Table 2 Experimental procedure.
From: Finding hotspots: development of an adaptive spatial sampling approach
Pseudo code for experiments | ||
|---|---|---|
1 | for rep in \(1\rightarrow 100\): | |
2 | for m in \(\{1, 10, 50\}\): | m = batch size |
3 | \({\mathscr {A}}_0 \leftarrow\) random selection: \({\mathscr {A}}_0 \subset {\mathscr {S}}\) with \(\Vert {\mathscr {A}}_0\Vert = 100\) | \({\mathscr {S}} =\) all villages |
4 | \({\mathscr {A}}^R_0 = {\mathscr {A}}^A_0 = {\mathscr {A}}_0\) | \(R/A=\)random/adaptive |
5 | \(steps = 100 / m\) + 1 | total number of iterations |
6 | for t in \(1 \rightarrow steps:\) | |
7 | \({y}^{\star}_{i} \sim \text{Binomial}{(100, \theta ({\mathscr{A}^{\star}_{t-1}}))}^{\dagger}\) | \(\star =\{R, A\}\) |
8 | \({\mathscr {D}}^\star _{t-1} = \{{\mathscr {A}}^\star _{t-1}, y^\star , \mathbf{x} ({\mathscr {A}}^\star _{t-1}) \}\) | \(\mathbf{x}\)=environmental data |
9 | find \(p(\theta > \vartheta | \cup _{k=0}^{t-1} {\mathscr {D}}^\star _k)\) | |
10 | compute validation statistics on \({\mathscr {S}} \setminus \cup _{k=0}^{t-1} {\mathscr {A}}^\star _k\) | |
11 | \({\mathscr {A}}^R_t \leftarrow\) random selection | |
12 | \({\mathscr {A}}^A_t \leftarrow\) acquisition function Eq. (6) | |
