• Yashraj Wable
• Yogiraj Wabale
• Yogita Pacharne
• Aakanksha Londhe
• H.R. Kulkarni, Nikita Bichitkar

Migraine is a complex neurological disorder affecting over a billion individuals globally, imposing significant personal, social, and economic burdens. It is clinically classified into migraine with aura (MwA) and migraine without aura (MwoA). Aura, characterized by transient visual, sensory, or speech disturbances, often precedes headache onset in MwA, whereas MwoA lacks these precursors but may present with longer or more frequent attacks. Accurate differentiation between these subtypes is critical for tailored clinical management and risk mitigation. This study evaluates the distinguishing features of MwA and MwoA using both statistical analyses and machine learning (ML) models. A dataset of 800 patient records (400 MwA, 400 MwoA) was analyzed, including demographic data, clinical symptoms, attack frequency, duration, and family history. Statistical tests (t-tests, chi-square, and logistic regression) identified significant differences between subtypes. ML models—including Random Forest, Support Vector Machine (SVM), Decision Tree, K-Nearest Neighbors (KNN), and Logistic Regression—were trained and tested to classify migraine subtype and determine feature importance. Results indicate that MwA patients exhibit significantly longer attack duration, higher prevalence of visual and sensory aura, and a greater family history association. Random Forest outperformed other models with 88% classification accuracy, and feature importance analysis highlighted visual aura, sensory symptoms, and family history as the most predictive features. Integrating ML with traditional statistical analysis enhances migraine subtype differentiation and offers clinical decision support opportunities. Recommendations include ML-assisted diagnostic tools, patient-monitoring systems, and incorporation of real-time symptom tracking. While this study is limited to retrospective secondary data, it establishes a framework for future research incorporating neuroimaging, wearable sensors, and longitudinal patient monitoring.

migraine, migraine with aura, migraine without aura, aura symptoms, visual aura, sensory aura, attack duration, attack frequency, family history, neurological disorder, statistical analysis, t-test, chi-square test, logistic regression, machine learning, Random Forest, SVM, Decision Tree, KNN, feature importance, classification accuracy, clinical decision support, diagnostic tools, symptom tracking, retrospective study, neuroimaging, wearable sensors, longitudinal monitoring

"A Comparative Study On Migraine With Aura And Without Aura Using Statistical And Machine Learning Models", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 1, page no.115-125, January-2026, Available :http://www.jetir.org/papers/JETIRHG06012.pdf

"A Comparative Study On Migraine With Aura And Without Aura Using Statistical And Machine Learning Models", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 1, page no. pp115-125, January-2026, Available at : http://www.jetir.org/papers/JETIRHG06012.pdf