A Scoring Model allows you to differentiate between
- A . Accept, Reject, Maybe Later
- B . Good, Bad
- C . Good, Better, Best
- D . Good, Bad, Unknown
C
Explanation:
A scoring model allows you to differentiate between Good, Better, and Best outcomes for a given proposition or action. A scoring model assigns a numerical value to each outcome based on its desirability or profitability for the business.
References: https://academy.pega.com/module/predictive-analytics/topic/using-scoring-models
As a data scientist, you have enabled capturing of historical data in an adaptive model.
Which two data elements are captured for every customer interaction? (Choose Two)
- A . The value of only the active predictors
- B . The outcome of the interaction
- C . The model metadata
- D . The propensity generated by the model
- E . The value of all predictors
B,E
Explanation:
When capturing historical data in an adaptive model, the outcome of the interaction and the value of all predictors are captured for every customer interaction.
What is the difference between predictive and adaptive analytics?
- A . Predictive models can predict a continuous value.
- B . Predictive models predict customer behavior.
- C . Adaptive models use the customer data as predict*
- D . Predictive models have evidence.
C
Explanation:
The difference between predictive and adaptive analytics is that adaptive models use the customer data as predictors, while predictive models use the customer data as outcomes. Adaptive models learn from real-time customer interactions and update their predictions accordingly. Predictive models use historical customer data to train and validate their predictions.
References: https://academy.pega.com/module/predicting-customer-behavior-using-real-time-data-archived/topic/adaptive-models-overview
The outcome of a scoring model indicates the likely
- A . write-off value of an arrears case
- B . claim value of a health insurance
- C . period in which a spare part has to be replaced
- D . response to an offer
D
Explanation:
The outcome of a scoring model indicates the likely response to an offer that is presented to a customer. For example, a scoring model can predict if a customer will accept, reject, or defer an offer for a credit card upgrade.
References: https://academy.pega.com/module/predictive-analytics/topic/using-scoring-models
The standardized machine learning process (MLOps) lets you replace a low-performing predictive model that drives a prediction with an updated model. When you approve the model, a change request is automatically generated in__________
- A . the business operations environment
- B . an external environment
- C . the production environment
C
Explanation:
When you approve the updated model in the standardized machine learning process (MLOps), a change request is automatically generated in the production environment.
What is the key difference between a predictive model and a human expert?
- A . Predictive models always outperform human experts.
- B . Humans are better at dealing with structured data and identifying patterns.
- C . Predictive models are more capable of detecting patterns in historical data.
- D . Humans make successful predictions on a large amount of data.
As a data scientist, you are tasked with configuring two predictions that are driven by an adaptive model: one for an inbound channel and one for an outbound channel.
To which setting do you need to pay extra attention?
- A . Response timeout
- B . Adaptive model
- C . Predictor fields
- D . Control group
B
Explanation:
As a data scientist, if you are tasked with configuring two predictions that are driven by an adaptive model, you need to pay extra attention to adaptive model settings.
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 . Case management_____
- B . Customer Decision Hub
- C . Text analytics
B
Explanation:
As a data scientist, you create a prediction that calculates the probability that a client is likely to cancel a subscription. The type of prediction you create is Customer Decision Hub.
Which component(s) do you use to calculate the average margin of four actions?
- A . one Set Property component
- B . one Group By component
- C . four Group By components
- D . four Set Property components
A
Explanation:
You can use one Set Property component to calculate the average margin of four actions by using an expression that sums up the margin values of each action and divides by four. You can then use this property in other components, such as Filter or Prioritize.
References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/setting-properties
An adaptive adaptive model component in a decision: propensity, performance, evidence, and positives.
What is evidence in the context of an adaptive model?
- A . The likelihood of a statistically similar behavior
- B . The number of customers who exhibited statistically similar behavior
- C . The number of statistical bins that arc generated by the system
- D . The number of outcomes that system registered
B
Explanation:
Evidence is the number of customers who exhibited statistically similar behavior. It indicates how much data the model has collected for a given predictor profile. The higher the evidence, the more reliable the model is.
References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-adaptivemodel/main.htm
You are the Decisioning Consultant on an Al-powered one-to-one Customer Engagement implementation project. You are asked to design the Next-Best-Action prioritization expression that balances the customer needs with the business objectives.
