1.
Pega Customer Decision Hub™ uses the P*C*V*L arbitration formula to select the
next best action for each customer. Which factor in the arbitration formula does AI
calculate?
A)Action value
B)Context weighing
C)Propensity
D)Business levers
2.To configure an adaptive model, you identify the responses that indicate specific
customer behavior. What types of behavior do you specify?
A)Positive behavior only
B)Positive, neutral, and negative behavior
C)Positive and negative behavior
D)Any behavior
3.Configuring an adaptive model involves selecting the potential predictors. What
is the best number of potential predictors for an adaptive model?
A)As many available fields that can contribute to the predictive power of the
model.
B)Up to 100 fields to limit the impact on model speed.
C)At least 100 fields to reach an acceptable level of model performance.
D)All fields that have been predictive in the past.
Why is leaving the default settings of an adaptive model unchanged important?
A)Adaptive Decision Manager adjusts the settings automatically adjusted during
adaptive learning.
B)Using the default settings is based on best practices.
C)Changing the settings is not a task for a data scientist.
D)Changing the settings does not effect on model performance.
You can use various data types in adaptive analytics. Some of these types require
preprocessing before you can use them as potential predictors, while others are
ready to use directly. Which two data types can you use without preprocessing?
(Choose Two)
A)Text data, such as Twitter messages.
B)Dates with absolute time and date values, such as birthdays.
C)Symbolic data with up to 200 distinct values, such as products that the customer
bought previously.
D)Numeric data, such as customer age and income.
E)Event stream data, such as recent transactions.
What is the purpose of shadow mode when you deploy a new model?
A)To test the new model with a sample of data.
B)To drive the prediction by both the active model and the new model.
C)To monitor the performance of the new model in a production environment before
you deploy it as an active model.
D)To deploy the new model immediately as an active model.
The system continuously monitors the predicting power of every predictor. When does
the system activate a previously dormant predictor?
A)When the predictive power of the predictor drops below a threshold.
B)When the data scientist activates the predictor.
C)When the predictive power of the predictor improves above a threshold.
D)When the number of responses improves above a threshold.
When predictors have a similar predicting performance, they are automatically
grouped. What happens when predictors are grouped?
A)All predictors are used by the model.
B)One predictor is used by the model, the others are deleted.
C)The predictors are merged into one.
D)One predictor is used by the model; the others are deactivated.
To confirm the continuing accuracy of your adaptive models, adaptive models must be
regularly inspected. Which two tasks are part of a regular inspection? (Choose Two)
A)Update the models.
B)Adjust the advanced settings.
C)Check the performance and success rate of the models.
D)Add the historical data collected since the last inspection.
E)Check the performance of individual predictors.
With the open-source GitHub repository Pega Data Scientist Tools, you can
______________.
A)build XGBoost models
B)upload predictive models to Prediction Studio
C)visualize the performance of a predictor
D)analyze your data in SQL
You enable the capture of historical data in an adaptive model. Which two data
elements does the system capture for every customer interaction? (Choose Two)
A)The propensity the model generates.
B)The model metadata.
C)The value of all predictors.
D)The outcome of the interaction.
E)The value of only the active predictors.
Which statement about the expected performance of a binary model is correct?
A)You set the expected performance before deploying the model.
B)The expected performance of a binary model can range from 0 to 100.
C)The expected performance is an optional field for binary models.
D)The system automatically calculates the expected performance.
What is the purpose of the Combiner function in the scorecard?
A)To set the Cutoff value to distinguish potential churners from loyal customers.
B)To select a method to calculate the score.
C)To create complex predictor field expressions.
D)To apply a data transform to run the scorecard for different customers.
Which statement about the PMML standard is correct?
A)The PMML standard facilitates the exchange of models between applications.
B)The PMML standard facilitates the exchange of scores between applications.
C)The PMML standard is a proprietary standard.
D)You can only use the PMML standard to describe tree, scorecard, and regression
models.
When you build a model by using Pega machine learning, Pega machine learning uses
the validation holdout set to ____________ and to ____________.
A)train the models
B)select the best model
C)compare the performance of candidate models
D)check for robustness of candidate models
E)analyze the performance characteristics of candidate models
The purpose of model templates in Pega machine learning is ____________.
