ServiceNow Predictive Intelligence (PI) Micro-Certification • Consultant

Mock Test 1: ServiceNow Predictive Intelligence (PI) Micro-Certification – Results

Question 1: Incorrect

Which Predictive Intelligence framework is a machine learning solution that identifies existing records that have similar values to a new record?

Similarity Framework

(Correct)
Classification Framework
Regression Framework

(Incorrect)
Clustering Framework

Explanation

Similarity Framework is a machine learning solution that identifies existing records that have similar values to a new record.
Classification Framework is a machine learning solution that sets field values during record creation based on user inputs.
Clustering Framework is a machine learning solution that groups similar records into clusters so you can address them collectively or identify patterns.
Regression Framework is a machine-learning framework that you can train with historic data to predict numeric outputs.

Question 2: Correct

What is the maximum classes a Predictive Intelligence classification solution can support?

100 classes
There is no class limit

(Correct)
50 classes
25 classes

Explanation

There is no limitation on the amount of classes a classification solution can support.

Question 3: Incorrect

What is the maximum number of records that can be used for training in Predictive Intelligence?

100,000

(Incorrect)
10,000
300,000

(Correct)
30,000

Explanation

The recommended number of records for training a good solution is from 30,000 through 300,000. If you submit more than 300,000 records, the most recent 300,000 records are used to train the solution. Use only authentic records from the database.
The minimum number of records required for classification and regression solution training is set at 10,000.
The number of records the Similarity Window can retrieve for lookup results is limited to 100,000.

Question 4: Incorrect

Which of the following ML solution parameters in Predictive Intelligence application are in the clustering parameter category?

Threshold

(Correct)
TF-IDF
Epsilon

(Correct)
Minimum Neighbors

(Correct)

Explanation

Here are the ML solution parameters that are in the clustering parameter category: Epsilon, Minimum Neighbors, Threshold, and Max-time-window-for-update.

Question 5: Incorrect

(P) What prediction output statistics appear in the regression solution after running the test in the Predictive Intelligence application? Select 3 Answers from the below options.

Point Estimate

(Correct)
Middle Bound
Range Estimate

(Incorrect)
Upper Bound

(Correct)
Lower Bound

(Correct)

Explanation

After you run the test within the Regression solution, the following prediction output statistics appear:

Point Estimate, Lower Bound and Upper Bound.

The Point Estimate on the screen is a single value at one point in time. The Lower and Upper Bounds on the screen signify a range accuracy value.

Question 6: Correct

What is ‘Cluster Quality’ in the cluster summary form in Predictive Intelligence application?

The percentage of how similar are the clustered records

(Correct)
The percentage of records that are clustered
A summary of changes to clusters for each clusters update interval
The number of records in the current cluster

Explanation

Cluster Quality field in the cluster summary form indicates the percentage of how similar are the clustered records.
Cluster size in the cluster summary form is the number of records in the current cluster.
Clusters Updates is a table that stores a summary of changes to clusters for each clusters update interval.
Solution coverage in clustering solution statistics is the percentage of records that are clustered.

Question 7: Incorrect

What system property does need to be set to true if you have already implemented Predictive Intelligence and want to save your trained solutions?

glide.platform_ml.artifact.cache_compression_scheme
glide.platform_ml.clone_artifacts

(Correct)
glide.platform_ml.override_training_lock
glide.platform_ml.use_new_schema

(Incorrect)

Explanation

If you have already implemented Predictive Intelligence and want to save your trained solutions, set the glide.platform_ml.clone_artifacts system property to true.

Question 8: Correct

Which of the following can reduce the solution’s accuracy or reduce the number of records where Predictive Intelligence can apply the solution? Select 3 Answers from the below options.

Old process results

(Correct)
Short training frequencies
Too many choices

(Correct)
Inaccurate selections

(Correct)

Explanation

The end goal is to maximise the accuracy of the model against the most useful data. Too many choices, inaccurate selections, or old process results can reduce the solution’s accuracy or reduce the number of records where PI can apply the solution. Setting short training frequencies has no negative impact.

Question 9: Incorrect

(P) What are the benefits of the Predictive Intelligence Workbench? Select 3 Answers from the below options.

To get started with machine learning-based use cases

(Correct)
To run intelligent business processes

(Incorrect)
To determine your business value

(Correct)
To monitor your models

(Correct)

Explanation

Predictive Intelligence Workbench provides pre-built use case templates and guides you through your machine-learning implementation to create intelligent business processes. Business process architects use Predictive Intelligence Workbench functionality to get started with machine learning-based use cases. You can determine your business value, monitor your models, and view related statistics using the Predictive Intelligence Workbench dashboards.

