Deck 10: Data Mining

Question 1

The data might be normalized so that each variable is expressed on a common scale.

Answer 1

True

Question 2

Classification refers to a type of data mining problem that uses the information available in a set of independent variables to predict the value of a discrete, or categorical, dependent variable.

Answer 2

True

Question 3

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What is the verbal test score value of the group centroid for group 3?

A) 697.71
B) 647.86
C) 587.67
D) 605.17

Answer 3

605.17

Question 4

Data mining tasks fall into three potential categories: Classification, Prediction and Association/Segmentation.

Question 5

A classification tree is a graphical representation of a set of rules for classifying observations into one group.

Question 6

Affinity analysis is a data mining technique that attempts to discover

A) what goes with what
B) the relationship between independent vatiables
C) multicollinearity
D) causality

Question 7

Suppose that an analyst classified a new record using the following sequential steps (i) find identical records in the training sample, (ii) determine a group, to which majority of these records belong, (iii) assign the new record to the group in step (ii). This technique is called

A) naive Bayes
B) Bayes
C) conditional probability estimation
D) posterior probability estimation

Question 8

​Data mining is the process of finding and extracting useful information and insights from large data sets.

Question 9

The Fisher classification scores can be converted to

A) a linear function for each of the groups in the classification problem
B) probabilities of group membership
C) a uniform distribution
D) a half-space

Question 10

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What is the quantitative test score value of the group centroid for group 1?

A) 697.71
B) 647.86
C) 587.67
D) 650.43

Question 11

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What percentage of observations is classified correctly?

A) 100%
B) 85.71%
C) 95%
D) 90%

Question 12

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What number of observations is classified correctly?

A) 19
B) 20
C) 7
D) 8

Question 13

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What is the verbal test score value of the group centroid for group 1?

A) 697.71
B) 647.86
C) 587.67
D) 650.43

Question 14

The Mahalanobis distance measure accounts for differences in the covariances between all possible pairings of the independent variables.

Question 15

If using the regression tool for two-group discriminant analysis, in the regression dialog box, the Input X-Range entry corresponds to

A) the Group values.
B) the independent variable values.
C) the predicted variable values.
D) the fitted variable values.

Question 16

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. What percentage of the observations is classified incorrectly?

A) 90%
B) 80%
C) 85%
D) 15%

Question 17

In hierarchical clustering, the measure of similarity between clusters is/are

A) single linkage
B) complete linkage
C) Ward's method
D) all of the above

Question 18

​The Get Data command is part of the XLMiner Platform in Excel add-on.

Question 19

The k-nearest neighbor (k-NN) technique identifies the k observations in the training data that are most similar (or nearest) to a new observation we want to classify.

Question 20

When purity is perfect, the Gini index is equal to

A) 0
B) 0.25
C) 0.5
D) 1

Question 21

In discriminant analysis the averages for the independent variables for a group define the

A) centroid.
B) median.
C) mode.
D) central tendency.

Question 22

One element in cleaning the data set in the mining process involves

A) removing unimportant variables
B) adding more variables to the data set
C) calculating the adjusted R²
D) calculating the coefficient of multiple correlation

Question 23

Discriminant analysis (DA) differs from most other predictive statistical methods because the dependent variable is

A) continuous
B) random
C) stochastic
D) discrete

Question 24

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. What is the quantitative test score value of the group centroid for group 2?

A) 683.8
B) 654.2
C) 610.7
D) 605.7

Question 25

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. The university has received applications from several new students and would like to predict which group they would fall into. What is the discriminant score for a student with a Quantitative score of 686 and a Verbal score of 601. Use five (5) significant figures in your coefficients.

A) 1.29 ≤ discriminant score ≤ 1.30
B) 1.69 ≤ discriminant score ≤ 1.70
C) 2.69 ≤ discriminant score ≤ 2.70
D) 6.05 ≤ discriminant score ≤ 6.06

Question 26

The dependent variable in the regression equation represents

A) the estimated value of the dependent variable.
B) the estimated value of the Group variable.
C) the estimated ranking of the subject
D) all of these are true.

