EXERCISES
CLARIFYING THE CONCEPTS
- Neural networks classification represents an attempt to imitate what?
- Using Figure 9.1, explain how an artificial neuron model imitates the actions of real neurons.
- What is the main benefit of neural networks for modeling? What gives neural networks this power?
- Describe the main drawback of neural network modeling.
- Explain what we mean when we say that a neural network is completely connected.
- Describe the benefits and drawbacks of using more or fewer nodes in the hidden layer.
- Referring to the example in the text, calculate netB = 1.5 and
--f ( n e t B ) = 1 1 + e − 1.5 = 0.8176 - Explain how the sigmoid function combines nearly linear behavior, curvilinear behavior, and nearly constant behavior.
- Describe the process of backpropagation.
- The essential problem for the neural network is to construct a set of weight that will minimize what?
WORKING WITH THE DATA
For the following exercises, work with the Framingham_training and Framingham_test data sets. Use either...