Table - PMC

Skip to main content

View full-text article in PMC

. 2020 Mar 30;20(7):1932. doi: 10.3390/s20071932

Algorithm 3 A

NNOTATE

Input:x, E, semi-label

L

,

{\vec{U}}_{t h}

Output: label ℓ

1:
$u \leftarrow 100$
2:
for all On-demand expert $e_{i} \in E$ do
3:
$L \leftarrow M AJORITY V OTE (x)$
4:
$u \leftarrow$ $M IN U NCERTAINTY (e_{i}, L)$ in labeling x
5:
end for
6:
while $u > U_{t h} (L)$ do
7:
$E_{c o n f} \leftarrow Ø$ , $C_{c o n f} \leftarrow Ø$
8:
for all perfect or imperfect expert $e_{i} \in E$ do
9:
if $U_{i} (L) \leq u_{t h}$ then
10:
add $e_{i}$ to the $E_{c o n f}$ /*confident experts*/
11:
add $c o s t (e_{i})$ to $C_{c o n f}$ /*cost of experts*/
12:
end if
13:
end for
14:
$e x p e r t$ ← $a r g m i n_{e_{i} \in E_{c o n f}} C_{c o n f}$
15:
$L \leftarrow$ label from $e x p e r t$ for instance x
16:
$u \leftarrow$ uncertainty of $e x p e r t$ in labeling x
17:
end while
18:
$ℓ \leftarrow L$