Algorithm 2. (Continuous) Correlated Beta Process Update Rule

This rule is for updating the beta distribution for the probability of observing response $r$ for some dose $d_i$ based on the $\left(n+1\right)th$ data point. Let this $\left(n+1\right)th$ data point have been at dose $d_j$.

BEGIN ALGORITHM

Calculate $K\left(d_i,d_j\right)$

If response $r$ was observed for individual $n+1$

Set ${\alpha }_{i,r}^{n+1}={\alpha }_{i,r}^n+K\left(d_i,d_j\right)$

Set ${\beta }_{i,r}^{n+1}={\beta }_{i,r}^n$

Else (response $r$ was not observed for individual $n+1$)

Set ${\alpha }_{i,r}^{n+1}={\alpha }_{i,r}^n$

Set ${\beta }_{i,r}^{n+1}={\beta }_{i,r}^n+K\left(d_i,d_j\right)$

END ALGORITHM