neuRecommend is a project focused on developing a classifier for semi-automated spike-sorting and channel recommendation. The primary goals are to remove noise during preprocessing to improve clustering and to recommend electrodes with neurons for user convenience.
- Classifiers: Utilizes Random Forest and XGBoost for robust classification.
- Feature Engineering: Incorporates advanced techniques to enhance model performance.
- Dimensionality Reduction: Employs PCA and UMAP to reduce feature space while retaining essential information.
- Model Interpretability: Includes SHAP analysis for understanding model predictions.
- Hyperparameter Tuning: Offers grid search capabilities for optimizing model parameters.
- Utilizes Random Forest and XGBoost classifiers
- Feature engineering and dimensionality reduction
- Model interpretability
To get started with neuRecommend, follow these simple steps:
- Clone the repository.
- Install the necessary dependencies.
- Run the application.
git clone https://github.com/abuzarmahmood/neuRecommend.git
cd neuRecommend
pip install -r requirements.txt
python main.pyExamples of how to use the classifier.
Guidelines for contributing to the project.
Information on how to reach the project maintainers.