Question 10: Correct

What target metrics can be configured for a trained classification solution in Predictive Intelligence?

Precision

(Correct)
Recall

(Correct)
Distribution
Coverage

(Correct)

Explanation

Predictive Intelligence provides three classification metric types: precision, coverage, and recall. You configure these metrics on the Solution Statistics tab of a trained classification solution form.

Question 11: Incorrect

Which of the following is a configuration of a Machine Learning solution which defines what is the outcome/output field for a solution, what data is used for training this solution and what is the input field(s) used for training?

Precision
Word Corpus

(Incorrect)
Coverage
Solution Definition

(Correct)

Explanation

Solution Definition is a configuration of an Machine Learning solution which defines what is the outcome/output field for a solution, what data is used for training this solution and what is the input field(s) used for training.
Word Corpus is a collection of words and phrases that functions as the vocabulary (dictionary) the system uses to compare your instance records based on their textual similarity.
Precision is the percentage of predictions that are correct for an ML solution.
Coverage is the percentage of incoming records that are predicted using an ML solution.

Question 12: Incorrect

Which statement best describes the use of update sets in Predictive Intelligence?

Update sets are used to track solution versions
Update sets are used to export solutions between instances

(Correct)
Update sets are used to store solution definitions
Update sets are used to move training data sets

(Incorrect)

Explanation

Update sets are used to export Predictive Intelligence solutions between instances. You can use update sets to export the changes to your production instance. This practice mitigates the risk of retraining solutions on your live production instance.

Question 13: Correct

What is the minimum role required to view Predictive Intelligence dashboards?

pa_data_collector
ml_report_user

(Correct)
pa_viewer
ml_admin

Explanation

ml_report_user is the minimum role required to view Predictive Intelligence dashboards. ml_admin role also manages the Predictive Intelligence application. pa_data_collector and pa_viewer roles are performance analytics roles and not specific to Predictive Intelligence.

Question 14: Incorrect

On what instance does Predictive Intelligence solution training take place?

Customer’s development instance
Customer’s production instance

(Incorrect)
Data centre training server

(Correct)
Data centre Prediction Server

Explanation

To train a solution, you export its solution definition and associated records to a centralised training server within the nearest ServiceNow data centre. When the training completes, the training server exports the solution back to your instance and deletes all of your training data from the server.

Question 15: Incorrect

Which statement is NOT true about word corpuses in Predictive Intelligence application?

PI uses word corpuses to predict field values

(Correct)
Word corpuses can use text from multiple tables
Word corpuses can be used in more than one solution
Similarity and Clustering solutions use a word corpus

(Incorrect)

Explanation

You can (re)use word corpus with your similarity and clustering solutions and apply it to various use cases. You can use a word corpus to help compare similar record text in a table or across multiple tables. A trained word corpus can be reused across multiple solutions and capabilities.

Continue

Settings

Report abuse

Fullscreen

Mock Test 2: ServiceNow Predictive Intelligence (PI) Micro-Certification

Question 1:

Skipped

Which of the following is the definition of the Classification Framework in Predictive Intelligence?

The output field values that the predictive model uses to make predictions through classification solutions
A trained solution which contains the underlying predictive models, that is generated when a solution goes through the training process
A machine learning solution that sets field values during record creation based on user inputs

(Correct)
The act of an instance processing given information, based on the parameters within the solution definition, to build its predictive model or word corpus

Explanation

Classification Framework is a machine learning solution that sets field values during record creation based on user inputs. For example, the incident category can be set based on the Short Description.
Class is the output field values that the predictive model uses to make predictions through classification solutions.
Solution is a trained solution which contains the underlying predictive models, that is generated when a solution goes through the training process.
Training is the act of an instance processing given information, based on the parameters within the solution definition, to build its predictive model or word corpus.

Question 2:

Skipped

(P) What ServiceNow application does provide pre-built use case templates and guide you through your machine-learning implementation to create intelligent business processes?

Predictive Intelligence Workbench

(Correct)
Natural Language Understanding
Predictive Intelligence
Performance Analytics

Explanation

Predictive Intelligence Workbench provides pre-built use case templates and guides you through your machine-learning implementation to create intelligent business processes.

Question 3:

Skipped

To find similar records, which type of Predictive Intelligence solution should be created?

Classification
Clustering
Similarity

(Correct)
Regression

Explanation

Similarity solutions are used to find similar records.

Question 4:

Skipped

What are the different ways to initiate the training process in Predictive Intelligence application? Select 4 Answers from the below options.

Train on local servers
Activate and Train

(Correct)
Update and Retrain

(Correct)
Submit and Train

(Correct)
Train from PI Homepage

(Correct)

Explanation

There are four ways to initiate the training process in ServiceNow: Activate and Train, Submit and Train, Train from PI Homepage, and Update and Retrain.

