Knowledge quiz

It's time to put what you've learned to the test, get 8 right to pass this week.

When you train a predictive model, you confirm that 10% of the observations are contained in the top decile. However, when you apply the model onto new data, you notice that 15% of the new observations are in the top decile. Why could this change have happened?

There are two correct answers.

There are more observations in the new data than there were in the data you used to train the model.

There is an underlying change in the population characteristics.

The model is no longer performing as you expect.

There are fewer observations in the new data than there were in the data you used to train the model.

Which of the following tasks are included in the Evaluation phase of the cross-industry standard process for data mining (CRISP-DM) methodology?

There are three correct answers.

Evaluate results

Review the process

Determine next steps

Plan the deployment

Assess the situation

What is the name of the process that is used to find the subset of explanatory variables that best explains the relationship of the explanatory variables with the target variable?

Choose the correct answer.

Feature selection

Data binning

Segmented modeling

Parameter tuning

In which task of the Evaluation phase of the cross-industry standard process for data mining (CRISP-DM) methodology do you highlight activities that have been overlooked?

Choose the correct answer.

Evaluate results

Review the process

Determine next steps

Describe data

How can you analyze the strength of a classification model that has a binary nominal target?

There are three correct answers.

Use a gains chart.

Use a lift chart.

Analyze a decile table.

Calculate the support.

Analyze the coefficient of determination.

How many responders would be selected (on the y-axis) when you analyze a gains chart random model line and randomly chose 40% of the total customer base (on the x-axis)?

Choose the correct answer.

33%

40%

50%

60%

What is the name of the confusion matrix statistic that represents the Type 2 errors?

Choose the correct answer.

True positive

False positive

True negative

False negative

What is the name of the confusion matrix statistic that represents the Type 1 errors?

Choose the correct answer.

True positive

False positive

True negative

False negative

What is specificity?

Choose the correct answer.

True positive rate

False positive rate

True negative rate

False negative rate

10.

Which metrics are most often used to assess the performance of a classification model, which has a binary nominal target?

There are three correct answers.

Type 1 and Type 2 errors

Area Under the Curve (AUC)

Lift

Mean absolute error

Mean squared error