A)to set the model outcomes
B)to streamline model development
C)to set the model context
D)to streamline model deployment
What is the purpose of shadow mode when you deploy a new model?
A)To deploy the new model immediately as an active model.
B)To test the new model with a sample of data.
C)To monitor the performance of the new model in a production environment before
you deploy it as an active model.
D)To drive the prediction by both the active model and the new model.
What is the purpose of using a validation data set when updating a predictive model
by using Machine Learning Operations (MLOps)?
A)To deploy the updated model to a staging environment.
B)To test the updated model in a development environment.
C)To compare the performance of the updated model to the active model.
D)To monitor the updated model in a production environment.
What is the purpose of the business operations environment (BOE) in a one-to-one
customer engagement project that uses Pega Customer Decision Hub™?
A)To create, modify, and test business artifacts and conduct simulations.
B)To collect customer responses.
C)To propagate the next best actions to external channels.
D)To add precalculated model scores to the data model.
U+ Telecom uses predictive analytics in its retention strategy. You have created a
predictive model based on recent historical company data and have placed the new
model in shadow mode.
Which statement is true about the new predictive model?
A)The active model does not affect business outcomes.
B)The new model affects business outcomes.
C)The active model and the new model affect business outcomes.
D)The new model does not affect business outcomes.
Who is responsible for adding precalculated model scores to the Customer Decision
Hub data model?
A)The technical team.
B)The Revision Manager.
C)The Data Scientist.
D)The Team Lead.
A strategy designer has created 10 actions in the Sales/Credit Cards group and 10
actions in the Sales/Mortgages group. He would like to import all 10 actions from
the Credit Cards group and only two actions from the Mortgage group into one
decision strategy. What is the minimum number of Proposition Data components he
needs to use in his strategy?
A)twelve
B)three
C)two
D)one
U+ Bank wants to offer a 10% discount for customers whose CLV value is higher than
400. Which strategy component should you use to meet the new requirement?
A)Prioritize
B)Set property
C)Filter
D)Group by
U+, a retail bank, offers the Standard card, the Rewards card and the Rewards Plus
card to its customers. The bank wants to display the banner for the offer that each
customer is most likely to click; therefore, their Arbitration uses Propensity from
the AI models. If you are debugging the Next-Best-Action decision strategy, which
strategy component will show you if the result of the Arbitration is correct?
A)Prioritize
B)Group by
C)Set property
D)Filter
U+ Bank shows mortgage offers on its website by using Pega Customer Decision Hub™.
As a data scientist at U+ Bank, you suspect that if the number of customer visits
to the mortgages web page increases, the increase might mean a higher interest in
the product. Currently, there is no such field in the customer data model. To
verify your hypothesis, you create ____________.
A)a non-parameterized predictor
B)a parameterized predictor
C)a strategy
D)an Interaction History (IH) summary
U+ Bank promotes credit card offers on its website and uses Pega Customer Decision
Hub™ to personalize the offer for every customer. Now, the bank wants to lower the
number of customers that leave the bank by showing a proactive retention offer to
high churn risk customers instead. As an NBA analyst, you are tasked with creating
a new applicability setting to comply with the new business rule. Which business
issue or issues do you modify?
A)Both the Sales issue and the Retention issue
B)Only the Retention issue
C)You do not modify any issue.
D)Only the Sales issue
A data scientist creates a predictive model to predict the likelihood of a customer
leaving the bank in the near future. The predictive model aggregates the classes
into three classification groups: high churn risk, medium churn risk, and low churn
risk. As a decisioning consultant, you need to reference the result of this
classification in a decision strategy. Which property allows you to do so?
A)pyPropensity
B)pyOutcome
C)PropertyHasValue
D)pxSegment
Model transparency is an important requirement for many businesses. In Prediction
Studio, you set model transparency thresholds for ________________.
A)a business issue
B)a model
C)a model type
D)a department
U+ Bank, a retail bank, recently implemented a project in which credit card offers
are presented to qualified customers when they log in to the web self-service
portal. The bank does not want any bias except to satisfy the eligibility condition
Age >=18.
As a decisioning consultant, how do you configure the ethical bias policy to allow
a minimum bias on age?
A)No detection
B)0.7 Gini
C)0 Gini
D)0.1 Gini
What is predictor importance in the context of machine learning?
