Question at position 17 Which activation function is commonly used in the target layer when modeling a binary target?exponentiallogisticidentitytanh单项选择题
A
exponential
B
logistic
C
identity
D
tanh

Question

Question at position 17 Which activation function is commonly used in the target layer when modeling a binary target?exponentiallogisticidentitytanh

BlackTom AI · Accepted Answer

The logistic (sigmoid) function sigma(z) = 1 / (1 + e^{-z}) maps any real-valued input to the interval (0, 1), making it the standard choice for outputs that represent probabilities in binary classification.

1. Relevant concepts
- Activation functions transform a network’s raw output (logit) into a range appropriate for the prediction target.
- For binary targets, we want an output interpretable as P(y=1 | x), i.e., constrained to [0,1].
- The logistic (sigmoid) activation corresponds to the Bernoulli/ logistic regression model and pairs naturally with binary cross-entropy (log-loss) as the likelihood-based loss.

2. Step-by-step reasoning
- Compute raw score z from final linear neuron: z = w^T x + b.
- Apply logistic activation: sigma(z) = 1 / (1 + e^{-z}).
- sigma(z) produces values strictly between 0 and 1; with a decision threshold (commonly 0.5) you classify into the two classes.
- Training via gradient descent uses cross-entropy loss L = -[y log(sigma(z)) + (1-y) log(1 - sigma(z))], which is convex with respect to z for a single neuron and yields well-behaved gradients.

3. Example calculations (intermediate steps)
- If z = 0: sigma(0) = 1 / (1 + e^0) = 1/2 = 0.5.
- If z = 1: sigma(1) = 1 / (1 + e^{-1}) ≈ 0.731.
- If z = -1: sigma(-1) ≈ 0.269.
These show outputs in (0,1), interpretable as probabilities.

4. Why logistic is correct
- Its output range and probabilistic interpretation make it ideal for modeling binary targets. It has a well-known probabilistic derivation (logit link for Bernoulli), stable gradients for learning, and direct compatibility with binary cross-entropy.

5. Why other options are incorrect
- Exponential: maps to (0, ∞) and is unbounded above, not a probability.
- Identity: outputs any real number, not constrained to [0,1], so not directly a probability.
- Tanh: maps to (-1, 1), includes negative values and is centered at zero; outputs would need extra transformation to represent probabilities, so it is less natural than logistic.

Conclusion: The correct activation for a binary target is logistic, matching the provided answer.

Question at position 17 Which activation function is commonly used in the target layer when modeling a binary target?exponentiallogisticidentitytanh单项选择题

类似问题

Which statements are correct regarding "Activation" Function?

Question at position 16 Which of following activation function can’t be used at output layer of a neural network to classify an image?TanhSigmoidIf(x>5,1,0)ReLUNone of the above

Which of the followings is not a widely used activation function?

Question at position 13 What activation function should we use for the output layer for the MNIST dataset?SoftmaxReLU

Question at position 13 What activation function should we use for the output layer for the MNIST dataset?SoftmaxReLU

Question at position 13 What activation function should we use for the output layer for the MNIST dataset?SoftmaxReLU

Question at position 11 Which of the following computation adds nonlinearity to an ANN?Performing activation functionTaking weighed sum of inputs

Question at position 11 Which of the following computation adds nonlinearity to an ANN?Performing activation functionTaking weighed sum of inputs

更多留学生实用工具

智能学习助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单