Question 5:

Skipped

What product lines are using Predictive Intelligence? Select three correct answers.

CSM

(Correct)
ITAM
ITSM

(Correct)
HR

(Correct)
Sec Ops

Explanation

Predictive Intelligence brings the power of AI to applications in IT Service Management (ITSM), HR Service Delivery (HR), Customer Service Management (CSM), Event Management (EM), and custom applications by providing prebuilt solutions and end-to-end analytics for delivering continuous improvement.

Question 6:

Skipped

Which of the following solution definition is an available similarity type template where both Predictive Intelligence and IT Service Management are active?

CSM Case Assignment
Major Incident Detection

(Correct)
HR Case Prioritization
Incident Categorization

Explanation

Major Incident Detection is an available similarity type solution definition where Predictive Intelligence and IT Service Management are active.

HR Case Prioritization, CSM Case Assignment, and Incident Categorization are classification type solution definitions.

Question 7:

Skipped

Word corpuses can be created from multiple tables and can be utilized in both the ———- and ———- solutions. Select 2 Answers from the below options.

Clustering

(Correct)
Regression
Classification
Similarity

(Correct)

Explanation

Word corpuses can be created from multiple tables and can be utilized in both the similarity and clustering solutions. Before you can create a similarity or clustering solution, you need to create a word corpus.

Question 8:

Skipped

(P) Which of the following resulting model statistics for the Predictive Intelligence Regression solutions does indicate the percentage of actual values between a predicted range?

Range Accuracy

(Correct)
Mean Absolute Error (MAE)
Average Interval Width
Symmetric Mean Absolute Percentage Error (SMAPE)

Explanation

Range Accuracy is the percentage of actual values between a predicted range. In other words, it’s the range between the upper and lower bounds of the prediction. For example, if four out of five actuals lie within the predicted range, the range accuracy is 80%. Mean Absolute Error (MAE) which measures the mean deviation of a predicted value from the actual value. Symmetric Mean Absolute Percentage Error (SMAPE) is a percentage value of the deviation from the predicted to the actual. Average Interval Width is the difference between the upper and lower bounds of the prediction.

Question 9:

Skipped

Which statement best describes what happens when a Predictive Intelligence solution definition is scheduled to train on an interval?

Administrators can schedule when the newly trained solution goes active
Newly trained solutions are automatically activated

(Correct)
Newly trained solutions must be manually activated
If the ML engine does not detect changes, a new solution is not created

Explanation

The system automatically activates the most recent version of the solution when it completes training a solution, and only allows one solution version to be active at a time. However, you can manually activate any previously trained solution version you want Predictive Intelligence to use to make predictions. The system creates a solution version each time you train a solution definition.

Question 10:

Skipped

What type of server does remove the 50 class limit for the classification framework and allow ServiceNow to support more complex machine learning models (like NLU and Clustering)?

MID server
Proxy server
Trainer server
Centralized prediction server

(Correct)

Explanation

The centralized prediction servers remove the 50 class limit for the classification framework and allow ServiceNow to support more complex machine learning models (like NLU and Clustering).
Trainer server generates a model based on the data that you defined in the solution definition.

Question 11:

Skipped

Which types of fields are recommended to be used as outputs for classification solutions in Predictive Intelligence? Select 2 Answers from the below options.

Reference
Script
String

(Correct)
Choice

(Correct)

Explanation

You can select any field available on the table. However, ServiceNow recommends using a string or a choice field with a finite set of possible values.

Question 12:

Skipped

How many closed incidents with populated category field is needed to build a viable solution that can predict the Category field of an incident?

3,000 to 30,000 records
30 to 90 days of data
3 to 30 months of data
30,000 to 300,000 records

(Correct)

Explanation

ServiceNow recommends that you have at least 30,000 to 300,000 historical records from which PI can learn. For example, if you want to build a solution that can predict the Category field of an incident, you need this many closed incidents which have the category field populated. The minimum number of records required for classification solution training is set at 10,000.

Question 13:

Skipped

Which of the following is a key Classification Framework table in Predictive Intelligence application?

Precision Coverage Lookup (ml_pc_lookup)

(Correct)
Similarity Example (ml_similarity_example)
Cluster Summary (ml_cluster_summary)
Regression Definition (ml_capability_definition_regression)

Explanation

Precision Coverage Lookup (ml_pc_lookup), Class Confidence (ml_class), and Excluded Classes (ml_excluded_classes) are the key Classification Framework tables in Predictive Intelligence application. Similarity Example (ml_similarity_example) is a Similarity Framework table. Cluster Summary (ml_cluster_summary) and Cluster Update (ml_cluster_sol_update) are Clustering Framework tables. Precision Coverage Lookup table stores a list of precision/coverage pairs for each class used in predictions and is primarily used in the classification framework.