A)The importance of each predictor in calculating propensity.
B)The importance of each predictor to the business team.
C)The importance of each predictor across all models.
D)The importance of each decision in calculating propensity.
As a decisioning consultant, you are tasked with configuring the ethical bias
policy.
Which context do you need to select to add bias fields?
A)Customer
B)Treatment
C)Action group
D)Action
In Pega Customer Decision Hub™, adaptive models calculate propensities that support
the optimization of customer interactions.
How does the system support explanations of the factors that drive a single
propensity calculation?
A)By showing the current prioritization of next-best-action results.
B)By offering built-in support for calculating global predictor importance.
C)By exporting the ADM predictor importance report for further analysis.
D)By exporting the ADM predictor importance report for further analysis.
E)By providing local explanations of the decisions made by Customer Decision Hub.
U+ Bank introduces a new credit card that has no historical customer behavior data.
U+ Bank wants to offer this credit card on the personalized web portal. Given the
scenario, which rule type must you use?
A)Decision table
B)Adaptive model
C)Pega machine learning model
D)When rule
Which of the following statements is true when describing Predictive Model Markup
Language (PMML) models?
A)PMML models are specialized models created by Predictive Analytic Director.
B)PMML-compliant models cannot be used in the same strategies as Pega’s predictive
models.
C)PMML-compliant models can be used in the same strategies as Pega’s predictive
models.
D)PMML models must be converted before they can be used.
An Adaptive model component in a decision strategy can output ____________.
A)the propensity of a customer to accept an action
B)the number of customers eligible for an Action
C)an Action with a high propensity
D)a sub-strategy that adapts to customer behavior
Adaptive models can learn without historical evidence. What is the starting
propensity of every action?
A)1
B)0.25
C)0.5
D)0
In a decision strategy, which component is necessary to enable access to primary
customer properties?
A)None; customer properties are available
B)Customer Import
C)Data Import
D)Set Property
What type of a predictor can you use in an adaptive model?
A)Logical
B)Customer identifiers
C)Symbolic
D)Page Type
In a decision strategy, what does a dotted line from a Group By component to a
Filter component indicate?
A)The Filter component references a property from Group By component.
B)Evaluate the Filter component first to evaluate the Group By component.
C)Information from the Group By component copies over to the Filter component.
D)There is a one-to-one relationship between the Group By and the Filter
components.
You have created a predictive model using R and have exported it to Predictive
Model Markup Language (PMML). Which component in a decision strategy do you use to
leverage the PMML model?
A)Third-party model
B)Scorecard model
C)Predictive model
D)Decision tree
A large online store wants to adapt to changing customer behavior but does not want
to be distracted by fads. Adaptive models can help accomplish both business
objectives. Which statement about adaptive models is correct?
A)Adaptive models are regression models.
B)Adaptive models require no historical data to get started.
C)Adaptive models require underlying predictive models.
D)Adaptive models perform a continuous model calculation.
Model transparency is becoming an important requirement for many businesses. In
Prediction Studio, you can set model transparency thresholds for ________________ .
A)a model type
B)a department
C)a business issue
D)a model
Pega Customer Decision Hub™ uses the P*C*V*L arbitration formula to select the next
best action for each customer. Which description best describes the purpose of the
formula?
A)To ensure that the customer always receives the best offer, regardless of the
business objective.
B)To ensure that every customer receives the same action.
C)To provide insight into business processes.
D)To balance customer needs with business objectives.
To configure an Adaptive Model, you must identify the responses that indicate
specific customer behavior.
What types of behavior do you need to identify?
A)Positive, neutral, and negative behavior
B)Positive behavior only
C)Any behavior
D)Positive and negative behavior
PegaPega Customer Decision Hub™ uses the P*C*V*L arbitration formula to select the
next best action for each customer. Which factor in the arbitration formula does AI
calculate?
A)Business levers
B)Propensity
C)Context weighing
D)Business value
Predictions combine predictive analytics and best practices in data science. As a
Data Scientist, what is a valid reason to adjust the default response timeout in a
Prediction that an Adaptive Model drives?
A)Increase lift.
B)Optimize the success rate.
C)Suit the use case.
D)Limit the number of responses.
