Development of a Machine Learning Based Clinical Decision Support System for Classification of Migraine Types: A Preliminary Study
International Journal of Advanced Natural Sciences and Engineering Researches, cilt.2, sa.8, ss.323-332, 2024 (Hakemli Dergi)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 2 Sayı: 8
- Basım Tarihi: 2024
- Doi Numarası: 10.5281/zenodo.14194306
- Dergi Adı: International Journal of Advanced Natural Sciences and Engineering Researches
- Derginin Tarandığı İndeksler: Index Copernicus, Root Indexing
- Sayfa Sayıları: ss.323-332
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- İnönü Üniversitesi Adresli: Evet
Özet
Migraine is a type of neurological headache that seriously affects daily life and is associated with different symptoms. Early diagnosis of migraine disease is important for the start of the treatment process. In this process, specialized physicians are always needed, but artificial intelligence-based clinical systems can save time in the diagnosis of migraine and other headache types and can help determine the right treatment methods by providing support to general practitioners. In this study, the classification of migraine typical with aura and migraine without aura, which are the most common types of migraine, and other types of migraine were performed. In the classification process, data from demographic and clinical questionnaires were used and five different machine learning models were applied. In this research, the Rotation Forest algorithm showed the most successful performance according to the classifier evaluation criteria. As a result of this algorithm, accuracy (95.14%), true positive (95.10%), false positive (2.40%), kappa statistics (92.71%) and mean absolute error (6.50%) rates were obtained.