Figure 1
From: Predicting Patient-ventilator Asynchronies with Hidden Markov Models
Time series of asynchronies count indexed each 15 min (a–c) and their corresponding histograms (d–f) from three representative patients. Note these are discrete time series of non-negative values, and positively skewed (histogram characterized by a long tail in the positive direction) with several observations having a value close to zero, which justifies the use of a Poisson distribution to model this kind of data.
