Free CPMAI_v7: Cognitive Project Management in AI CPMAI v7 - Training & Certification Exam Questions and Answers

A project manager is briefing a new team member on supervised and unsupervised approaches. Which statement BEST distinguishes classification from clustering?

• AA. Classification assigns records to predefined categories learned from labeled examples, while clustering groups records by similarity without labels.
• BB. Classification groups records by similarity without labels, while clustering assigns records to predefined categories using labeled examples.
• CC. Classification works only on unlabeled data sets, while clustering requires a labeled training set to form its groups.
• DD. Classification predicts a continuous numeric value from past records, while clustering assigns records to predefined categories.

A data scientist tells the team that she will use k-means clustering to segment the company's customers. In k-means, what does the value k represent?

• AA. The number of clusters the algorithm will form
• BB. The number of nearest neighbors used to vote on a customer's group
• CC. The number of features used to describe each customer record
• DD. The number of passes the algorithm makes over the data before it stops

Which of the following lists the six phases of the CPMAI methodology in the correct order?

• AA. Business Understanding, Data Understanding, Data Preparation, Model Development, Model Evaluation, Model Operationalization
• BB. Business Understanding, Data Preparation, Data Understanding, Model Development, Model Evaluation, Model Operationalization
• CC. Data Understanding, Business Understanding, Data Preparation, Model Development, Model Evaluation, Model Operationalization
• DD. Business Understanding, Data Understanding, Data Preparation, Model Evaluation, Model Development, Model Operationalization

A fraud detection model reviews 10,000 transactions and flags 200 of them as fraud. Of the flagged transactions, 150 are confirmed as fraud. The data contains 300 fraudulent transactions in total. What are the precision and recall of the model?

• AA. Precision 0.50, recall 0.75
• BB. Precision 0.60, recall 0.75
• CC. Precision 0.75, recall 0.50
• DD. Precision 0.75, recall 0.60

A bank has deployed a fraud detection model behind a public API. Security monitoring shows that attackers are submitting transactions whose values have been altered in small, deliberate ways so that the model classifies them as legitimate. The training data and the model files have not been accessed. Which type of attack is the bank experiencing?

• AA. Adversarial input (evasion) attack
• BB. Membership inference attack on records
• CC. Data poisoning of the training set
• DD. Model inversion against the prediction API

An insurer wants to identify claims that differ from typical claim behavior without writing a predefined rule for each type of suspicious claim. Which pattern of AI does this problem represent?

• AA. Patterns and anomalies
• BB. Recognition
• CC. Predictive analytics and decision support
• DD. Conversational and human interaction

Which pattern of AI is aimed at creating systems that complete tasks with minimal or no human involvement?

• AA. Goal-driven systems
• BB. Hyperpersonalization
• CC. Recognition
• DD. Autonomous systems

A streaming service with 12 million subscribers wants each subscriber's home screen to show titles, artwork and ordering tailored to that individual's viewing history. Which pattern of AI is the service applying?

• AA. Predictive analytics and decision support
• BB. Autonomous systems
• CC. Hyperpersonalization
• DD. Recognition

During model evaluation, which metric measures the proportion of actual positive cases that the model correctly identified?

• AA. Precision
• BB. Accuracy
• CC. Specificity
• DD. Recall

Which statement correctly distinguishes clustering from classification in machine learning?

• AA. Clustering finds natural groupings in unlabeled data, while classification learns to assign records to known categories.
• BB. Clustering predicts a continuous value for each record, while classification assigns each record to a category.
• CC. Clustering needs labeled examples of every group, while classification discovers groups in data that has no labels.
• DD. Clustering and classification both depend on labeled training data, but clustering returns only a probability score.