TimeREISE: Time Series Randomized Evolving Input Sample Explanation

. 2022 May 27;22(11):4084. doi: 10.3390/s22114084

Algorithm 2 Mask application-Attribution.

1:
Define: x as input, $θ$ as classifier, $σ$ as perturbation function, S as list of scores and N as feature occurrences across all masks M.
2:
for $m = 1, \dots, M$ do
3:
$x_{m_{i}} = σ (x, m_{i})$
4:
$y^{'} = θ (x_{m_{i}})$
5:
Append $y^{'}$ to S
6:
end for
7:
$S = \frac{S^{T} \times M}{N}$
8:
$S = \frac{S - min (S)}{max (S - min (S))}$