|
Algorithm 1 Ensemble fusion strategy with score normalization |
Require:
-
1:
• Test sample x
• Raw Shape Template Matching score:
• Raw TimeVAE anomaly score:
• Binary output from Enhanced DTW:
• Validation set (for min-max bounds)
• Secondary fusion threshold:
Ensure: Final binary decision , where 1 denotes anomalous and 0 normal.
-
2:
▹ Step 0: Normalize raw scores to
-
3:
-
4:
-
5:
▹ Step 1: Preserve all DTW positives
-
6:
if
then
-
7:
-
8:
else ▹ Step 2: Fuse deep learning and shape scores
-
9:
▹ Step 3: Apply secondary threshold
-
10:
if then
-
11:
-
12:
else
-
13:
-
14:
end if
-
15:
end if
-
16:
return
|