Tuning Hyper Parameters of Deep Learning Model to Monitor Obstructive Sleep Apnea (OSA)

Abstract

Currently, a series of developing diseases in nations like India's powers to look for new answers to a continuing observation of health registry. Visiting emergency clinics has become a necessity. Even now for specialist's meeting, which has turned out to be monetarily related and a tedious procedure. Beside the above-mentioned lines, a non-stop checking of this problem is a primary need in medicinal offerings arrangements. There are some diseases which affects the quality of the lifestyle in a very slow manner. Sleep is considered to be most important activity in human day to day activities. During sleep most of the essential processes happens which benefits human body. Number of people affected by sleeping problems, is increasing due to current lifestyle. One such problem commonly found in humans is Obstructive Sleep Apnea (OSA). There are a few frameworks for OSA recognition. Hence, this exploration displays framework for both to acknowledge and help for the treatment of OSA of aged, home alone persons by observing various factors, like sleeping position, rest status, physical activities and physical parameters just as the utilization of open information accessible in smart urban communities. Our framework engineering performs two sorts of handling. From one perspective, a pre-preparing dependent on guidelines that empowers the sending of continuous notifications to the attendee, in case of a crisis circumstance. In this paper, we discuss various tuning parameters for constructing deep learning model by using the data received from the conducted experiments.