Investigation of inflammatory risk factors in restless legs syndrome using logistic regression analysis and machine learning algorithms

Introduction: In recent years, the rapid advancement of technology and the proliferation of large and complex datasets have made artificial intelligence increasingly fascinating within the field of statistics. Especially in health-related studies, interest in comparing logistic regression analysis and machine learning methods has increased. Our study aimed to investigate the potential role of systemic inflammatory markers in the pathogenesis of restless legs syndrome (RLS) in patients with sleep disturbances using logistic regression analysis and machine learning methods. Method: Between January 2022 and June 2023, 230 patients aged 18-65 diagnosed with RLS and 230 healthy controls were included. Demographic and laboratory data of the patient and control groups were recorded retrospectively. Independent variables investigated for possible role in disease prediction were analyzed with logistic regression, support vector machines (SVM) and K nearest neighbor (KNN) algorithms. After separating the data into training and test data using R statistical software, the findings were compared. Results: The patient group had a median age of 51.0 years (IQR: 44.0–58.0) and a female rate of 73.9%, whereas the control group had a median age of 43.5 years (IQR: 37.0–49.0) and a female rate of 58.7%. The accuracy percentage of the model on the test data was found to be 65.21% for the KNN algorithm, 51.44% for the SVM algorithm, and 50.0% for the LR model. In our study, when comparing the metric measures of all three models, the KNN model demonstrated higher performance than the other two models on both the training and test data. Conclusion: In our study, we aimed to investigate the potential role of systemic inflammatory markers in the pathogenesis of restless legs syndrome (RLS) in patients with sleep disorders using logistic regression analysis and machine learning methods. We believe that the development of machine learning algorithms will contribute to advancements in the healthcare field as well.