# PrevhClassifier
This package implements the Prevh classification algorithm.
> The algorithm is based in the follow [research](https://zenodo.org/record/6090322#.Yj98bKbMKUk) **Pages 71-76**.
> [Package Documentation](https://pypi.org/project/prevhlib/).
> [](https://github.com/JCGCosta/Prevh/actions/workflows/python-publish.yml)
# User Guide
> This package can be installed with the following command: **pip install prevhlib**
## Python example:
```python
import numpy as np
import pandas as pd
from __init__.PrevhClassifier import PrevhClassifier
from sklearn.preprocessing import StandardScaler, LabelEncoder
if __name__ == '__main__':
iris = pd.read_csv('Datasets/iris.csv')
X = iris.iloc[:, 0:4].values
y = iris.iloc[:, 4].values
header = (iris.columns[:-1], iris.columns[-1])
prevh = PrevhClassifier(distance_algorithm="euclidean")
prevh.fit(X, y, header=header, encoder=LabelEncoder(), scaler=StandardScaler())
print(prevh)
# Outputs:
# {
# "dataset": {
# "header": {
# "features": "Index(['sepal length', 'sepal width', 'petal length', 'petal width'], dtype='object')",
# "classes": "class"
# },
# "encoder": "LabelEncoder()",
# "scaler": "StandardScaler()"
# },
# "distance": "euclidean"
# }
print(prevh.classify(np.array([5.1, 3.5, 1.4, 0.2]), K=3))
# Outputs: (array(['Iris-setosa'], dtype=object), np.float64(0.2653212465045153))
kfold_split_arguments = {
"n_splits": 5,
"random_state": 42,
"shuffle": True
}
Evaluation_Results = prevh.evaluate(1, "kfold_cross_validation", kfold_split_arguments)
print(Evaluation_Results.get_metrics())
# Outputs:
# accuracy precision recall f1-score
# 0 0.966667 0.972222 0.962963 0.965899
# 1 0.966667 0.969697 0.952381 0.958486
# 2 0.966667 0.962963 0.966667 0.962848
# 3 0.900000 0.911681 0.905556 0.907368
# 4 0.966667 0.972222 0.972222 0.971014
# Mean 0.953333 0.957757 0.951958 0.953123
Evaluation_Results.plot_confusion_matrices()
```
<img src="https://raw.githubusercontent.com/JCGCosta/Prevh/refs/heads/main/confusion_matrix_example.png" width = "600">
## Next Steps
- In the next steps I will add support to other split, evaluation, encoder, and decoder methods.
- I expect to in new versions to comparisons between other machine learn method and the prevh classifier.
Change Log
===============
0.1.6 (27/09/2024)
------------------
MAJOR CHANGES
- Update in the README.md
- Update the required libraries for security reasons
- Change code structure to be more readable
- Added the Evaluator object, and metric plots
------------------
Raw data
{
"_id": null,
"home_page": "https://github.com/JCGCosta/Prevh",
"name": "prevhlib",
"maintainer": null,
"docs_url": null,
"requires_python": null,
"maintainer_email": null,
"keywords": "DataMining",
"author": "J\u00falio C\u00e9sar Guimar\u00e3es Costa",
"author_email": "juliocesargcosta123@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/1b/e2/1a34e852491c4db6e1877b507d4f5e89a33a9281bb1db1cdb3fa9c2ca39a/prevhlib-0.1.6.tar.gz",
"platform": null,
"description": "# PrevhClassifier\nThis package implements the Prevh classification algorithm.\n> The algorithm is based in the follow [research](https://zenodo.org/record/6090322#.Yj98bKbMKUk) **Pages 71-76**.\n \n> [Package Documentation](https://pypi.org/project/prevhlib/).