In PROC HPNEURAL, which activation function is specified for hidden layer neurons using the HIDDEN statement?
Select an answer to reveal the explanation.
Short Explanation
You specify the activation function for hidden neurons with the ACTFUNC= option in the HIDDEN statement, such as TANH, LOGISTIC, or RELU. PROC HPNEURAL doesn't lock you into sigmoid or force a global setting.
Full Explanation
The ACTFUNC= option in the HIDDEN statement specifies the activation function for hidden layer neurons, such as TANH, LOGISTIC, or RELU. The claim that activation functions cannot be specified and sigmoid is used for all layers is wrong. The claim that activation functions are set globally rather than per layer is wrong because HIDDEN controls each layer. The claim that the INPUT statement's TYPE= option determines hidden activation is wrong because TYPE= describes input normalization, not activation.