Question 14:

Skipped

What is the goal of the clustering framework in Predictive Intelligence?

Identify existing records that have similar values to a new record
Set field values based on text from specific fields
Group similar records

(Correct)
Understand the meaning of words and their context

Explanation

You can group similar records into clusters so you can address them collectively or identify patterns in a clustering solution.

Question 15:

Skipped

Which statements are true about language processing in Predictive Intelligence application? Select 3 Answers from the below options.

If the dominant language is English, the Processing Language field can be set to –None–
PI supports the following languages: English, French, German, Japanese, Dutch, Spanish, Italian, and Brazilian Portuguese

(Correct)
PI processes the data set in English by default

(Correct)
Language support is available for the classification, similarity, regression, and clustering solutions

(Correct)
Additional languages can be added by activating additional language choices

Fullscreen

Fullscre

Fullscreen

Mock Test 3: ServiceNow Predictive Intelligence (PI) Micro-Certification – Results

Question 1:

Skipped

What is ‘Similarity Score Threshold’ in Predictive Intelligence?

The score that identifies the anomalies in the similarity solution
The score that records must exceed in order to be returned by the solution

(Correct)
The level of prediction confidence in similarity solutions
The degree of similarity to the text of the test record

Explanation

Similarity Score Threshold is the score that records must exceed in order to be returned by the solution.

Question 2:

Skipped

Which of the following solution definitions are the available similarity type templates where both Predictive Intelligence and Customer Service Management are active? Select 3 Answers from the below options.

Major Issue Detector

(Correct)
CSM Case Prioritization
Recommended Resolved Cases

(Correct)
Recommended Open Cases

(Correct)

Explanation

Here are the solution definitions that are the available similarity type templates where both Predictive Intelligence and Customer Service Management are active: Major Issue Detector, Recommended Resolved Cases and Recommended Open Cases. CSM Case Prioritization is a classification type solution definition.

Question 3:

Skipped

What details need to be defined for each Word Corpus Content in Predictive Intelligence application?

Type
Table

(Correct)
Input
Field List

(Correct)
Output

Explanation

Table and Field List are the available mandatory fields in the Word Corpus Content form (i.e. [ml_word_vector_corpus_details] table). Input, Output and Type do not exist in this table.

Question 4:

Skipped

What implementation steps do you follow when you installed Predictive Intelligence plugin in you sup-production? Select 3 Answers from the below options.

Set retraining frequency
Create solution definition

(Correct)
Clone production to sub-prod
Submit train

(Correct)
Validate/tune solution

(Correct)

Explanation

Once you installed Predictive Intelligence plugin in you sup-production, you should follow the these implementation steps: Create solution definition, Submit train and Validate/tune solution. You clone production to sub-prod before installing Predictive Intelligence plugin. You normally set retraining frequency in Production, once you validated/tuned your solution in Production.

Question 5:

Skipped

A Class record in Predictive Intelligence is a record for each ———-.

Input field value
Output field value

(Correct)
Mandatory field on the form
Field on the solution definition table

Explanation

The solution classes are the output field values for which the model can make predictions. Each class is an output field value with a list of possible precision, coverage, and distribution metrics to choose from.

Question 6:

Skipped

What is the purpose of the prediction server for Predictive Intelligence?

Store ML solutions and training data set
Schedule training for ML solutions
Train ML solutions
Generate predictions for ML solutions

(Correct)

Explanation

Predictions occur on a centralised prediction server within the same data centre as the instance. The trained model artefacts are sent from the instance server to the prediction server when the prediction is invoked for the first time. After that, the trained model artefacts are cached on the prediction server for subsequent predictions.

Question 7:

Skipped

Which statements are true regarding training solutions in Predictive Intelligence? Select 2 Answers from the below options.

Retrained solutions are more accurate than older versions of the solution
Of all of your data is retrained on the training server
Solutions are exported from the training servers back to your ServiceNow instance

(Correct)
Old versions of the solution are inactivated

(Correct)

Explanation

When a solution is trained or retrained, older versions of the solution are set to inactive. After the solution is trained, it is exported back to your ServiceNow instance from the training server.

Question 8:

Skipped

What are the most important aspects of configuring solution definitions in Predictive Intelligence? Select 3 Answers from the below options.

Configuring advanced settings on all ML solutions
Choosing the right data filters

(Correct)
Choosing the right refresh frequency

(Correct)
Choosing the right input fields

(Correct)

Explanation

Each solution follows a similar process for configuration. The most important aspects of configuring solution definitions are: Choosing the right data filters, Choosing the right input fields, Choosing the right refresh frequency, and Determining the need for advanced settings.
Configuring advanced settings on your ML solutions is optional and without any additional configurations, ServiceNow will use the default algorithm(s) to train your models.