What factors do you consider in the prioritization expression?
- A . product eligibility rules
- B . customer contact rules
- C . product compatibility rules
- D . business levers
D
Explanation:
Business levers are factors that you consider in the prioritization expression to balance the customer needs with the business objectives. They can include revenue, cost, risk, retention, satisfaction, or any other custom metric that reflects the value of an action.
References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/using-business-levers
What type of a predictor can you use in an adaptive model?
- A . Symbolic
- B . Integer
- C . Page Type
- D . Logical
B
Explanation:
Integer type predictors can be used in an adaptive model, as they represent quantifiable or measurable attributes.
An online store is interested in increasing its revenues from cross-selling and wants to predict the acceptance rate of the offers presented on their website. A customer’s propensity to accept an offer increases when_________.
- A . Similar offers were rejected by the customer
- B . The offer was rejected by similar customers
- C . Similar offers were accepted by the customer
- D . The offer was accepted by similar customers
C
Explanation:
This is because a customer’s propensity to accept an offer depends on their past behavior and preferences. If a customer has accepted similar offers in the past, they are more likely to accept a new offer that matches their interests https://academy.pega.com/sites/default/files/media/documents/2020-12/Mission20301-2-EN-StudentGuide.pdf
When implementing a Next-Best-Action project, which step is recommended to be taken first?
- A . Define Issue and Group hierarchy
- B . Define propositions
- C . Define business rules
- D . Define prioritization formula
A
Explanation:
When implementing a Next-Best-Action project, the recommended first step is to define Issue and Group hierarchy, which are used to organize and categorize propositions based on business objectives and customer needs. This step helps to align the project with the business vision and goals.
References: https://academy.pega.com/module/one-one-customer-engagement/topic/next-best-action-designer
U+ Insurance uses Pega Process AI™ to assess the complexity of the claims and route a claim to the best-suited user. In the case type that handles claims, the application developer wants to use AI to route claims that are likely to miss their deadline to an expert. As a data scientist, what task do you first perform to allow the application developer to reference the AI output in the case type?
- A . Create a predictive model.
- B . Add a decision step to the case type.
- C . Configure an adaptive model to drive the prediction.
- D . Create a prediction.
A
Explanation:
Before a developer can reference the AI output in a case type, a data scientist would first need to create a predictive model. This model will be used to predict which claims are likely to miss their deadline.
In a strategy defined in the "Retension" issue and the "X-Sell" group, you can import________
- A . Actions from all groups under the "Retention" issue
- B . Actions from "X-Sell" group
- C . Actions from the "Sales" issue
- D . All active actions
A
Explanation:
According to the Pega Academy1, a strategy is a unit of reasoning that defines how to select an action for a customer. A strategy can be defined in different issues and groups, which are categories that help organize actions. An issue represents a business goal (such as retention or sales), and a group represents a subcategory of an issue (such as cross-sell or up-sell).
In a strategy defined in the “Retention” issue and the “X-Sell” group, you can import actions from all groups under the “Retention” issue1. This allows you to use actions that are relevant to your business goal and compare them with other actions in the same issue.
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 . The Retention issue
- B . The Sales issue
- C . The Sales issue and the Retention issue
- D . No modification is required
A
Explanation:
To comply with the new business rule of showing a proactive retention offer to high churn risk customers, you should modify the Retention issue.
Which statement about predictive models is true?
- A . Predictive models need historical data to be created
- B . Predictive models need to be specified in a data attribute
- C . Predictive models are always associated with an action
- D . Predictive models need unstructured bie data
A
Explanation:
Predictive models need historical data to be created. Predictive models are statistical models that use historical data to learn patterns and trends and make predictions for future outcomes. Predictive models can be built with Pega machine learning or imported from third-party tools such as PMML or H2O.
References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-predictivemodel/main.htm
Configuring an adaptive model involves selecting the potential predictors.
How many potential predictors are recommended for an adaptive model?
- A . At least 100 fields to reach an acceptable level of model performance
- B . All fields that have been predictive in the past
- C . All available uncorrected fields
- D . Up to 100 fields to limit the impact on model speed
D
Explanation:
Up to 100 fields to limit the impact on model speed
Reference:
When configuring an adaptive model, it is recommended to select up to 100 potential predictors to limit the impact on model speed.
