PEGACPDS25V1

PEGACPDS25V1 practice questions, page 2 of 2, with answer review and explanations.

Question 6

A global insurance company is implementing Pega Customer Decision Hub and wants to maximize the predictive accuracy of their models for determining customer propensity to purchase additional coverage. The data science team is evaluating different modeling techniques.What is the key difference between Gradient Boosting models and Bayesian models in Adaptive Decision Manager that would influence the choice between them?

  1. Gradient boosting models typically offer higher predictive accuracy but with lower transparency compared to Bayesian models.
  2. Gradient boosting models can only be used for numeric predictors, while Bayesian models work with both numeric and symbolic predictors.
  3. Gradient boosting models have significantly longer runtime compared to Bayesian models, making them impractical for real-time decisions.
  4. Bayesian models can only be used for web channel predictions, while Gradient boosting models work across all channels.

Question 7

An e-commerce company employs data scientists who need to perform advanced analysis on Adaptive Models using Python. The data resides in Prediction Studio, and the data scientists require access to this data within their Python environment for detailed analysis.How can a data scientist export Adaptive Model snapshot data from Prediction Studio for offline analysis in a Python environment?

  1. Use the GET Describe Model API for each model, followed by the POST Export Model API.
  2. Use the POST Trigger Datamart Export API.
  3. Create a Python client, connect directly to the Adaptive Data Mart database, and then query the model snapshots table.
  4. Export the Adaptive Model configuration, then extract the model data from the exported archive file.

Question 8

When selecting the list of predictors for an adaptive model you should __________.

  1. always use numeric type for integer properties
  2. consider properties from a wide range of sources
  3. select at least one date property
  4. select up to a maximum of 500 predictors

Question 9

A large online store uses Pega Customer Decision Hub™ to smoothly adapt to changing customer behavior. Adaptive models help accomplish this business objective as the models learn from customer responses.Which one of the following statements about adaptive models is correct?

  1. Adaptive models require underlying predictive models.
  2. Adaptive models require a historical dataset to start learning.
  3. Adaptive models are often used to predict a continuous value outcome, like the number of items purchased by a customer.
  4. Adaptive models and predictive models can work in conjunction with one another.

Question 10

What is the output of a Bayesian adaptive model?

  1. The propensity score for a unique treatment or action.
  2. The rank of a unique treatment or action.
  3. The propensity score for all unique treatments or actions.
  4. Decisions.