PEGACPDS88V1 Certified Pega Data Scientist 88V1 Questions and Answers

Positive and negative behavior Reference:

To configure an adaptive model, you must identify positive and negative behavior that indicate specific customer behavior.

Given the scenario where U+ Bank introduces a new credit card that has no historical customer behavior data and wants to offer this credit card on the personalized web portal, you must use an adaptive model.

Lift is a key metric for the performance of the adaptive models. To measure lift, you need a control group with random actions. Lift is the ratio of the performance of the adaptive model to the performance of the control group. A control group is a subset of customers who receive random actions instead of the ones suggested by the model. References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#data-/data-adm-/data-adm-model/main.htm

Balance customer needs with business objectives Reference:

The goal of Next-Best-Action is to balance customer needs with business objectives.

Performance and evidence are two of the results of an adaptive model. Performance is the percentage of positive responses that the model predicts for a given predictor profile. Evidence is the number of customers who exhibited statistically similar behavior. References: https://community.pega.com/sites/default/files/help_v82/procomhelpmain.htm#rule-/rule-decision-/rule-decision-adaptivemodel/main.htm

The likelihood that a proposition will be accepted by the customer is stored in the strategy property called pyBehavior. This property is calculated by an adaptive model or a predictive model and reflects the customer’s propensity to respond to an offer. References: https://academy.pega.com/module/creating-and-understanding-decision-strategies-archived/topic/using-predictions-decision

When an adaptive model has not collected any evidence, its performance is 0.5, which means that it has no predictive power and is equivalent to a random guess. As more evidence is collected, the performance can increase or decrease depending on how well the model predicts customer behavior. References: https://academy.pega.com/module/predicting-customer-behavior-using-real-time-data-archived/topic/adaptive-models-overview

The purpose of predictions is to build adaptive models.

The adaptive model component in a decision strategy computes a propensity value for each 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

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

Predictive analytics is a branch of data science that uses statistical techniques and machine learning algorithms to find repeatable patterns in data and make predictions about future outcomes or behaviors. References: https://academy.pega.com/module/predictive-analytics/topic/predictive-analytics-overview

To identify emails that voice a complaint, the text prediction uses an entity extraction model.

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

As a data scientist, a valid reason to adjust the default response timeout in a prediction is to suit the use case.

According to the Pega Academy1, decision strategies drive the next best action and comprise a unit of reasoning represented by decision components. You use the Proposition Data component to import actions into a strategy canvas. The sequence of the components in the canvas determines which action is selected for a customer.

The Group By component2 is used to group a list of ranked items based on a field and retain only one element in each group. The function available when configuring the Group By component is Count2, which returns the number of elements in each group.

When building a model using Pega machine learning, the validation hold-out set is used to train the models and to check for robustness of candidate models.

Two tasks that a system architect needs to perform to export historical data are export the data set and create a data set.