Question 9:

Skipped

What is the maximum number of classes that a classification model can support in Predictive Intelligence?

25
No fixed class limit

(Correct)
100
50

Explanation

There is no limitation on the amount of classes a classification solution can support.

Question 10:

Skipped

(P) Which of the following statistics are delivered by the resulting model of the Regression solutions in Predictive Intelligence? Select 3 Answers from the below options.

Mean Absolute Error (MAE)

(Correct)
Estimated Solution Precision
Average Interval Width

(Correct)
Range Accuracy

(Correct)

Explanation

Regression solutions enable you to predict a point estimate and prediction interval. The resulting model delivers the following statistics: Mean Absolute Error (MAE), Symmetric Mean Absolute Percentage Error (SMAPE), Range Accuracy, and Average Interval Width. ‘Estimated Solution Precision’ is used in the Classification solutions.

Question 11:

Skipped

To predict field values, which type of Predictive Intelligence solution should you create?

Classification

(Correct)
Regression
Similarity
Clustering

Explanation

Classification solutions are used to predict field values. It is also used to leverage the power of machine learning and historical data to categorize tasks, incidents, and cases at scale.
Similarity solutions are used to surface similar tasks and content to predict new major issues and recommend critical actions.
Clustering is used to group similar records into clusters so you can address them collectively or identify patterns.
Regression is a machine-learning framework that you can train with historic data to predict numeric outputs.

Question 12:

Skipped

Which statements are true regarding the clustering solutions in Predictive Intelligence? Select 2 Answers from the below options.

Clustering solutions groups data based on similarity

(Correct)
Clustering solutions support multiple languages
Clustering solutions generate classes
Clustering solutions can reuse an existing word corpus

(Correct)

Explanation

Clustering solutions can reuse an existing word corpus and group data based on how similar they are.

Question 13:

Skipped

What solution parameter does enable you to steer a solution’s training to bias a specific class in Predictive Intelligence?

DBSCAN
Class-Recall

(Correct)
TF-IDF
Threshold

Explanation

The class recall solution parameter enables you to steer a solution’s training to bias a specific class. The parameter relies on the recall metric, which indicates how accurately a model can predict a particular class/value correctly.
TF-IDF is used to evaluate how ‘important’ a word is to a document in the document set or word corpus.
DBSCAN algo is used to group together points that are close to each other based on a distance measurement and a minimum number of points.
Threshold is a clustering parameter that is used to target solution coverage.

Question 14:

Skipped

What does the size of the bubbles in the solution visualization indicate when you access the bubble chart by opening the solution record in Predictive Intelligence application?

Estimated Precision
Estimated Recall
Distribution

(Correct)
Estimated Coverage

Explanation

Each solution has a bubble chart that shows the estimated precision and coverage for each of the classes covered by the solution. The size of the bubble indicates the % of records (distribution) that belong to the class. Access the bubble chart by opening the solution record and clicking Solution Visualization.

Question 15:

Skipped

(P) Which type of Predictive Intelligence solution do you use to estimate the time it takes to resolve an incident or a case?

Classification
Regression

(Correct)
Similarity
Clustering

Explanation

Regression solutions enable you to predict a point estimate and prediction interval. For example, you can use regression to estimate the time it takes to resolve an incident or a case.

Mock Test 4: ServiceNow Predictive Intelligence (PI) Micro-Certification – Results

Question 1:

Skipped

How does the training server process and prepare the Predictive Intelligence input data to generate the solutions? Select 3 Answers from the below options.

By removing stop words

(Correct)
By removing any problematic characters

(Correct)
By adding context to the input data
By identifying and indexing duplicate records

(Correct)

Explanation

The training server processes and prepares the Predictive Intelligence input data to generate the solutions by: removing any problematic characters, identifying and indexing duplicate records, tokenizing and breaking text input into n-grams, removing stop words, and stemming.

Question 2:

Skipped

(P) What encoding is recommended to be applied to your classification, similarity, and regression solutions to return better prediction results for records that have machine-generated content, such as alerts and error messages in log files?

Density-based spatial clustering of applications with noise (DBSCAN)
Term Frequency-Inverse Document Frequency (TF-IDF)

(Correct)
Levenshtein Distance (LD)
Class-Recall

Explanation

Apply optional TF-IDF encoding to your classification, similarity, and regression solutions. Predictive Intelligence uses paragraph vector word embedding by default in these solutions, which is highly effective for processing data that contains primarily human readable content. However, TF-IDF sometimes returns better prediction results for records that have machine-generated content, such as alerts and error messages in log files.

