PEGACPDS88V1 Certified Pega Data Scientist 8.8 Questions and Answers

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.

Two results of an adaptive model are propensity and performance. Propensity is the probability that the customer will accept or respond to an offer. Performance is a measure of how well the adaptive model predicts customer behavior over time. References: https://academy.pega.com/module/predicting-customer-behavior-using-real-time-data-archived/topic/adaptive-models-overview

Pega AI is primarily based on self-learning models Reference:

Pega AI is primarily based on  self-learning models .

When using machine learning, the Must keywords function as  positive features .

When defining outcomes for an adaptive model, you must define one or more positive behavior values, which indicate that the customer accepted or responded to the offer. You can also define negative and neutral behavior values, but they are optional. References: https://academy.pega.com/module/predicting-customer-behavior-using-real-time-data-archived/topic/configuring-adaptive-models

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

A prediction helps in proactive retention by predicting the customer’s churn risk. A prediction is an estimate of the likelihood of a future outcome based on historical data and statistical models. A prediction can help identify customers who are at risk of leaving and target them with appropriate actions to retain them. References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#decisioning-/decisioning-strategies-/decisioning-strategies-proactive-retention/main.htm

 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

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

Using Prediction Studio to build Pega machine learning models on historical data, you can build two types of models:  binary models  and  adaptive models .

The Monte Carlo method is a way to create a data set that includes mock-up customer data. You can use this method to generate random values for customer properties, such as age, gender, income, etc., based on predefined probabilities and distributions. References: https://academy.pega.com/module/demonstrating-adaptive-learning-archived/topic/creating-monte-carlo-data-set

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 solid arrow from a “Set Property” component to a “Filter” component means that a property from the “Set Property” component is referenced by the “Filter” component. For example, you can use a “Set Property” component to calculate a customer’s age and then use a “Filter” component to remove actions that are not suitable for that age group. References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/setting-properties

Sentiment refers to the general attitude of the author towards a subject and can be  negative .

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

Import a third party PMML model 1 : PMML stands for Predictive Model Markup Language, which is an XML-based standard for representing predictive models. You can import a PMML model that was created by another tool or platform into Pega and use it in your strategies.

Create an adaptive self-learning model 1 : An adaptive model is a type of predictive model that learns from customer responses and adapts its predictions over time. You can create an adaptive model in Pega and configure its parameters, such as learning rate, decay rate, and performance goal.

Use Pega machine learning to build a model 1 : Pega machine learning is a feature that allows you to build predictive models using various algorithms, such as decision trees, logistic regression, neural networks, and random forests. You can use Pega machine learning to build a model from your data and evaluate its performance.

A predictive model is a mathematical function that estimates the probability of an outcome based on input data.  When building a predictive model, the use of testing and validation samples  increases the accuracy of models 1 2 3 . Testing and validation samples are subsets of data that are used to evaluate how well a model performs on new data that was not used to train the model. Testing and validation samples help prevent overfitting, which is when a model learns too much from the training data and fails to generalize to new data.

 When using a text prediction based on machine learning with keywords configured, the Not keywords function as negative features, meaning that they reduce the probability of detecting the topic if they appear in the text. The Must keywords and May keywords do not have any impact on the machine learning model. References: https://academy.pega.com/module/text-analytics/topic/configuring-keywords

This is because shadow mode allows you to test a new model in parallel with an existing model without affecting the decision outcomes. You can compare the performance of both models and decide whether to replace or keep the existing model. 

https://academy.pega.com/sites/default/files/media/documents/2020-12/Mission20301-2-EN-StudentGuide.pdf

The use of an imported third-party model in a decision strategy is similar to the use of a model built with Pega machine learning. You can use a predictive model component in a decision strategy to reference an imported third-party model and pass the input parameters and receive the output score. You do not need to convert the third-party model into a Pega machine learning model or Pega markup language. References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-predictivemodel/main.htm

When developing a predictive model, the outcome value of a continuous model type can represent  the purchase value of an offer .

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

Predictions combine predictive analytics and best practices in data science. Two best practices included in Pega Customer Decision Hub™ predictions are  using a control group  and  selecting the model with the highest performance .

Customer name is usually not appropriate to be used as a predictor. A predictor is a property that influences the customer behavior and can be derived from various sources such as customer profile, interaction history, proposition details, etc. Customer name is not likely to have any impact on the customer’s preferences or responses, and it may also violate privacy regulations. References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-adaptivemodel/main.htm

To associate an adaptive model with each action in a decision strategy created from scratch, you must define the adaptive model instances in the Adaptive Decision Manager.

Adaptive model predictors are selected from the customer profile, which contains information about the customer’s attributes and behavior. Predictors can be either scalar or aggregate properties that capture customer context, such as channel, location, time, etc. References: https://academy.pega.com/module/predicting-customer-behavior-using-real-time-data-archived/topic/configuring-adaptive-models

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

Predictor selection and The update frequency of the models Reference:

As a highly experienced data scientist, you have access to advanced settings such as  predictor selection  and  the update frequency of the models .

In Customer Decision Hub™, the factor in the arbitration formula that is calculated by using AI is  Propensity .

To predict if a customer is likely to churn, you use a model of type decision tree. A decision tree is a type of predictive model that uses a set of rules to classify customers into different categories based on their attributes and behavior. A decision tree can predict a binary outcome (such as churn or not churn) or a multi-class outcome (such as low risk, medium risk, or high risk). References: https://academy.pega.com/module/predictive-analytics/topic/using-decision-tree-models

Pega machine learning supports the creation of two distinct types of predictive models: categorical and binary. Categorical models predict the outcome of a variable that can have multiple values, such as product category or customer segment. Binary models predict the outcome of a variable that can have only two values, such as yes or no, accept or reject, etc. References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-predictivemodel/main.htm

 Through analysis of customer lifecycles, Next-Best-Action anticipates retention issues and takes proactive actions to prevent customer churn. It uses predictive analytics to identify customers who are at risk of leaving and offers them incentives or solutions to retain them. References: https://academy.pega.com/module/one-one-customer-engagement/topic/proactive-retention

In a decision strategy, the Adaptive Model decision component belongs to the Decision Analytics category. This category contains components that use advanced analytics techniques, such as adaptive models, predictive models, text analytics models, etc., to make predictions or recommendations. References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/decision-analytics-components

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 prioritize component is used to order actions based on a numerical value, such as priority, propensity, or custom expression. You can choose to sort the actions in descending or ascending order. References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/prioritizing-actions

Timely and contextual Reference:

Next-Best-Action ensures that communication between the business and the customer is timely and contextual.