Using Prediction Studio to build Pega machine learning models on historical data,
you can build two types of models: ____________ and ____________. (Choose Two)
A)adaptive models
B)binary models
C)voice to text model
D)continuous models
When developing a predictive model, the outcome value of a continuous model type
can represent _________________.
A)the purchase value of an offer
B)acceptance of an offer
C)customer churn
D)customer loan default
Which statement about Pega Process AI is correct?
A)Pega Process AI is restricted to self-learning models.
B)Pega Process AI lets you design strategies.
C)Pega Process AI is restricted to external models.
D)Pega Process AI lets you use your own predictive models.
The goal of Pega Process AI is to interpret the incoming data and then decide on
the best action to take. In which two ways can you better analyze the incoming
data?
A)Predictive analytics
B)Decision strategies
C)Business rules
D)Text analytics
E)Event processing
U+ Insurance uses a fraud model to route suspicious claims to an expert for closer
inspection. In the case type that processes incoming claims, what type of process
step does AI support?
A)Approve/Reject step
B)Assignment step
C)Change to a stage step
D)Decision step
U+ Insurance wants to predict successful completion of car insurance claims but has
no historical data. Which rule type do you use in this scenario?
A)Adaptive model
B)PMML model
C)When rule
D)Decision table
As a data scientist, you are tasked with creating a new prediction that estimates
the likelihood that a claim is fraudulous. The application developer wants to
proceed and use the prediction in a case type to test the application. To unblock
the application developer, which task do you prioritize?
A)Create the prediction
B)Create a placeholder scorecard to drive the prediction
C)Create the customer data model
D)Create the predictive model that drives the prediction
U+ Insurance uses Pega Process AI™ to optimize case management and applies adaptive
models to predict the outcome of claim cases. In this scenario, the system creates
an adaptive model for each _____________.
A)Case type stage
B)Case
C)Case type step
D)Case type
U+ Insurance wants straight-through processing of claims with a low fraud risk. As
a data scientist, you create a prediction that calculates the probability that a
claim is fraudulent. What type of prediction do you create?
A)Strategic
B)Customer Decision Hub
C)Text analytics
D)Case management
The objective of Pega Process AI™ is to leverage AI and machine learning
capabilities to automate and optimize business processes. How does Pega Process AI
use AI to optimize business processes? (Choose Two)
A)It uses AI to self-optimize processes and apply your own AI in Pega case
management.
B)It uses AI to predict future business trends.
C)It uses AI to replace human decision-making in case management.
D)It uses AI to automate all business processes.
It uses real-time, adaptive case outcome predictions and your own AI models in
custom predictions.
Which statement about the PMML standard is correct?
A)The PMML standard is a proprietary standard.
B)The PMML standard facilitates the exchange of models between applications.
C)The PMML standard describes only tree, scorecard, and regression models.
D)The PMML standard facilitates the exchange of scores between applications
What types of analyses occur at the high level of natural language processing
(NLP)? (Choose Three)
A)Semantic analysis
B)Inferential analysis
C)Linguistic analysis
D)Casual analysis
E)Syntactic analysis
F)Exploratory analysis
Prediction Studio supports keyword-based topic detection, model-based topic
detection and the combination of these. When using machine learning,
______________________________.
A)the Must keywords function as positive features
B)keywords have a higher impact on the model than the training data
C)the keywords are ignored
D)the Must keywords are required to detect the topic
To optimize their customer interactions, U+ Bank routes all emails that are
complaints to a specialized department. To identify emails that voice a complaint,
the text prediction uses ___________.
A)a sentiment model
B)a language model
C)a topic model
D)an entity extraction model
Which of the following statements about chatbots is true?
A)A text prediction that aims to detect the topic and entities in the message
drives the chatbot.
B)The chatbot can detect only entities that follow a strict text pattern.
C)The prediction needs to be created in the Prediction Studio by the data scientist
to enable the chatbot.
D)After the digital messaging channel creation, the chatbot maps the entities to
case properties and saves them in a case that CSRs can view in the Customer Service
portal.
The goal of entity extraction is to interpret the incoming data and detect the
necessary entity. Which two methods can you use to achieve this business outcome?
