|
LE GOURRIEREC Titouan
Report a bug · Request Feature
Table of Contents
This project focuses on classifying chest X-rays as either healthy or pneumonia-affected. Several machine learning algorithms were tested, including K-Nearest Neighbors (KNN), Logistic Regression, Decision Trees, Support Vector Classifier (SVC), Random Forest, Gradient Boosting Classifier, and a Voting Classifier.
To address the issue of imbalanced classes, various techniques were employed, such as Random Over Sampling, SMOTE (Synthetic Minority Over-sampling Technique), Random Under Sampling, and Weight Modification.
For each algorithm, Grid Search was used to fine-tune the hyperparameters. The final model was evaluated using cross-validation, and learning curves were analyzed to assess whether further evaluation was needed.
The final model chosen is a soft voting classifier, which combines the outputs of the different models, configured as follows:
Below are the evaluation metrics for this model:
You can find the project report here: report.pdf
Before running this project, make sure you have installed the necessary dependencies. You can do this by installing the packages listed in the requirements.txt file:
pip install -r requirements.txtEnsure you have Python installed and that you're using a virtual environment if needed.
-> For the data, please follow the instructions in the Dataset_Link.pdf file.
To use this project, just change the variable path in the Images Import part of the file project.ipynb
Distributed under the MIT License. See LICENSE for more information.
LE GOURRIEREC Titouan - titouanlegourrierec@icloud.com
Project Link: https://github.com/titouanlegourrierec/PneumoniaXRayClassification