How Validation Methodology Influences Human Activity Recognition Mobile Systems

. 2022 Mar 18;22(6):2360. doi: 10.3390/s22062360

Algorithm 1 SHAP basic algorithm

1:
Required: M: Number of interactions; x: Instance of interest; j: Features index; X: Data matrix; f: Machine learning model.
2:
procedureSHAP( $M, x, j, X, f$ )
3:
for m=1 in M do
4:
Draw random instance z from the data matrix X
5:
Choose a random permutation o of the feature values
6:
Order instance x: $x_{0} = (x_{(1)}, . . ., x_{(j)}, . . ., x_{(p)})$
7:
Order instance z: $z_{0} = (z_{(1)}, . . ., z_{(j)}, . . ., z_{(p)})$
8:
Construct two new instances
9:
Compute marginal contribution $ϕ_{j}^{m} = \hat{f} (x_{+ j}) - \hat{f} (x_{- j})$
10:
end for
11:
Compute Shapley value as the average $ϕ_{j} (x) = \frac{1}{M} \sum_{m = 1}^{M} ϕ_{j}^{m}$
12:
end procedure