Hi all, I have a general question regarding the use of a recurrent neural network to model data. I have time dependent data (clinical measurements taken at 5 different time points) for a number of patients. I've therefore assumed a recurrent neural network of some sort would be useful in capturing the time dependency. I'm a little confused however, of what constitutes a reasonable number of time points to warrant a RNN. Is this a genuine consideration when deciding on a model?