IBM A1000-119 Intermediate Practice Exam: Medium Difficulty 2025

A retail company wants to implement an AI system that can predict customer churn by analyzing purchase history, browsing behavior, and customer service interactions. Which combination of AI concepts would be most appropriate for this scenario?

A healthcare organization is developing an AI diagnostic tool that will assist radiologists in detecting tumors. During the model evaluation phase, they find that the model has 95% accuracy but only 60% recall for tumor detection. What should be their primary concern and action?

An insurance company wants to use AI to automate claim approval decisions. Their data science team discovers that the training data contains historical biases where claims from certain ZIP codes were systematically denied more often. What ethical AI principle is at risk, and what should be the approach?

A manufacturing company implements a computer vision system to detect defective products on an assembly line. After deployment, the system performs well during day shifts but shows significantly reduced accuracy during night shifts. What is the most likely cause and solution?

A financial services company is building a chatbot to handle customer inquiries about account balances, transaction history, and loan applications. Which combination of AI technologies would be most appropriate for this multi-functional chatbot?

During model training, a data scientist observes that the training accuracy reaches 99% while the validation accuracy plateaus at 75%. What problem does this indicate, and what are appropriate mitigation strategies?

A transportation company wants to optimize delivery routes in real-time based on traffic conditions, delivery priorities, and fuel efficiency. The system needs to learn and improve its decisions over time through trial and error. Which type of machine learning approach is most suitable?

An e-commerce platform implements a recommendation system that suggests products to users. They need to balance between recommending items the user is likely to buy (exploitation) and introducing new product categories to discover user preferences (exploration). Additionally, they must ensure the system doesn't create filter bubbles. What approach addresses these requirements?

A company is preparing to deploy an AI-powered credit scoring system. Regulators require that loan rejections be explainable to applicants. The data science team has developed a deep neural network with superior accuracy but limited interpretability, and a decision tree model with good accuracy and high interpretability. What should be the governance approach?

A sentiment analysis model trained on social media data performs well on product reviews but poorly when applied to customer support tickets. The model shows 85% accuracy on the original test set but only 60% on support tickets. What phenomenon is occurring, and what is the appropriate solution?