Deep Learning in Time Series Analysis

9 Deep Learning of Cyclic Time Series

Real-world activities in many cases are examples of cyclic activities, resembling cyclic time series, when recorded by appropriate instruments reflecting the behaviour of the activities. It is evident that a level of noise is always associated with the recording due to many factors, such as electrical induction and thermal condition. In biological activities such as heart sound signal, even more factors come into the phenomenon such as biochemical status of the blood, heart rhythm and respiration, all constituting different sources of the noise, that can resemble cyclic characteristics, even at the presence of the noises. One should pay attention to the difference between periodic and cyclic time series ...