Question 27

Useful data mining techniques can be found in Excel under ___________ drop menu

A) Data/(Data Analysis)
B) Regression
C) Histogram
D) Insert/Chart

Question 28

In hierarchical clustering, the measure of similarity between clusters is/are

A) single linkage
B) average linkage
C) average group linkage
D) all of the above

Question 29

The k-means clustering algorithm is available

A) in XLMiner Excel add-in
B) as a regression option in Excel
C) as ANOVA option in Excel
D) as one of the options under non-parametric statistics in Excel

Question 30

Overfitting refers to

A) placing too much emphasis on the sample-specific noise
B) fitting the model too tightly
C) fitting the model too loosely
D) underestimating model parameters

Question 31

Prediction step in data mining is an option available in

A) XLMiner Excel add-in
B) Excel
C) data mining software
D) nonlinear multivariate regression

Question 32

Suppose that two variables are found to be significantly correlated. A researcher may

A) remove one variable from the data set
B) replace the two variables by their product
C) replace the two variables by their squared difference
D) remove both variables from the data set

Question 33

Steps in the data mining process include the following (in sequence)

A) (identify opportunity), (collect data), (explore, understand and prepare data), (identify tasks and tools, (partition data), (build and evaluate models), (deploy models)
B) (identify opportunity), (collect data), (explore, understand and prepare data), (identify tasks and tools, (build and evaluate models), (deploy models)
C) (collect data), (explore, understand and prepare data), (identify tasks and tools, (partition data), (build and evaluate models), (deploy models)
D) (identify opportunity), (collect data), (identify tasks and tools, (partition data), (build and evaluate models), (deploy models)

Question 34

Suppose that the correlation coefficient between X₁ and X₂ is equal to -1. This means that

A) X₁ and X₂ are perfectly positively correlated
B) X₁ and X₂ are perfectly negatively correlated
C) X₁ and X₂ are weakly and positively correlated
D) X₁ and X₂ are weakly and negatively correlated

Question 35

____________ is a classification technique that estimates the probability of an observation belonging to a particular group

A) logistic regression
B) binary regression
C) multivariate analysis
D) ANCOVA

Question 36

The discriminant score is denoted by

A)
B)
C) Y_i
D)

Question 37

The Fisher linear discriminant function

A) identifies a linear function for each of the groups in the classification problem
B) fits a nonlinear function for each of the groups in the classification problem
C) defines a hyperplane
D) defines a half-space

Question 38

In a two-group discriminant analysis problem using regression, why is the midpoint cut-off value used to determine group classification?

A) Because the value minimizes the absolute misclassification error.
B) Because the value minimizes the probability of misclassification error.
C) Because the value represents an equal division between the groups.
D) Because the value incorporates problem specific knowledge.

Question 39

In the k nearest neighbor technique, a small value of k produces classifications that are

A) very sensitive to the sample-specific characteristics of the training data
B) not sensitive to the sample-specific characteristics of the training data
C) robust
D) reliable

Question 40

Suppose that there are 3 variables in a data set. Approximately how many data records are required using a rule of thumb discussed in the textbook?

A) 30 to 45
B) 20 to 30
C) 45 to 60
D) 50 to 100

Question 41

The parameters of the logistic regression model

A) are derived through a nonlinear maximum likelihood estimation procedure
B) are negative
C) are positive
D) are negative fractions

Question 42

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What percentage of observations is classified incorrectly?

