Suppose a time series data is generated with random noise. Due to random behavior of noise, it is very complex to model these noisy data. Which and how DL denoising methods handle this problem?
BTW @Babek response reminded me that robust time series analysis exists. You might look at Robust Statisics by maronna.martin and yohai in the z-library google instructions for it's use. this isn't yet in R but is in S-plus I believe. Best wishes, David Booth