Chapter 2

Question 1

Binning

Answer 1

Categorizing a metric variable into a smaller number of categories/bins and thus converting the variable into a nonmetric form

Question 2

Boxplot

Answer 2

box reporesents the major portion of the distribution and the extensions (whiskers) reach to teh extreme points of the distribution
- useful in making comparisons of one or more metric variables across groups formed by a nonmetric variable

Question 3

Cardinality

Answer 3

the number of distinct values for a variable

Question 4

Censored data

Answer 4

observations that are incomplete in a systematic and known way
- censored data are an example of “ignorable missing data”

Question 5

curse of dimensionality

Answer 5

problems associated with including a very large number of variables in the analysis
- distance measures becomign less useful
- higher potential for irrelevant variables
- differing scales of measurement for the variables

Question 6

Data management

Answer 6

all activities associated with assembling a dataset for analysis

Question 7

Data quality

Answer 7

accuracy of the informatin in a dataset
● 8 dimensions → completeness / availability and accessibility / currency / accuracy / validity / usability and interpretability / reliability and credibility / consistency

Question 8

dCor

Answer 8

newer measure of association that is distance-based and more sensitive to nonlinear pattenrs in the data

Question 9

dichotomization

Answer 9

dividing cases into classes basedon being above or below a specific value

Question 10

elasticity

Answer 10

measure of ratio of % change in Y for a % change in X.
Obtained by using a log-log transformation of both DVs and IVs

Question 11

Heat map

Answer 11

form of a scatterplot of nonmetric variables where frequency within each cell is color-coded to depic relationships

Question 12

Assumptions of equal variance of the population error E

Answer 12

E is estimtaed from e

Question 13

Homoscedasticity

Answer 13

variance of the error terms (e) appears constant over a range of predictor variables

Question 14

Heteroscedasticitiy

Answer 14

error terms have increasing or modulating variance

Question 15

Histogram

Answer 15

graphical display of the distribution of a single variable

Question 16

Hoeffding’s D

Answer 16

new measure of association/correlation, based on distance measures between the variables and thus more likely to incorporate nonlinear components