Azure Machine Learning 1.      Which of the following is...

Azure Machine Learning

1.      Which of the following is false about Train Data and Test Data in Azure ML Studio ________

a.      Train and Test data random split is reproducible.

b.      Train and Test data are used to Score the Model.

c.      Cross validation data is taken from train data.

d.      Train data and Test data split should follow a thumb-rule of 80 : 20.

2.      What are the different aspects of data that need to be analyzed when understanding data

a.      Data Category

b.      Variable Type

c.      Data Type

d.      All the options

3.      The data preparation step enables which of the following

a.      Imports data.

b.      Trains and evaluates the model.

c.      Cleans missing data.

d.      Filter-based feature selection.

4.      Age can be Normalised through _________.

a.      Zscore

b.      MinMax

c.      Logistic

d.      LogNormal

5.      Which of the following is the preferred model to find the best fit line for a data set using linear regression

a.      Maximum likelihood and Logarithmic loss

b.      Logarithmic loss

c.      Maximum likelihood

d.      Least square error

6.      When predicting if the patient has cancer or not, which parameter must be given importance

a.      Threshold

b.      Precision

c.      Recall

d.      Accuracy

7.      Quality of our classifier can be determined through _________

a.      Recall

b.      Confusion Matrix

c.      Accuracy

d.      Precision

8.      What is the Azure CLI command to view the score of a model

a.      az ml service show score -n $service

b.      az ml score-n $service

c.      az ml service show -n $service

d.      None of the options

9.      An environment for working with Azure Machine Learning in an Azure subscription does not automatically create ___________

a.      A storage account

b.      A pricing plan

c.      A workspace

d.      A web service

10.  Data Visualization in Azure ML Studio is possible through ________

a.      Both inbuilt and jupyter notebook

b.      Import module and jupyter notebook

c.      Import module

d.      Jupyter Notebook

e.      In-built visualisation options

11.  What is the Azure CLI command to create machine learning workspace

a.      az ml workspace -w $workspace -g $resource

b.      Cannot create workspace using cli

c.      az workspace -w $workspace -g $resource

d.      az ml workspace create -w $workspace -g $resource

12.  A 5GB data file attained from a U-SQL job can be used for a training experiment through __________

a.      Any of the options

b.      Statistical Sampling

c.      Manual Upload

d.      Azure Blob Storage

13.  Launching Azure ML in an Azure subscription does not automatically create ___________

a.      Web service

b.      Pricing plan

c.      Workspace

d.      Storage account

14.  All our published web services can be considered as _________

a.      Azure Apps

b.      Azure Resources

c.      Azure App Services

d.      None of the options

15.  Which of the following is the best model to predict the number of earthquakes occurring at a place based on historical data

a.      None of the options

b.      Fast forest quantile regression

c.      Poisson Regression

d.      Logistic Regression

16.  The support vector machine algorithm is used for which of the following

a.      More precision

b.      Faster predictions (comprimises accuracy)

c.      Slower predictions (more accuracy)

d.      More accuracy

17.  Average amount of money spent by a user at a Supermarket can be predicted through ____________

a.      Clustering

b.      None of the options

c.      Regression

d.      Classification

18.  To prevent running Import Data module each time, we run an experiment, when data is imported from Azure Blob storage, we can use _____________

a.      Cached Results

b.      Not possible to prevent Import Data module from running.

c.      Stored Results

d.      Generated Results

19.  A client application needs to fetch reference data based on customer location and data is shared on multiple databases. Servicing the application through a single prediction web service is possible through ______________

a.      Parametrising Database Connection

b.      Any of the options

c.      Parametrising Database location

d.      Parametrising Webservice URL

20.  Predictive Experiment can be used directly after creating it from a training experiment. Select the appropriate option as answer

a.      False, a new predictive experiment must be created from scratch.

b.      False, changes must be done to facilitate proper web service inputs/outputs.

c.      True, changes can be done only for customisation of web service for client application.

d.      True, but need changes with data transformation.