Data Science MCQ
Data Science MCQ
3.0 Pts
In which task of the Evaluation phase of the cross-industry standard process for data mining
(CRISP-DM) methodology do you assess the degree to which the model meets the business
objectives?
Describe data
Evaluate results
Which of the following phases of the cross-industry standard process for data mining (CRISP-
DM) methodology can follow the Evaluation phase?
Note: There are 2 correct answers to this question.
Modeling
Business understanding
Data preparation
Data understanding
Deployment
Question 3
2.0 Pts
What is the name of the confusion matrix statistic that represents the Type 1 errors?
True positive
True negative
False negative
False positive
Question 4
3.0 Pts
How many responders would be selected (on the y-axis) if you analyze a gains (detected) chart
and randomly choose 40% of the total customer base (on the x-axis)? 2
33%
50%
60%
40%
Question 5
3.0 Pts
What is the lift if you select 20% of the total customer base (on the x-axis) and identify 60% of the
responders? 2
6x
2x
4x
3x
Question 6
4.0 Pts
Which of the following values for the predictive power (KI) and prediction confidence (KR) metrics
would indicate a potential problem with the model? 13 2
Note: There are 2 correct answers to this question.
Which of the following metrics is most often used to assess the performance of a regression
model with a continuous target? 16 2
Lift
You use a predictive model that predicts that 500 customers will respond if you select the top
10% of scores, 400 customers will respond with the next 10% of scores, 300 customers with the
next 10% of scores, and 200 customers with the next 10% of scores. How many responders do
you have if you select the top 3 deciles (together)? 3
900
700
1200
1400
Question 9
3.0 Pts
What is the name of the process that is used to find the subset of explanatory variables that best
explain the relationship of independent variables with the target variable?
Parameter tuning
Segmented modeling
Feature re-engineering
Feature selection
Question 10
3.0 Pts
How many times do you run the training algorithm if you use a k-fold cross-validation process
and split the data into 10 subsets?
10
5
11