The Use of Data Mining Techniques for Analysis of Menstrual Cycle Parameters and Prognosis of Migraine Symptoms in Reproductive Age Women

摘要

The article represents the results of using the data mining algorithms to determine probability of occurrence migraine symptoms related to changes of menstrual cycle parameters in women of reproductive age. The analysis was carried out on the basis of logistic regression models, developed by three different methods. Relations between menstrual amount changes, regularity cycle and onset of migraine symptoms were determined. Sensitivity and specificity of developed model was counted. Its quality assessment was appraised by means of ROC (Receiver Operating Characteristics) analysis.