Deep Learning Model Highly Accurate at Classifying Migraine Types
-
Machine learning models like SVM, RF, KNN, Decision Tree, and Deep Neural Networks were implemented to classify migraine into 7 types using a publicly available dataset. Data augmentation using SMOTE was applied to balance the dataset.
-
The Deep Neural Network model with 2 hidden layers and 512 neurons achieved the best accuracy of 99.66% after data augmentation, outperforming other ML models.
-
The study demonstrates the potential of using machine learning for accurate prediction and classification of migraine to assist diagnosis in countries lacking medical resources.
-
Data augmentation significantly improved model performance, indicating the need for larger datasets to train sophisticated deep learning models.
-
Future work involves collecting more migraine data and exploring advanced algorithms like BERT to further improve classification accuracy.