_page": "https://github.com/mike-mendoza/py_banshee",
"name": "py-banshee",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "pypi,py_banshee,tutorial",
"author": "Paul Koot, Miguel Angel Mendoza Lugo, Oswaldo Morales Napoles",
"author_email": "mendozalugo@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/6f/27/fc7d6850f23c311cd1c88c0dac3ce9ae2ad8144b35df49afb7dc42e22fa8/py_banshee-1.1.1.tar.gz",
"platform": null,
"description": "# BANSHEE\r\n\r\nBayesian Networks (BNs) are probabilistic, graphical models for representing complex dependency structures. They have many applications in science and engineering. Their particularly powerful variant \u2013 Non-Parametric BNs \u2013 are implemented as a Matlab toolbox and an open-access scriptable code, in the form of a Python-based package.\r\n\r\nThis repository contains:\r\n* PyBanshee a Python-based open source of the MATLAB toolbox.\r\n \r\nThe supporting SoftwareX paper for PyBanhsee, _PyBanshee version (1.0): A Python implementation of the MATLAB toolbox BANSHEE for Non-Parametric Bayesian Networks with updated features_, can be found at https://doi.org/10.1016/j.softx.2022.101279 \r\n\r\nThese codes are an update of the original version supporting SoftwareX paper: https://doi.org/10.1016/j.softx.2020.100588. The latest version of MATLAB BANSHEE (v1.3) supprting SoftwareX papper can be found at https://doi.org/10.1016/j.softx.2023.101479\r\n\r\nThe packages allows for quantifying the BN, validating the underlying assumptions of the model, visualizing the network and its corresponding rank correlation matrix, sampling and finally making inference with a BN based on existing or new evidence. MATLAB BANSHEE v1.3 and Py_BANSHEE have the same features.\r\n\r\n# PyBanshee\r\n\r\nPyBanshee is a Python-based open source of the MATLAB toolbox [BANSHEE](https://doi.org/10.1016/j.softx.2020.100588). \r\n\r\n## Installation and updating\r\nUse the package manager [pip](https://pip.pypa.io/en/stable/) to install last stable version of the package. \r\n\r\n```bash\r\npip install py-banshee\r\n```\r\nHowever, if you are looking for the latest updates, consider installation directly from sources.\r\n```\r\ngit clone https://github.com/mike-mendoza/py_banshee.git\r\ncd py_banshee\r\npython setup.py install\r\n```\r\n\r\n## Usage\r\nFeatures:\r\n* py_banshee.rankcorr.bn_rankcorr --> Create a Bayesian Network rank correlation matrix \r\n* py_banshee.bn_plot.bn_visualize --> Visualize the structure of a defined Bayesian Network\r\n* py_banshee.copula_test.cvm_statistic --> Goodness-of-fit test for copulas\r\n* py_banshee.d_cal.gaussian_distance --> Measure the distance between Gaussian densities\r\n* py_banshee.sample_bn.generate_samples --> Make samples using the defined Bayesian Network\r\n* py_banshee.sample_bn.sample_base_conditioning --> Make samples based conditioning on intervals\r\n* py_banshee.prediction.inference --> Make predictions using a non-parametric Bayesian Network\r\n* py_banshee.prediction.conditional_margins_hist --> Visualize the un-conditional and conditional marginal histograms\r\n\r\n#### Demo of some of the features:\r\n\r\n```python\r\nfrom py_banshee.rankcorr import bn_rankcorr\r\nfrom py_banshee.bn_plot import bn_visualize\r\nfrom py_banshee.prediction import inference,conditional_margins_hist\r\n\r\n#Defining the variables of the BN\r\nnames = ['V1','V2','V3'] #names of the variables (nodes)\r\nN = len(names) \t\t #number of nodes\r\n\r\n#parametric distributions of the nodes\r\ndistributions = ['norm','genextreme','norm']\t\r\nparameters = [[100,23],[-0.15,130,50],[500,100]]\r\n\r\n#Defining the structure of the BN\r\nParentCell = [None]*N\r\nParentCell[0] = []\r\nParentCell[1] = [0]\r\nParentCell[2] = [0,1]\r\n\r\n#Defining the rank correlation matrix\r\nRankCorr = [None]*N\r\nRankCorr[0] = []\r\nRankCorr[1] = [.1]\r\nRankCorr[2] = [.41,-.25]\r\n\r\n#Conditional rank correlation matrix\r\nR = bn_rankcorr(ParentCell,RankCorr,var_names=names,is_data=False, plot=True)\r\n\r\n#Plot of the Bayesian Network\r\nbn_visualize(ParentCell,R,names,fig_name='BN_TEST')\r\n\r\n# Inference\r\ncondition_nodes = [0] #conditionalized variables (node V1)\r\ncondition_values = [181] #conditionalized value (node V1)\r\n\r\nF = inference(Nodes = condition_nodes,\r\n Values = condition_values,\r\n R=R,\r\n DATA=[],\r\n SampleSize=100000,\r\n empirical_data=False, \r\n distributions=distributions,\r\n parameters=parameters,\r\n Output='full')\r\n\r\n#Conditional and un-conditional histograms \r\nconditional_margins_hist(F,[],names,condition_nodes,\r\n empirical_data = False,\r\n distributions=distributions,\r\n parameters=parameters)\r\n```\r\n## Contributing\r\nPull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.\r\n\r\n## License\r\n[GNU](https://choosealicense.com/licenses/gpl-3.0/)\r\n",
"bugtrack_url": null,
"license": "GNU",
"summary": "Py_Banshee allows for quantifying non-parametric Bayesian Networks",
"version": "1.1.1",
"project_urls": {
"Bug Tracker": "https://github.com/mike-mendoza/py_banshee/issues",
"Download": "https://github.com/mike-mendoza/py_banshee/archive/refs/tags/v1.1.1.tar.gz",
"Homepage": "https://github.com/mike-mendoza/py_banshee"
},
"split_keywords": [
"pypi",
"py_banshee",
"tutorial"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "6f27fc7d6850f23c311cd1c88c0dac3ce9ae2ad8144b35df49afb7dc42e22fa8",
"md5": "4471d7ea53160251ae28a78de7c1cee3",
"sha256": "c11e9812612480fff73e0c91cdfda554151a8bdac6ae63c446e2e29c17d1f616"
},
"downloads": -1,
"filename": "py_banshee-1.1.1.tar.gz",
"has_sig": false,
"md5_digest": "4471d7ea53160251ae28a78de7c1cee3",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 34346,
"upload_time": "2023-11-14T12:12:34",
"upload_time_iso_8601": "2023-11-14T12:12:34.327032Z",
"url": "https://files.pythonhosted.org/packages/6f/27/fc7d6850f23c311cd1c88c0dac3ce9ae2ad8144b35df49afb7dc42e22fa8/py_banshee-1.1.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-11-14 12:12:34",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "mike-mendoza",
"github_project": "py_banshee",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "py-banshee"
}
JSON
