https://github.com/caselles/classic_classifiers
Comparison of 5 predictive algorithms on a imbalanced 0-1 dataset.
https://github.com/caselles/classic_classifiers
Last synced: over 1 year ago
JSON representation
Comparison of 5 predictive algorithms on a imbalanced 0-1 dataset.
- Host: GitHub
- URL: https://github.com/caselles/classic_classifiers
- Owner: Caselles
- Created: 2017-01-09T12:18:40.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2017-02-12T09:54:19.000Z (over 9 years ago)
- Last Synced: 2025-02-04T17:55:09.765Z (over 1 year ago)
- Language: Jupyter Notebook
- Size: 435 KB
- Stars: 2
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Classic classifiers
Author : Hugo Caselles-Dupré
Disclaimer : several syntax errors can exist in the code. Use it wisely.
---------------------------------------
Comparison of 5 predictive algorithms on a imbalanced 0-1 dataset.
---------------------------------------
In this directory you can find a Jupyter Notebook with the study of a binary classification problem on this dataset : https://archive.ics.uci.edu/ml/datasets/Wilt
For this, I compared the results obtained with the following algorithms :
- k-Nearest Neighbors (scikit-learn)
- Support Vector Machine (scikit-learn)
- Random Forest (scikit-learn)
- 2 hidden layers fully-connected Neural Network (Keras with TensorFlow backend)
- Gradient Boosting (XGBoost)
The Notebook provides results in terms of accuracy and ROC curves (AUC), and a discussion about the results (in French, for a course in ENSAE ParisTech).
------------------------------------------
If you have any questions : casellesdupre.hugo@gmail.com.