Deck 5: Machine-Learning Techniques for Predictive Analytics

Question 1

In the opening vignette, the high accuracy of the models in predicting the outcomes of complex medical procedures showed that data mining tools are ready to replace experts in the medical field.

Answer 1

False

Question 2

The use of hidden layers and new topologies and algorithms renewed waning interest in neural networks.

Answer 2

True

Question 3

In the mining industry case study, the input to the neural network is a verbal description of a hanging rock on the mine wall.

Answer 3

False

Question 4

In supervised learning techniques, such as backpropagation, the training data consist of vector pairs-an input vector and a target vector.

Question 5

A disadvantages of Hopfield neural networks is that their structure cannot be replicated on an electronic circuit board.

Question 6

The most complex problems solved by neural networks require one or more hidden layers for increased accuracy.

Question 7

Naïve Bayes is a simple probability-based classification method derived from the Bayes theorem.

Question 8

The strong assumption of independence among the input variables in the Naïve Bayes method is realistic.

Question 9

The Naïve Bayes method is a powerful tool for representing dependency structure in a graphical, explicit, and intuitive way.

Question 10

Pearl won the prestigious ACM's A.M. Turing Award for his contributions to the field of artificial intelligence and the development of BN.

Question 11

Model ensembles tend to be more robust against outliers and noise in the data set than individual models.

Question 12

The k-nearest neighbor algorithm is overly complex when compared to artificial neural networks and support vector machines.

Question 13

The k-nearest neighbor algorithm appears well-suited to solving image recognition and categorization problems.

Question 14

Because of their complexity, it is more difficult to understand the inner structure of model ensembles (how they do what they do) than individual models.

Question 15

Ensemble models can be quickly characterized based on their use of a bagging or boosting method type.

Question 16

Which of the following are advantages of the Naïve Bayes method or classification?

A) developed very efficiently
B) developed effectively and accurately in a supervised machine learning environment
C) absent any underlying assumptions that may affect output
D) ability to quickly create models and outputs

Question 17

Some of the benefits of the BN model include:

A) ease of adaptability.
B) extent of applicability.
C) both A and B.
D) none of these.

Question 18

Backpropagation requires the of vector pairs, with the pairs consisting of:

A) an input vector and a control vector.
B) binary input vectors.
C) a control vector and a target vector.
D) an input vector and a target vector.

Question 19

Homogeneous-type ensembles combine the outcomes of:

A) two or more of the different type of models.
B) only two of the different type of models.
C) two or more of the same type of models.
D) only two of the same type of models.

Question 20

Bagging can be used for:

A) classification-type prediction problems.
B) regression/estimation-type prediction problems.
C) both A and B
D) none of these

Question 21

The random forest (RF) model is a modification to what algorithm?

A) complex bagging
B) simple bagging
C) complex boosting
D) simple boosting

Question 22

BN is a powerful tool for representing dependency structure in a ________, explicit, and intuitive way.

Question 23

The methodology employed in the traffic case follows a very well-known standardized analytics process know by its acronym _________.

Question 24

Model ensembles tend to be more ________ against outliers and noise in the data set than individual models.

Question 25

Why have neural networks shown much promise in many forecasting and business classification applications?

Question 26

How is a general Hopfield network represented architecturally?

Question 27

Describe the Tree Augmented Naïve (TAN) Bayes method.

Question 28

Describe the taxonomy for model ensembles.

Question 29

List the pros and cons of model ensembles compared to individual models.

Question 30

In 1992, Boser, Guyon, and Vapnik suggested a way to create nonlinear classifiers by applying the kernel trick to maximum-margin hyperplanes. How does the resulting algorithm differ from the original optimal hyperplane algorithm proposed by Vladimir Vapnik in 1963?