Question 3:

Skipped

If you want to recluster your records periodically, which field should you update in Predictive Intelligence application?

Cluster Update
Update Frequency
Training Request Schedule
Training Frequency

(Correct)

Explanation

With ‘Training Frequency’, you can select how often you want the system to discard all previous cluster results and recreate clusters from the beginning. With ‘Update Frequency’, you can select how often you want the system to update your clusters with new and updated records.

Question 4:

Skipped

Predictive Intelligence is available to customers who

Have on-premise solutions
Customers who do not have a PI subscription
Have more than 1 million records
Use the ServiceNow cloud solution

(Correct)

Explanation

ServiceNow Predictive Intelligence is a platform function that empowers features and capabilities across applications which means it is not specific to customers who use its cloud solution.

Predictive Intelligence is available to customers who use the ServiceNow cloud solution.

It can also be made available to on-premise customers if they coordinate it with their account manager.

Question 5:

Skipped

In ServiceNow, Predictive Intelligence will determine how to: Select 3 Answers from the below options.

Recommend actions and content

(Correct)
Categorise

(Correct)
Create new tickets
Discover hidden patterns

(Correct)

Explanation

In ServiceNow, Predictive Intelligence will determine how to categorise, recommend actions and content, and discover hidden patterns. PI analyses thousands of historical records (i.e., incidents, cases, requests) and applies the correct categorisation, priority, and routing information. It suggests relevant tasks and content to agents working on tickets. And, it continuously segments and groups similar items so you can identify and highlight areas of improvement. Predictive Intelligence will not determine how to create new tickets.

Question 6:

Skipped

What are the three scores you can find in your trained models as you are evaluating and tuning them? Select 3 Answers from the below options.

Regression(%)
Coverage(%)

(Correct)
Net Automation(%)

(Correct)
Precision(%)

(Correct)

Explanation

Here are the three scores you can find in your trained models as you are evaluating and tuning them:

* Coverage(%),

* Precision(%) and

* Net Automation(%).

Question 7:

Skipped

Why would a class be added to the exclusion list for classification solutions in Predictive Intelligence?

The class is inactive or not used anymore
There are less than 30 records with that specific class

(Correct)
The class was excluded from the data set
There is not enough variation of the data for that specific class

Explanation

Classes that have less than 30 records in your training dataset are excluded from solution training.

Question 8:

Skipped

What does ‘Update Frequency’ option in the similarity solution definition form in Predictive Intelligence application indicate?

How often you want the system to discard all previous cluster results and recreate clusters from the beginning
How often the system regenerates the solution based on records matching the filter
How often you want the system to group new and updated records into clusters
How often you want to refresh the data that you use to retrieve your similarity results

(Correct)

Explanation

Update Frequency’ option in the similarity solution definition form indicates how often you want to refresh the data that you use to retrieve your similarity results.

Question 9:

Skipped

To surface similar tasks and content to predict new major issues and recommend critical actions, which type of Predictive Intelligence solution should you create?

Regression
Similarity

(Correct)
Classification
Clustering

Explanation

Similarity solutions are used to surface similar tasks and content to predict new major issues and recommend critical actions.
Classification is used to leverage the power of machine learning and historical data to categorize tasks, incidents, and cases at scale.
Clustering is used to group similar records into clusters so you can address them collectively or identify patterns.
Regression is a machine-learning framework that you can train with historic data to predict numeric outputs.

Question 10:

Skipped

Predictive Intelligence offers out-of-the-box solutions for which applications? Select 4 Answers from the below options.

IT Service Management

(Correct)
Customer Service Management

(Correct)
Event Management

(Correct)
Governance, Risk, and Compliance
HR Services

(Correct)
IT Asset Management

Explanation

Predictive Intelligence offers out-of-the-box solutions for ITSM, CSM, HR, and Event Management.

Question 11:

Skipped

What Predictive Intelligence table does store a list of solutions which are produced after the solution is processed on ServiceNow’s training servers?

Advanced Solution Setting (ml_advanced_solution_settings)
ML Solution (ml_solution)

(Correct)
Training Request Schedule (ml_training_request_schedule)
ML Solution Definition (ml_capability_definition_classification)

Explanation

ML Solution (ml_solution) table stored a list of solutions which are produced after the solution is processed on ServiceNow’s training servers.
ML Solution Definition (ml_capability_definition_classification) table stores the data records used to train the classification, similarity, and clustering solutions.
Training Request Schedule (ml_training_request_schedule) table stores the schedule based on the selected training frequency.
Advanced Solution Setting (ml_advanced_solution_settings) table stores a list of parameters that PI will use to invoke other ML technologies and algorithms to produce a solution.

