Résumé:
Migraine is a severe neurological condition characterized by intense headaches and accompanying
symptoms, impacting millions of people worldwide. Accurately categorizing migraine types is crucial
for effective diagnosis and treatment. Traditional classification methods often rely on subjective clin-
ical judgment, leading to inconsistencies. This thesis explores advanced machine learning and deep
learning techniques to improve the precision of migraine classification using a comprehensive dataset.
The research comprises two main approaches. The first approach utilizes ensemble learning algo-
rithms, particularly XGBoost, to classify migraine types based on diverse features. This method com-
bines descriptive, diagnostic, and predictive analyses to offer a comprehensive understanding of the
dataset. Extensive experimentation demonstrates that XGBoost performs better than traditional ma-
chine learning models in terms of accuracy and efficiency. The second approach introduces a unique
method for converting tabular data into image format, enabling the use of Convolutional Neural Net-
works (CNNs) for classification. This innovative technique harnesses the powerful feature extraction
capabilities of CNNs, resulting in superior classification accuracy compared to traditional and ensem-
ble learning models. The integration of these advanced methodologies significantly enhances the ac-
curacy and reliability of migraine classification. The implications of these findings are substantial for
medical informatics and migraine management, enabling more targeted and effective treatment plans.
Moreover, the novel methodologies presented in this thesis can be adapted to other medical classifica-
tion issues, further extending their potential impact. Overall, this research demonstrates the effective-
ness of advanced machine learning and deep learning techniques in enhancing the classification of
migraine types, laying the groundwork for more precise and reliable diagnostic tools in healthcare.
