Feature Selection and Extraction in Microsoft Azure

Question 1

Which techniques can be used to prepare features for machine learning?

A. Feature detection, feature extraction, and feature normalization

B. Feature selection, feature perfection, and feature normalization

C. Feature selection, feature extraction, and feature qualification

D. Feature selection, feature extraction, and feature normalization

Answer 1

D. Feature selection, feature extraction, and feature normalization

Question 2

Which is the internal data representation used by Azure ML Studio Classic?

Data Table

Data Package

Data Frame

Data Set

Answer 2

Data Table

Question 3

What is a reason to perform feature extraction?

To convert features into a scale that is usable by a ML model

To generate features from raw data to use in ML models

To select features that are optimal to use in a ML model

To calculate data points so that they can be consumed by a ML model

Answer 3

To generate features from raw data to use in ML models

Question 4

What is one of the first recommended steps for working with images?

Convert to black and white

Resize the image

Convert to png files

Extend the images to maximum resolution

Answer 4

Resize the image

Question 5

What are some clear advantages of feature normalization?

Calculates data points so that they can be consumed by a ML model

Generates features that were previously unavailable from raw data to use in ML models

Improves training time and prevents overfitting

Selects the best features to use in the ML model

Answer 5

Improves training time and prevents overfitting

Question 6

Which is the golden rule of machine learning?

Use all available features as input

Keep the scales normalized

Combine features to reduce dimensions

Keep the model simple

Answer 6

Keep the model simple

Question 7

Which module can you use to create columns for categories when encoding features?

Normalize data

Convert to indicator values

Edit metadata

Select columns in data set

Answer 7

Convert to indicator values

Question 8

What is one of the advantages of performing feature selection?

Select those data points that help create better predictions.

Ability to create new features that are retrieved from existing data.

Converting raw features into tabular data that can be used by models.

Select data points that are availalbe in a common scale.

Answer 8

Select those data points that help create better predictions.

Question 9

Which of these are all feature scoring methods?

A. Pearson Correlation, Mutual Relation, Kendall Correlation, Linear Correlation, Chi Squared, Fisher Score, and Count Based

B. Pearson Correlation, Mutual Information, Kendall Correlation, Spearman Correlation, Chi Squared, Fisher Score, and Count Based

C. Pearson Correlation, Mutual Information, Kendall Correlation Spearman Correlation, Chi Squared, Fisher Score, and Count Based

D. Pearson Correlation, Mutual Information, Kendall Correlation, Spearman Correlation, Chi Root, Fisher Score, and Count Based

Answer 9

B. Pearson Correlation, Mutual Information, Kendall Correlation, Spearman Correlation, Chi Squared, Fisher Score, and Count Based

Question 10

Which pretrained model module is used for a binary classification in the permutation feature importance demo?

Two-Class support Vector Machine module

Binary-Class support Vector Machine module

Dual-Class support Vector Machine module

Binary Decision Tree

Answer 10

Two-Class support Vector Machine module