Question 1

This type of supervised network architecture does not contain a hidden layer.&#10;A) backpropagation&#10;B) perceptron&#10;C) self-organizing map&#10;D) genetic

Accepted Answer

The perceptron is a type of neural network architecture that contains only one layer of output nodes and no hidden layer. It is a supervised learning algorithm that is used for binary classification problems.

Question 2

The test set accuracy of a backpropagation neural network can often be improved by&#10;A) increasing the number of epochs used to train the network.&#10;B) decreasing the number of hidden layer nodes.&#10;C) increasing the learning rate.&#10;D) decreasing the number of hidden layers.

Accepted Answer

Increasing the number of epochs used to train the network can improve the model's generalization performance by allowing it to learn more complex relationships between input and output values. This can help the model identify patterns in the data that it might have missed during earlier training epochs. However, increasing the learning rate or decreasing the number of hidden layers or nodes can lead to overfitting or underfitting of the data, which can result in poor test set accuracy.

Question 3

The total delta measures the total absolute change in network connection weights for each pass of the training data through a neural network. This value is most often used to determine the convergence of a

A) perceptron network.
B) feed-forward network.
C) backpropagation network.
D) self-organizing network.

C

Accepted Answer

The total delta measure is commonly used in backpropagation networks to determine if the network weights have converged to a stable state. It is calculated as the sum of the absolute changes in weights for each training example. Therefore, the correct choice is C, backpropagation network.

Question 4

Two classes each of which is represented by the same pair of numeric attributes are linearly separable if&#10;A) at least one of the pairs of attributes shows a curvilinear relationship between the classes.&#10;B) at least one of the pairs of attributes shows a high positive correlation between the classes.&#10;C) at least one of the pairs of attributes shows a high positive correlation between the classes.&#10;D) a straight line partitions the instances of the two classes.

Accepted Answer

Linear separability means that we can draw a straight line that separates the instances of two different classes. Therefore, option D is the best choice. Option A is incorrect because a curvilinear relationship may make the classes harder to separate. Option B is not necessarily true, and option C is essentially the same as option B.

Deck 9: Building Neural Networks With Ida