Developing Fine-Grained Actigraphies for Rheumatoid Arthritis Patients from a Single Accelerometer Using Machine Learning

. 2017 Sep 14;17(9):2113. doi: 10.3390/s17092113

Algorithm 2 Dichotomy mapped forest—Test

Input: Z (test set),

m o d e l

S e l_{n o d e s}^{f e a t}

m a p

Output:

Y^{p r e}

1:
for Each node in logic three according to the dichotomy in Figure 3 do
2:
$Z_{n o d e}^{s e l}$ ← select $S e l_{n o d e}^{f e a t}$ features in $Z_{n o d e}$
3:
${m a p p e d}_{Z}$ ← map input $Z_{n o d e}^{s e l}$ with $m a p (n o d e)$
4:
[ $Y_{n o d e}^{p r e}$ , $C_{n o d e s}$ ] ← test ${m a p p e d}_{Z}$ in $m o d e l (n o d e)$
5:
end for
6:
[ $Y^{p r e}$ , $C_{n o d e s}^{n o r m}$ ] ← apply Dichotomy (predictions = $Y_{n o d e s}^{p r e}$ , confidence = $C_{n o d e s}$ ) as Figure 3 & Equation (7)
7:
${b o u n d}_{n o d e s}$ ← get confidence bound using $C_{n o d e s}^{n o r m}$ as Equation (9)
8:
if $C_{n o d e}^{n o r m}$ < ${b o u n d}_{n o d e}$ then
9:
$Y^{p r e}$ ← check logic filter as in (11)
10:
end if