Fig. 9From: Representation learning for clinical time series prediction tasks in electronic health recordsThe effect of different training data sizesBack to article page