S: signed sigmoid function 𝑆(𝑎)=𝑠𝑖𝑔𝑛[𝜎(𝑎)−0.5]=𝑠

[answerhappygod]

S: signed sigmoid function 𝑆(𝑎)=𝑠𝑖𝑔𝑛[𝜎(𝑎)−0.5]=𝑠𝑖𝑔𝑛[
1/(1+𝑒𝑥𝑝(−𝑎) )−0.5]
L: linear function 𝐿(𝑎)=𝑐𝑎
Where in both cases, 𝑎=∑ 𝜔𝑖𝑋𝑖𝑖
Assign proper activation functions (S or L) for each unit in the
following graph so this neural network simulates a boosting
classifier that combines two logistic regression classifiers, 𝑓1:𝑋
→𝑌1 and 𝑓2:𝑋 → 𝑌2, to produce its final prediction: 𝑌=𝑠𝑖𝑔𝑛[𝛼1𝑌1
+𝛼2𝑌2]. Use the same definition in problem (b) for the logistic
regression functions 𝑓1 and 𝑓2.

1 (17.63 KiB) Viewed 21 times

W1 X1 05 W2 -Y W3 W6 X2 W4