A) 5%
B) 15%
C) 95%
D) 90%

Question 43

Suppose that a data set contains a variable EDUCATION, which has 7 discrete levels. EDUCATION is an example of

A) a categorical variable
B) a classification variable
C) a continuous variable
D) an exponential variable

Question 44

Logistic regression in XLMiner add-in can be used for ______ groups

A) 2
B) 3
C) 4
D) 5

Question 45

Data mining tasks fall in the following categories

A) classification, prediction, association
B) categorization, segmentation
C) prediction, association, mining
D) observation, categorization, association

Question 46

Cluster analysis is a data mining technique used for

A) grouping together similar data
B) segmentation of records within a data set
C) designing effective marketing strategies
D) all of the above

Question 47

Neural networks classification methodology

A) is one of the options available in XLMiner Excel add-in
B) is superior to other classification schemes
C) is inferior to other classification schemes
D) produces results that are identical to the outcomes of the DA technique

Question 48

Technique(s) used in classification step of data mining include

A) discriminant analysis
B) logistic regression
C) neural networks
D) all of the above

Question 49

In the ________ step of data mining, a researcher attempts to form logical groupings of data in the set

A) classification
B) prediction
C) categorization
D) association/segmentation

Question 50

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. Based on the regression output, what is the discriminant score for a student with a quantitative score of 635 and a verbal score of 570?

A) 1.72 ≤ discriminant score ≤ 1.73
B) 2.02 ≤ discriminant score ≤ 2.03
C) 3.04 ≤ discriminant score ≤ 3.05
D) 6.12 ≤ discriminant score ≤ 6.14

Question 51

Logistic regression is a classification technique that

A) outperforms other techniques accros a variety of data collections
B) is not reliable
C) is not robust
D) is not feasible for most data sets

Question 52

___________ and _________ must be chosen each time a partition is subdivided

A) an independent variable and splitting value
B) a dependent variable and cutoff value
C) a significance level and an upper bound
D) a significance level and a lower bound

Question 53

A graphical representation of a set of rules for classifying observations into 2 or more groups is called

A) a classification tree
B) a binary tree
C) a Pareto diagram
D) a branch-and-bound tree

Question 54

Given the following confusion matrix what is the correct classification rate?

A) 9/13 = 69%
B) 10/14 = 86%
C) 19/25 = 76%
D) 6/19 = 32%

Question 55

Plots useful in data mining analysis can be accessed in Excel using the _______ add-in

A) XLMiner
B) Charts
C) Data Analysis
D) Visual Basic

Question 56

Neural networks are

A) a pattern recognition technique
B) a physical model representation of interrelationships
C) a non-directed graph
D) a binary network

Question 57

Suppose that all observations belong to the same class. The entropy measure for this situation is equal to

A) 0
B) 0.25
C) 0.5
D) 1

Question 58

Two approaches to clustering discussed in the text are

A) k-means clustering and hierarchical clustering
B) intuitive clustering and methodical clustering
C) MIPS clustering and pixel clustering
D) centroid clustering and VMI

Question 59

Exhibit 10.2
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful (Group 1), marginally successful (Group 2) or not-successful (Group 3) in their graduate studies. The officer has data on 20 current students, 7 successful (Group 1), 6 marginally successful (Group 2) and 7 not successful (Group 3). ​ ​ ​ ​
Refer to Exhibit 10.2. What number of observations is classified incorrectly?

A) 19
B) 20
C) 7
D) 1

Question 60

Suppose that a data set contains a variable EDUCATION, which has 7 discrete levels. EDUCATION can be represented by ____ binary variables

A) 6
B) 7
C) 8
D) 9

Question 61

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. What is the verbal test score value of the group centroid for group 2?

A) 683.8
B) 654.2
C) 610.7
D) 605.7

Question 62

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. What is the quantitative test score value of the group centroid for group 1?

