CLASSIFICATION OF FAMILIAL AND SPORADIC NEUROGENETIC DISEASE CASES BY USING CLINICAL SYMPTOMS AND MACHINE LEARNING


Değirmenci Akdeniz M.

10th International Conference on Engineering Technology and Innovation (ICETI 2026) , Sarajevo, Bosna-Hersek, 14 - 18 Ekim 2026, ss.11, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Sarajevo
  • Basıldığı Ülke: Bosna-Hersek
  • Sayfa Sayıları: ss.11
  • Kütahya Sağlık Bilimleri Üniversitesi Adresli: Evet

Özet

Neurofibromatosis type 1 (NF1) is a complex neurogenetic disorder characterized by genetic and neurological features, with clinical manifestations and disease progression that can vary significantly among individuals. This variability in clinical symptoms can complicate the accurate and early detection of the disease and the appropriate classification of patients. This study evaluates the performance of various machine learning methods for detecting NF1 and classifying patients into familial and sporadic cases based on clinical findings. A total of 11 different machine learning algorithms were employed, and their classification performances were compared. Principal Component Analysis (PCA) was applied for the feature selection stage to facilitate more effective evaluation of dataset variables and to achieve dimensionality reduction. The clinical data used in the study were obtained from the UCI Machine Learning Repository. The results indicate that, for the task of detecting NF1, the highest accuracy rate (75.7%) was achieved using Quadratic Discriminant Analysis and Naive Bayes methods on the PCAprocessed data. Regarding the classification of familial versus sporadic cases, the Naive Bayes method yielded the highest accuracy rate at 63.5%. These findings demonstrate that machine learning methods represent potential approaches for detecting NF1 based on clinical findings and for classifying cases.