I have an ECG signal and want to predict the nth instance of signal based on previous m samples.
Which of these (or some other) solution will be best suitable for it:
- Using past m samples as features and (m+1)th as the y
- Using Signal Modelling algorithms (like Padé approximant or Levinson-Durbon Recursion)
If using Signal Modelling algorithms, which algorithm (and order of poles) will be most suitable to ECG as it can vary quickly. Thanks!