When you build a decision strategy, what property do you use to access the output of a prediction that is driven by a predictive model markup language (PMML) model?
- A . pxEvidence
- B . pxResult
- C . pxSegment
- D . nxOutcome
B
Explanation:
The pxResult property is used to access the output of a prediction that is driven by a PMML model. It contains the predicted value or class for each record in the input data set.
References: https://academy.pega.com/module/predictive-analytics/topic/using-pmml-models
The purpose of predictions is to______________
- A . build adaptive models
- B . monitor the success rate of individual actions
- C . add best data scientist practices to adaptive models
- D . add predictors to adaptive models
B
Explanation:
The main purpose of predictions is to monitor the success rate of individual actions, by estimating the likelihood of certain outcomes.
When building a predictive model, at what stage do you compare the performance of predictive models?
- A . Model Development stage
- B . Model Analysis stage
- C . Model Export stage
- D . Model Comparison stage
B
Explanation:
Comparing the performance of predictive models typically occurs during the Model Analysis stage.
When building a predictive model, the Data Analysis stage is where you
- A . create data samples
- B . select the input data
- C . group predictors
- D . determine the output field
C
Explanation:
During the Data Analysis stage of building a predictive model, predictors are grouped
A telecom company is interested in improving customer engagement on social media. However, there are hundreds of relevant messages posted every day, and it is not practical for customer service representatives (CSRs) to review and respond to all messages. Instead, CSRs should focus on negative messages.
What do you need to analyze the incoming messages?
- A . Predictive model
- B . Adaptive model
- C . Text categorization model
- D . Text extraction model
C
Explanation:
A text categorization model is a type of text analytics model that can analyze the incoming messages and assign them to predefined categories, such as positive, negative, or neutral sentiment. This way, CSRs can focus on negative messages that require immediate attention or escalation.
References: https://academy.pega.com/module/text-analytics/topic/creating-text-categorization-model
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 . Group By
- B . Filter
- C . Set Property
- D . Prioritize
B
Explanation:
To offer a 10% discount for customers whose CLV value is higher than 400, you should use the Filter strategy component.
As a highly experienced data scientist, which two advanced settings are available to you? (Choose Two)
- A . Outcomes
- B . Predictor types
- C . The parameters used to bin the responses
- D . Predictor selection
- E . The update frequency of the models
C,D
Explanation:
These advanced settings allow a highly experienced data scientist to fine-tune an adaptive model.
Which decision component allows you to monitor the real-time performance of a third- party Churn Model?
- A . Scorecard Model
- B . Adaptive Model
- C . PMML Model
- D . Predictive Model
C
Explanation:
The PMML Model can be used to monitor the real-time performance of a third-party Churn Model
The Predictive Model Markup Language (PMML) allows for predictive models to
- A . Perform better
- B . Be easily shared between applications
- C . Use the same modeling process
- D . Be developed faster
B
Explanation:
The Predictive Model Markup Language (PMML) allows for predictive models to be easily shared between applications. PMML is a standard XML format that describes the input parameters, output score, and mathematical formulas of predictive models. PMML enables interoperability between different tools and platforms that support PMML, such as Pega Customer Decision Hub.
References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#data-/data-predictivemodel-/data-predictivemodel-pmml/main.htm
Which decision component allows you to use a third-party Credit Risk Model 80% of the time and a Pega Credit Risk Model 20%?
- A . Filter
- B . Champion Challenger
- C . Adaptive Model
- D . Switch
B
Explanation:
The Champion Challenger component is used to implement a setup where multiple models (e.g., a third-party Credit Risk Model and a Pega Credit Risk Model) are used with different weights (80% and 20% in this case).
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 . Any behavior
- C . Positive behavior only
- D . Positive, neutral, and negative behavior
A
Explanation:
Positive and negative behavior
Reference:
To configure an adaptive model, you must identify positive and negative behavior that indicate specific customer behavior.
When implementing Next-Best-Action, the Customer Lifetime Value Threshold is typically used to_________.
- A . prioritize high value propositions
- B . prioritize customers
- C . determine if the customer is eligible
- D . calculate the customer’s lifetime value
B
Explanation:
The Customer Lifetime Value Threshold is typically used to prioritize customers based on their expected long-term value to the business. Customers who have a higher lifetime value than the threshold are considered high-value customers and receive more personalized and relevant offers.