(Choose Two)
A)Ruta script
B)Topic detection
C)Machine learning
D)Keywords
A telecommunications company wants to apply text analysis to incoming emails to
understand how satisfied its customers are with various products and services. The
setup requires natural language processing (NLP) of the email texts. What is one of
the types of analysis that occurs during natural language processing?
A)Error analysis
B)Subjective analysis
C)Semantic analysis
D)Presumptive analysis
A data scientist and an application developer must configure a chatbot to improve
customer service response time. The application developer created the digital
messaging channel. What is the next step for the data scientist?
A)Modify pySystemEntities model to meet business needs, and then associate it with
the digital messaging channel that the application developer created.
B)Train the text prediction that the system generates.
C)Create and train a new text prediction and inform the application developer.
D)Create a new text prediction and associate it with the digital messaging channel
that the application developer creates.
E)Reference module: Using entity extraction with chatbot channel
An application developer and a data scientist are working together on a chatbot
project. They want the chatbot to detect that the customer wants to change the
account home address. To achieve this goal, the chatbot must:
-Detect the topic of the message
-Run the correct case type
-Extract 2 entities: new address and account number.
What main tasks must they do to achieve this business scenario? (Choose Three)
A)The data scientist trains the text prediction: several topic detection models and
one entity extraction model.
B)The application developer creates the case type and designs the conversation
flow.
C)The data scientist trains the text prediction: One topic detection model and
several entity extraction models.
D)The application developer creates the digital messaging channel.
E)The data scientist uses the entity extraction model from pySystemEntities model
and trains the topic detection model.
F)The data scientist creates the text prediction and builds one topic detection and
two entity extraction models.
U+ Air recently enabled a chatbot on its website to interact with customers. The
airline asked a data scientist to configure the chatbot to recognize all ticket
number patterns that the airline uses. The company wants the chatbot to be
resistant to human error and detect also unusual ticket numbers. Which combination
of entity extraction methods can the data scientist implement to achieve this
business requirement? (Choose Two)
A)Keywords
B)Machine learning
C)Topic detection
D)Ruta script
A data scientist must create an entity extraction model to recognize bank account
numbers for U+ bank. The account numbers follow a strict pattern: 2 letters
followed by 10 digits. What is the best choice for the data scientist to make to
implement this requirement?
A)Create an entity extraction model based on machine learning, and then train it
with a dataset that contains emails from customers.
B)Create a new Ruta script-based entity in the pySystemEntities model.
C)Create an entity extraction model based on Ruta script.
D)Use the out-of-the-box pySystemEntities model, which contains an entity
extraction model specific for this requirement.
A large online store wants to adapt to changing customer behavior but does not want
to be distracted by fads. Adaptive models can help accomplish both business
objectives. Which statement about adaptive models is correct?
A)Adaptive models do not require predictors to start learning.
B)Adaptive models can predict a continuous numeric value.
C)Adaptive models require underlying predictive models.
D)Adaptive models require no historical data to get started.
Customer Decision Hub overview uses the P*C*V*L arbitration formula to select the
next best action for each customer. Which factor in the arbitration formula is
calculated by using AI?
A)Propensity
B)Context weighing
C)Business levers
D)Action value
Adaptive models start learning without historical evidence. What is the starting
propensity of every action?
Responses
A)1
B)0.25
C)0
D)0.5
Adaptive model strategy components can output ____________.
A)the customer's propensity to accept an action
B)an action
C)an optimized strategy
D)the number of customers eligible for an action
Predictions combine predictive analytics and best practices in data science. As a
data scientist, what is a valid reason to adjust the default response timeout in a
prediction?
A)Increase lift
B)Limit the number of responses
C)Suit the use case
D)Optimize the success rate
When developing a predictive model, the outcome value of a continuous model type
can represent _________________.
A)customer loan default
B)customer churn
C)acceptance of an offer
D)the purchase value of an offer
Model transparency is becoming an important requirement for many businesses. In
Prediction Studio, model transparency thresholds can be set for ________________ .
A)a department
B)a business issue
C)a model type
D)a model
In a decision strategy, which component is required to enable access to primary
customer properties?
A)Customer Import
B)None; customer properties are available.
C)Set Property
D)Data Import
Which statement about the PMML standard is correct?
A)The PMML standard describes only tree, scorecard, and regression models.