Question 12:

Skipped

What is the purpose of a word corpus in Predictive Intelligence?

The word corpus is the vocabulary the system uses to compare textual similarity between records

(Correct)
The word corpus identifies similar existing records that have similar values to a new record
The word corpus is the output field values that the predictive model uses to make predictions through classification solutions
The word corpus is the subset of records used to train the similarity solution

Explanation

The word corpus functions as the vocabulary the system uses to compare textual similarity between records.

Question 13:

Skipped

What metric scores are being tracked in the classification solution statistics after your Predictive Intelligence solution is trained? Select 3 Answers from the below options.

Estimated Solution Precision

(Correct)
Estimated Solution Distribution
Estimated Solution Recall

(Correct)
Estimated Solution Coverage

(Correct)

Explanation

Following target metrics are being tracked in the classification solution statistics after your Predictive Intelligence solution is trained: Estimated Solution Precision, Estimated Solution Recall and Estimated Solution Coverage.
Estimated Solution Precision is the estimated percentage of records that will be correct for this solution.
Estimated Solution Recall is the estimated percentage of true positives.
Estimated Solution Coverage is the estimated percentage of records that will receive a prediction.
Estimated Solution Distribution does not exist.

Question 14:

Skipped

What field in clustering solution definition is used to choose fields from your table that can help the system identify the class which is most frequent in the cluster in Predictive Intelligence application?

Fields (fields)
Purity Fields (purity_fields)

(Correct)
Word Corpus (word_vector_corpus)
Group By (segmentation_field)

Explanation

Purity Fields (purity_fields) is used to choose fields from your table that can help the system identify the class which is most frequent in the cluster.

Question 15:

Skipped

(P) What type of Predictive Intelligence solution definition do you create and train to predict the amount of time it takes to restore a cloud database?

Regression

(Correct)
Clustering
Similarity
Classification

Explanation

Regression solutions enable you to predict a point estimate and prediction interval. You create and train a regression solution definition to predict the amount of time it takes to restore a cloud database.

Additional Test: ServiceNow Predictive Intelligence (PI) Micro-Certification – Results

Question 1:

Skipped

Predictive Intelligence adds a class to the list when the output value is assigned to ———- or more records (in your data set).

10
1
30

(Correct)
3

Explanation

Predictive Intelligence generates a class list, which is all the possible values for a specific output field. This list is based on your data set. A class is added to the list when the output value is assigned to 30 or more records (in your data set).

Question 2:

Skipped

Which types of fields are recommended to be used as inputs for training Predictive Intelligence solutions? Select 3 Answers from the below options.

Choice

(Correct)
String

(Correct)
Script
Reference

(Correct)
HTML

Explanation

The system will allow you to select any field available on the table. However, ServiceNow recommends using choice, string, or reference fields.

Question 3:

Skipped

For classification solutions in Predictive Intelligence, how many possible outputs can a single solution include?

There is no output limit
10
1

(Correct)
5

Explanation

A solution can only have one output.

Question 4:

Skipped

What plugin do you need to activate in your instance before using Predictive Intelligence?

Predictive Intelligence – Enhanced UI [com.snc.ml_ui]
Predictive Intelligence Reports [com.glide.platform_ml_pa]
Predictive Intelligence Workbench [com.sn_piwb_ml]
Predictive Intelligence [com.glide.platform_ml]

(Correct)

Explanation

The Predictive Intelligence plugin [com.glide.platform_ml] requires activation through the HI Service Portal. If you are implementing PI for Customer Service Management, Event Management, HR Service Delivery, or Knowledge Management, these plugins must be installed to use the prebuilt templates.

Question 5:

Skipped

Which statements are true about similarity solutions in Predictive Intelligence? Select 3 Answers from the below options.

Learn unstructured language
PI groups records based on textual similarity

(Correct)
PI can compare different record types (i.e. incident vs. knowledge)

(Correct)
Recommends similar/relevant records

(Correct)
PI can do a comparison between small and large text

Explanation

With similarity solution, you can create and train a machine-learning solution to collect and compare your existing records to new similar records. For example, you can compare the Short Description in an Incident record to a Knowledge article to find similarities. Similarity solution uses word corpus to compare your instance records based on their textual similarity. Natural Language Understanding (NLU) can learn unstructured language but not Predictive Intelligence.

Question 6:

Skipped

What is the minimum number of records per cluster when you are creating and training a clustering solution in Predictive Intelligence?

2

(Correct)
1
10
100

Explanation

The minimum number of records per cluster when you are creating and training a clustering solution must be greater than or equal to 2.

Question 7:

Skipped

Which is NOT a framework of Predictive Intelligence?