References: https://academy.pega.com/module/one-one-customer-engagement/topic/next-best-action-designer
A legal firm wants to use text analytics for easier and faster access to information to helo with compliance related issues. The legal firm needs a taxonomy of legal concepts.
What is a taxonomy?
- A . A list of business rules
- B . The output of an expert survey
- C . A sentiment analysis model
- D . A list of valid categories
D
Explanation:
A taxonomy is a list of valid categories that can be used to classify text documents or entities. A taxonomy can be hierarchical or flat, depending on the level of detail required.
References: https://academy.pega.com/module/text-analytics/topic/creating-taxonomy
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 . An entity extraction model
- B . a topic model
- C . a language model
- D . a sentiment model
B
Explanation:
Topic models are typically used to identify the main themes or topics in a body of text. In this case, a topic model could be used to identify emails that pertain to the topic of "complaints".
As a data scientist, you are asked to create a prediction to optimize the click-through rate of a web banner.
What type of prediction do you need to create in Prediction studio?
- A . Adaptive prediction
- B . Case management
- C . Customer Decision Hub
- D . Text analysis
A
Explanation:
Predictive analytics tools like Prediction Studio are used to build predictive models. To optimize the click-through rate of a web banner, you would likely use an adaptive prediction, which uses machine learning to improve predictions based on past data.
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 . Suit the use case
- B . Optimize the success rate
- C . Increase lift
- D . Limit the number of responses
A
Explanation:
As a data scientist, a valid reason to adjust the default response timeout in a prediction is to suit the use case.
Prediction Studio supports keyword-based topic detection, model-based topic detection and the combination of these.
When using machine learning,
- A . keywords have a higher impact on the model than the training data
- B . the Must keywords are required to detect the topic
- C . the Must keywords function as positive features
- D . the keywords are ignored
C
Explanation:
When using machine learning, the Must keywords function as positive features.
What happens when you increase the performance threshold setting of an adaptive model rule?
- A . The number of active predictors increases
- B . The performance of the model is increased
- C . The correlation threshold decreases
- D . The number of active predictors may decrease
D
Explanation:
When you increase the performance threshold setting of an adaptive model rule, the number of active predictors may decrease. The performance threshold is the minimum performance that a predictor must have to be included in the model. If you increase this value, some predictors may not meet the criteria and be excluded from the model.
References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-adaptivemodel/main.htm
To enable an assessment of its reliability, the Adaptive Model produces three outputs: Propensity, Performance and Evidence. The performance of an Adaptive Model that has not collected any evidence is_________.
- A . 1-0
- B . null
- C . 0.5
- D . 0.0
B
Explanation:
The performance of an Adaptive Model that has not collected any evidence (i.e., hasn’t been trained on any data yet) is typically indicated as null, as it doesn’t have any basis for making accurate predictions yet.
As a data scientist, you are tasked with creating a new prediction that estimates a customers’ likelihood to leave the business in the near future. The NBA analyst wants to move forward and use the prediction in Pega Customer Decision Hub™ to test the application.
To unblock the NBA specialist, which task do you prioritize?
- A . Create the prediction
- B . Create the customer data model
- C . Create a placeholder scorecard to drive the prediction
- D . Create the predictive model that drives the prediction
D
Explanation:
To unblock the NBA specialist, as a data scientist, you should prioritize creating the predictive model that drives the prediction.
Which property is automatically recomputed for each decision component?
- A . Property
- B . Rank
- C . Order
- D . Priority
B
Explanation:
The rank property is automatically recomputed for each decision component. It indicates the order in which the actions are presented to the customer, based on their priority and propensity.
References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/ranking-actions
Pega Decision Management enables organizations to make next-best-action decisions.
To which types of decisions can next-best-action be applied?
- A . Determining how to optimize the product portfolio to increase market share
- B . Determining why response rates for a campaign in one region are below average
- C . Determining the cause of a customer’s problem
- D . Determining which banner to show on a web site to increase click rate
D
Explanation:
Pega Process AI™ lets you bring your own predictive models to Pega and use predictions in case types to optimize the way your application processes work and meet your business goals.
