Fig. 3

The model hierarchy for the Bayesian Clearance Estimator. Note that each blue box represents a patient and within the blue box of patient i, there are \(n_i\) red boxes representing her associated parasitemia measurements. In this graphical model, a collection of variables A all pointing to a variable b simply means the distribution of b is described by the variables in A. Observe that, following the notation introduced in the previous section, a set of parameters \(\theta _i = \{ \alpha _i, \beta _i, \delta _i^{\ell }, \delta _i^\tau \}\) describes the distribution of the measurement vector \(\varvec{y}_i = \{ y_{ij} \}_{j=1}^{n_i}\) for patient i. Furthermore, each element of \(\theta _i\) is assumed to be a sample of a distribution described by the hyperparameters appearing in the figure. For example, \(\beta _i\) is assumed to be generated from a distribution which is described by \(\gamma\) and \(\sigma _\beta ^2\). This distribution and others are all introduced in the text