B)The PMML standard is a proprietary standard.
C)The PMML standard facilitates the exchange of scores between applications.
D)The PMML standard facilitates the exchange of models between applications.
A telecommunications company wants to apply text analysis to incoming emails to
understand how satisfied its customers are with various products and services. That
setup requires natural language processing (NLP) of the email texts. What is one of
the types of analysis that occurs during natural language processing?
A)Error analysis
B)Semantic analysis
C)Presumptive analysis
D)Subjective analysis
Predictions combine predictive analytics and best practices in data science. Which
two best practices are included in Pega Customer Decision Hub™ predictions? (Choose
Two)
A)Selecting the model with the highest performance
B)Setting the response timeout
C)Using a control group
D)Selecting the audience
U+ Bank introduces a new credit card that has no historical customer behavior data.
U+ Bank wants to offer this credit card on the personalized web portal. Given the
scenario, which rule type must you use?
A)When rule
B)Pega machine learning model
C)Adaptive model
D)Decision table
The U+ Insurance wants straight-through processing of claims with a low fraud risk.
As a data scientist, you create a prediction that calculates the probability that a
claim is fraudulent. What type of prediction do you create?
A)Customer Decision Hub
B)Text analytics
C)Strategic
D)Case management
A company wants to capture the sentiment of messages to allow its customer service
representatives to focus on only the negative messages. Sentiment refers to the
general attitude of the author towards a subject and can be _______________.
A)absent
B)negative
C)offensive
D)defensive
U+ Telecom wants to engage in proactive retention to reduce churn. As a data
scientist, you create a prediction that calculates the probability that a client is
likely to cancel a subscription. What type of prediction do you create?
A)Text analytics
B)Customer Decision Hub
C)Case management
To configure an adaptive model, the responses that indicate specific customer
behavior must be identified. What types of behavior need to be identified?
A)Positive and negative behavior
B)Positive behavior only
C)Positive, neutral, and negative behavior
D)Any behavior
Customer Decision Hub overview uses the P*C*V*L arbitration formula to select the
next best action for each customer. Which description best describes the purpose of
the formula?
A)To provide insight into business processes.
B)To balance customer needs with business objectives.
C)To ensure that the customer is always given the best offer, regardless of the
business objective.
D)To ensure that every customer receives the same action.
U+ Insurance wants straight-through processing of claims with a low fraud risk. As
a data scientist, you create a prediction that calculates the probability that a
claim is fraudulent. What type of prediction do you create?
A)Text analytics
B)Customer Decision Hub
C)Case management
You can replace a low-performing predictive model that drives a prediction with an
updated model. When you approve the model, the system automatically generates a
change request in _________.
A)the development environment
B)the business operations environment
C)an external environment
D)the production environment
As a data scientist, you are tasked with creating a new prediction that estimates
the likelihood that a claim is fraudulent. The application developer wants to
proceed and use the prediction in a case type to test the application. To unblock
the application developer, which task do you prioritize?
A)Create the prediction
B)Create a placeholder scorecard to drive the prediction
C)Create the predictive model that drives the prediction
D)Create the customer data model
Creating predictive models Using Prediction Studio to build Pega machine learning
models on historical data, you can build two types of models: ____________ and
____________. (Choose Two)
A)adaptive models
B)voice to text model
C)continuous models
D)binary models
U+ Telecom uses predictive analytics for customer engagement. As a data scientist,
you have created a predictive model based on recent historical company data and
have placed the new model in shadow mode. Which statement is true about the new
predictive model?
A)The new model affects business outcomes.
B)The new model does not affect business outcomes.
C)The active model does not affect business outcomes.
D)The active model and the new model affect business outcomes.
U+ Bank relies on a text prediction to route incoming emails to the correct
department. To detect an account number in the text, the text prediction uses
____________.
A)an entity extraction model
B)a language model
C)a sentiment model
D)a topic model
The objective of Pega Process AI™ is to interpret the incoming data and then decide
on the best action to take. In which two ways can you better analyze the incoming
data? (Choose Two)
A)Predictive analytics
B)Text analytics
C)Event processing
D)Decision strategies
E)Business rules
What type of a predictor can you use in an adaptive model?
A)Logical
B)Integer
C)Symbolic
D)Page Type