To use the outcome of a predictive fraud model in the case type that processes the incoming claim, you need to use the model outcome in the condition of a decision step2. This way, you can route suspicious claims to a fraud expert for closer inspection based on the model’s prediction.
The management team at U+ Insurance wants to improve the experience of dissatisfied customers. The customers send the feedback through email.
To detect the sentiment of the incoming emails, which type of prediction do you need to configure in Prediction Studio?
- A . Pega Customer Decision Hub™ prediction.
- B . Sentiment detection does not require any predictions.
- C . Case management prediction.
- D . Text analytics prediction.
D
Explanation:
To detect the sentiment of the incoming emails, you need to configure a text analytics prediction1234 in Prediction Studio. A text analytics prediction is a type of prediction that uses natural language processing (NLP) to analyze text data and extract insights, such as topics, entities, and sentiments. You can use a text analytics prediction to detect the sentiment of an email based on its content and assign a score ranging from -1 (negative) to 1 (positive). This can help you improve the customer experience by identifying dissatisfied customers and taking appropriate actions.
In a decision strategy, to remove propositions based on the current month, you use a
- A . Calendar component
- B . Filter component
- C . Data Strategy property
- D . Calendar strategy property
B
Explanation:
In a decision strategy, a filter component would be used to remove propositions based on specific criteria, such as the current month.
The result of a Predictive Model is stored in a property called__________.
- A . pyPrediction
- B . pxResult
- C . pyOutcome
- D . pxSegment
Which value is output by an Adaptive Model?
- A . Score
- B . Performance
- C . Behavior
- D . Lift
A
Explanation:
An Adaptive Model outputs a score, which is a quantified estimate of a certain behavior, such as the likelihood of a customer to accept an offer or the likelihood of a customer to churn
Proactive retention is applicable when a customer is
- A . Initiating contact to churn
- B . A high value customer
- C . In a collections process
- D . Likely to churn
D
Explanation:
Proactive retention is applicable when a customer is likely to churn. Proactive retention is a strategy that aims to prevent customer attrition by identifying customers who are at risk of leaving and offering them incentives or solutions to retain them. Proactive retention requires predicting the customer’s churn risk and selecting the next best action accordingly.
References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#decisioning-/decisioning-strategies-/decisioning-strategies-proactive-retention/main.htm
Adaptive model components can output__________
- A . An option___________
- B . An optimized strategy
- C . The number of customer’s eligible for an action
- D . The customer’s propensity to accept an action
D
Explanation:
Adaptive model components can output the customer’s propensity to accept an action. Propensity is the likelihood of a positive response for a given action and predictor profile. It ranges from 0 to 100.
References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-adaptivemodel/main.htm
U+ Insurance uses Pega Process AI™ to route complex claims to an expert. As a data scientist, you have used the wizard to create a prediction with Case completion as the outcome to help with decision routing. You are tasked with monitoring the adaptive models.
When you open the monitoring tab of the adaptive model rule, you see the following chart:
In this scenario, the system creates an adaptive model for each
- A . case type instance
- B . case type
- C . case type step
- D . case type stage
B
Explanation:
In this scenario, the system creates an adaptive model for each case type, such as claim or complaint. The adaptive model learns from the outcomes of each case type and predicts the probability of case completion for each customer.
References: https://academy.pega.com/module/predicting-customer-behavior-using-real-time-data-archived/topic/adaptive-models-case-management
Which statement about the PMML standard is correct?
- A . The PMML standard is designed to facilitate the exchange of models between applications
- B . The PMML standard can only be used to describe tree, scorecard and regression models.
- C . The PMML standard is a proprietary standard
- D . The PMML standard is designed to facilitate the exchange of scores between applications
A
Explanation:
The PMML standard is designed to facilitate the exchange of models between applications.
The standardized model operations process (MLOps) lets you replace a low-performing predictive model that drives a prediction with a superior one.
When you place the new model in shadow mode in the production environment, the current model___________
- A . uses the outcomes of the new model as predictors
- B . is automatically replaced
- C . drives the prediction
- D . no longer drives the prediction
C
Explanation:
When you place the new model in shadow mode in the production environment, the current model still drives the prediction, but the new model runs in parallel and collects performance data for comparison.
References: https://academy.pega.com/module/predictive-analytics/topic/mlops