\n\n> [](https://github.com/JCGCosta/Prevh/actions/workflows/python-publish.yml)\n\n# User Guide\n\n> This package can be installed with the following command: **pip install prevhlib**\n\n## Python example:\n\n```python\nimport numpy as np\nimport pandas as pd\nfrom __init__.PrevhClassifier import PrevhClassifier\nfrom sklearn.preprocessing import StandardScaler, LabelEncoder\n\nif __name__ == '__main__':\n iris = pd.read_csv('Datasets/iris.csv')\n\n X = iris.iloc[:, 0:4].values\n y = iris.iloc[:, 4].values\n header = (iris.columns[:-1], iris.columns[-1])\n\n prevh = PrevhClassifier(distance_algorithm=\"euclidean\")\n prevh.fit(X, y, header=header, encoder=LabelEncoder(), scaler=StandardScaler())\n\n print(prevh)\n # Outputs: \n # { \n # \"dataset\": {\n # \"header\": {\n # \"features\": \"Index(['sepal length', 'sepal width', 'petal length', 'petal width'], dtype='object')\",\n # \"classes\": \"class\"\n # },\n # \"encoder\": \"LabelEncoder()\",\n # \"scaler\": \"StandardScaler()\"\n # },\n # \"distance\": \"euclidean\"\n # }\n\n print(prevh.classify(np.array([5.1, 3.5, 1.4, 0.2]), K=3))\n # Outputs: (array(['Iris-setosa'], dtype=object), np.float64(0.2653212465045153))\n\n kfold_split_arguments = {\n \"n_splits\": 5,\n \"random_state\": 42,\n \"shuffle\": True\n }\n\n Evaluation_Results = prevh.evaluate(1, \"kfold_cross_validation\", kfold_split_arguments)\n\n print(Evaluation_Results.get_metrics())\n # Outputs:\n # accuracy precision recall f1-score\n # 0 0.966667 0.972222 0.962963 0.965899\n # 1 0.966667 0.969697 0.952381 0.958486\n # 2 0.966667 0.962963 0.966667 0.962848\n # 3 0.900000 0.911681 0.905556 0.907368\n # 4 0.966667 0.972222 0.972222 0.971014\n # Mean 0.953333 0.957757 0.951958 0.953123\n\n Evaluation_Results.plot_confusion_matrices()\n```\n\n<img src=\"https://raw.githubusercontent.com/JCGCosta/Prevh/refs/heads/main/confusion_matrix_example.png\" width = \"600\">\n\n## Next Steps\n\n- In the next steps I will add support to other split, evaluation, encoder, and decoder methods.\n- I expect to in new versions to comparisons between other machine learn method and the prevh classifier.\n\n\nChange Log\n===============\n0.1.6 (27/09/2024)\n------------------\nMAJOR CHANGES\n- Update in the README.md\n- Update the required libraries for security reasons\n- Change code structure to be more readable\n- Added the Evaluator object, and metric plots\n------------------\n\n\n",
"bugtrack_url": null,
"license": "GNU General Public License v3 (GPLv3)",
"summary": "Data Mining package implementing the PrevhClassifier algorithm.",
"version": "0.1.6",
"project_urls": {
"Homepage": "https://github.com/JCGCosta/Prevh"
},
"split_keywords": [
"datamining"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "63e8cd714ed6ca37b4ac7e91a106be7a7f8b20754a0411e103f504a69f2a0d45",
"md5": "0a4ccee51a924f314532e1b83f73616c",
"sha256": "ca8103e2a9d1e08b64bdf81b7ad4e3f6d3d63ffd633a581ff02d6d45ad4e9f81"
},
"downloads": -1,
"filename": "prevhlib-0.1.6-py3-none-any.whl",
"has_sig": false,
"md5_digest": "0a4ccee51a924f314532e1b83f73616c",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 22662,
"upload_time": "2024-09-27T16:08:20",
"upload_time_iso_8601": "2024-09-27T16:08:20.882166Z",
"url": "https://files.pythonhosted.org/packages/63/e8/cd714ed6ca37b4ac7e91a106be7a7f8b20754a0411e103f504a69f2a0d45/prevhlib-0.1.6-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "1be21a34e852491c4db6e1877b507d4f5e89a33a9281bb1db1cdb3fa9c2ca39a",
"md5": "f79802e2d1496ecba5249272d26ee7a0",
"sha256": "c5f933500988035c5066d08ac9eb52eaf034223db3a0b6bcd97bf34b5c5fb39c"
},
"downloads": -1,
"filename": "prevhlib-0.1.6.tar.gz",
"has_sig": false,
"md5_digest": "f79802e2d1496ecba5249272d26ee7a0",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 21863,
"upload_time": "2024-09-27T16:08:22",
"upload_time_iso_8601": "2024-09-27T16:08:22.270171Z",
"url": "https://files.pythonhosted.org/packages/1b/e2/1a34e852491c4db6e1877b507d4f5e89a33a9281bb1db1cdb3fa9c2ca39a/prevhlib-0.1.6.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-09-27 16:08:22",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "JCGCosta",
"github_project": "Prevh",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"requirements": [],
"lcname": "prevhlib"
}
JSON
