This project aims to predict seizures in epilepsy patients before they occur using machine learning techniques. By leveraging ensemble learning methods that combine multiple classification models, we enhance the accuracy and reliability of predictions, providing valuable insights for patients and healthcare professionals.
- Predictive Modeling: Utilizes advanced machine learning algorithms to analyze patient data and forecast potential seizures.
- Ensemble Learning: Implements a Voting Classifier that combines various classifiers, such as Logistic Regression, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Random Forest, and Gradient Boosting, to improve predictive performance.
- Data Preprocessing: Incorporates techniques such as scaling and feature selection to ensure optimal model training.
- User Interface: A user-friendly interface is built using Gradio to allow users to input data and receive real-time seizure predictions.
- Data Collection: The model is trained on a dataset containing relevant features such as heart rate, blood pressure, and seizure history.
- Model Training: The ensemble model is trained using cross-validation to ensure robustness and minimize overfitting.
- Prediction: Users can input their current physiological data, and the model predicts whether a seizure is likely to occur.
To run this project locally, follow these steps:
- Clone the repository:
git clone https://github.com/jnvw/pre_seizure_prediction.git