The threshold rule includes an equality choice. This lesson pins the row whose score equals the threshold and shows why it becomes class 1 under the stated policy.
highlighted = computed this step
Pin the equality policy
Use exact scores -1, 0, and 1 with threshold 0. The stated policy is score >= threshold -> class 1.
scores=(−1,0,1),t=0
Apply it row by row
The score -1 is below the threshold, so it gets class 0. The scores 0 and 1 are at or above the threshold, so they get class 1.
(−1,0,1)⇒(0,1,1)
The boundary row is included
The row with score 0 becomes class 1 because equality is included in the stated policy. No sigmoid decimal or probability value is computed.
0≥0⇒class 1
Boundary
A different stated policy, score > threshold, would put the zero row on the other side. This lesson pins the >= policy only: not a claim of accuracy, calibration, training, generalization, or probability.