Abstract
Biphasic property in female body temperature during menstrual cycle is estimated by a discrete Hidden Markov Model (HMM)-based approach. Estimation procedure includes three steps - preprocessing, HMM-based main processing and postprocessing. The HMM is supposed to have two hidden phases to describe the biphasic property in body temperature during a menstrual cycle. Three kinds of different body temperature data were collected daily from four female volunteers over six months. Skin and core body temperatures were measured at intervals of ten and four minutes respectively and automatically by two separated wearable devices during sleep. Oral basal body temperature was measured in the morning right after wakeup. Estimation results of biphasic property from different body temperatures were evaluated by quantifying the alignment between estimated phase transitions and volunteers' menstruation records. Results showed that the estimation performance, in terms of sensitivity and positive predictability, varies with different kind of body temperature. Among 21 menstrual cycles in four participants during six months, both overall sensitivity and positive predictability of estimation by oral basal body temperature reach the highest 95.2%; those of skin body temperature have the lowest 81.0% and 77.3%, respectively; while those of core body temperature are 90.5% and 82.6%, respectively, straddling between oral and skin body temperatures.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have