A) 683.8
B) 654.2
C) 610.7
D) 605.7

Question 63

Suppose that the correlation coefficient between X₁ and X₂ is equal to 1. This means that

A) X₁ and X₂ are perfectly positively correlated
B) X₁ and X₂ are perfectly negatively correlated
C) X₁ and X₂ are weakly and positively correlated
D) X₁ and X₂ are weakly and negatively correlated

Question 64

A graphical representation of clustering outcomes showing which items should be classified to which clusters is called a(n)

A) dendrogram
B) hierarchical chart
C) horizontal multi layer chart
D) vertical bar chart

Question 65

A major challenge in affinity analysis is to

A) identify the most meaningful rules used to determine what goes with what
B) the conditional probabilities
C) the posterior probabilities
D) the probabilities of states of nature

Question 66

The objective of classification tree algorithms is to

A) minimize the average weighted impurity of the resulting partitions
B) maximize the average weighted impurity of the resulting partitions
C) maximize the average weighted purity of the resulting partitions
D) maximize the likelihood estimator of the resulting partitions

Question 67

The first step in creating a classification tree involves

A) recursively partitioning the independent variables using the outcome of previous partitions
B) creating a subset of variables
C) clustering the variables into a larger superset
D) deciding the number of partitions

Question 68

In the ________ step of data mining, a researcher attempts to predict the value of a continuous response variable based on the data set

A) classification
B) prediction
C) categorization
D) association/segmentation

Question 69

If using the regression tool for two-group discriminant analysis, in the regression dialog box, the Input Y-Range entry corresponds to

A) the Group values.
B) the independent variable values.
C) the predictor variable values.
D) (b) and (c) are both correct

Question 70

In the k nearest neighbor technique, a large value of k produces classifications that

A) are very sensitive to the sample-specific characteristics of the training data
B) place all observations into the most frequently occurring group in the training data
C) are robust
D) are reliable

Question 71

A ______________ provides a visual summary of the improvements that a data mining project provides on a binary classification problem compared to a random guess

A) lift chart
B) Pareto chart
C) Gantt chart
D) histogram

Question 72

Which of the following is not true regarding discriminant analysis?

A) The classification rule translates discriminant scores into group membership.
B) Discriminant analysis is based on discrete or categorical dependent variables.
C) The classification rule selected perfectly classifies the data used to derive the classification rule.
D) The confusion matrix summarizes classification results.

Question 73

The k-nearest neighbor classification technique

A) identifies the k observations in the training data that are most similar to a new observation we want to classify
B) sorts the data in an increasing order
C) sorts the data in a decreasing order
D) works very much like the ranking algorithm

Question 74

A ___________ algorithm is used during the training process to adjust weights in a neural network

A) backpropagation
B) branch-and-bound
C) Simplex
D) cutting plane

Question 75

Affinity analysis is a data mining technique used in marketing research to determine

A) which products are purchased together
B) product clusters
C) the likelihood that a product will be purchased
D) causality

Question 76

Exhibit 10.1
The following questions are based on the problem description and the output below.
A college admissions officer wants to evaluate graduate school applicants based on their GMAT scores, verbal and quantitative. Students are classified as either successful or not-successful in their graduate studies. The officer has data on 20 current students, ten of whom are doing very well (Group 1) and ten who are not (Group 2). ​ ​ ​ ​
Refer to Exhibit 10.1. Suppose that for a given observation, the difference between Mahalanobis distances between group 1 and 2 (G1-G2) is big and negative. This means that

A) The observation is likely to be classified correctly to group 2
B) The observation is likely to be classified correctly to group 1
C) The observation is likely to be classified incorrectly to group 2
D) The observation is likely to be classified incorrectly to group 1

Question 77

To create the training and validation data set for the model use the __________ option in the XLMiner tab

A) partition with oversampling
B) partition without oversampling
C) partition
D) sample

Question 78

A way to detecting and avoiding overfitting is to

A) use the validation sample to calibrate the model
B) use repeated runs of the model and averaging the results
C) use computer simulation
D) use rigorous statistical tools

Question 79

Suppose that all observations in partition j belong to the same group. The Gini index for this situation is equal to

A) 0
B) 0.25
C) 0.5
D) 1

Question 80

In using neural networks, an analyst must decide __________ and ___________

A) how many hidden layers to use and how many nodes to use in each of the hidden layers
B) how accurate the prediction should be and how many iterations to use
C) how many nodes to use and how much time to allow for simulation
D) how many iterations to use and what is the limit on the budget constraint