Classification
Clustering
Similarity
Natural Language Understanding

(Correct)

Explanation

Predictive Intelligence frameworks are Classification, Similarity, Clustering, and Regression.

Question 8:

Skipped

What is the minimum number of records needed to create a Predictive Intelligence solution?

50k
75k
100k
30k

(Correct)

Explanation

The recommended number of records for training a good solution is from 30,000 through 300,000.

Question 9:

Skipped

What is Word Corpus in Predictive Intelligence?

A collection of words and phrases that functions as the vocabulary (dictionary) the system uses to compare your instance records based on their textual similarity

(Correct)
The output field values that the predictive model uses to make predictions through classification solutions
A trained solution which contains the underlying predictive models, that is generated when a solution goes through the training process
The frequency in which the solution is refreshed/re-trained (bring in new data).

Explanation

Word Corpus is a collection of words and phrases that functions as the vocabulary (dictionary) the system uses to compare your instance records based on their textual similarity.
Solution is a trained solution which contains the underlying predictive models, that is generated when a solution goes through the training process.
Class is the output field values that the predictive model uses to make predictions through classification solutions.
Training Frequency is the frequency in which the solution is refreshed/re-trained (bring in new data).

Question 10:

Skipped

(P) Which of the following resulting model statistics for the Predictive Intelligence Regression solutions does indicate the difference between the upper and lower bounds of the prediction?

Mean Absolute Error (MAE)
Average Interval Width

(Correct)
Range Accuracy
Symmetric Mean Absolute Percentage Error (SMAPE)

Explanation

Average Interval Width is the difference between the upper and lower bounds of the prediction. This metric explains the interval informativeness. The smaller the average width, the better the model performance. Range Accuracy is the percentage of actual values between a predicted range. Mean Absolute Error (MAE) which measures the mean deviation of a predicted value from the actual value. Symmetric Mean Absolute Percentage Error (SMAPE) is a percentage value of the deviation from the predicted to the actual.

Question 11:

Skipped

Which of the following REST APIs for Predictive Intelligence can retrieve records meeting the filter conditions for a specific solution?

GetRecordsForSolutionDefintion (GET)

(Correct)
GetOOBSolutionDefinitions (GET)
Prediction for multiple solutions (GET)
Prediction (GET)

Explanation

GetRecordsForSolutionDefintion (GET) API can retrieve records meeting the filter conditions for a specific solution.
GetOOBSolutionDefinitions (GET) API can get a list of solution definitions that are out of the box.
Prediction for multiple solutions (GET) API can get predictions for multiple solutions.
Prediction (GET) API can predict an output field value using a specific solution.

Question 12:

Skipped

At a minimum, which role is required to manage Predictive Intelligence solutions?

ml_report_user
security_admin
admin
ml_admin

(Correct)

Explanation

At a minimum, users need the ml_admin role to manage Predictive Intelligence solutions.

Question 13:

Skipped

(P) What field can be used to train and maintain a Predictive Intelligence classification solution that covers more than one data area, such as geographical location or domain?

Class (sys_class_name)
Group-By Field (groupby_field)

(Correct)
Fields (fields)
Output Field (output_field)

Explanation

Use APIs to simultaneously submit multiple classification solutions for training based on the Group By field. You can use the optional Group By capability to train and maintain one classification solution covering more than one data area, such as geographical location or domain.

Question 14:

Skipped

Which of the following are the benefits of the ServiceNow Predictive Intelligence? Select 3 Answers from the below options.

Quicker major issue detection

(Correct)
Optimise resources and reduce costs

(Correct)
Increase resolution times and errors
Increase fulfiller productivity

(Correct)

Explanation

PI provides agents and organizations with many benefits such as Increase Fulfiller Productivity, Optimise Resources and Reduce Costs, Improves Overall Business Efficiency, Improves Service and Customer Satisfaction, and Quicker Major Issue Detection.
Reducing assignment error rates enables resolution and fulfilment teams to increase productivity by focusing on the work that needs to be done, resulting in less downtime for customers.
Organisations can reduce the number of required interactions by correctly assigning manual tasks and automating tasks when possible.
Fulfillers can quickly find similar open issues and determine if escalation is required.

Question 15:

Skipped

To group similar records into clusters so you can address them collectively or identify patterns, which type of Predictive Intelligence solution should be created?

Regression
Similarity
Clustering

(Correct)
Classification

Explanation

Clustering solutions are used to group similar records into clusters so you can address them collectively or identify patterns.
Similarity solutions are used to surface similar tasks and content to predict new major issues and recommend critical actions.
Classification is used to leverage the power of machine learning and historical data to categorize tasks, incidents, and cases at scale.
Regression is a machine-learning framework that you can train with historic data to predict numeric outputs.

Continue

Fullscreen