Understand Classification Performance Metrics
Defective Product Detection in a Factory: A manufacturing company deploys an AI model to
classify products as defective (positive) or non-defective (negative). The results from 1,000
products are: Defective products: 150, non-defective products: 850. The model correctly
classified 120 defective products as defective. The model incorrectly classified 30 non-defective
products as defective. The model correctly classified 820 non-defective products as non-
defective.
Using this data, answer the following research question
• Construct the confusion matrix based on the given data.
• Using the confusion matrix, calculate the accuracy of the model.
• Using the confusion matrix, calculate: Precision (defective detection accuracy)
• Using the confusion matrix, calculate: Recall (defectives correctly classified)
Understand AI application Scenarios
• A doctor uses an ML model to classify patients as having diabetes, heart disease, or no
illness based on test results. What kind of classification is this?
• A bank wants to predict whether a loan applicant will default or not default on their
loan. What kind of classification is this?
• A bank wants to predict whether a loan applicant will default or not default on their
loan. What kind of classification is this?