Route each request to ONE primary department: billing, delivery, or cancellation. Separately, extract product names and dates mentioned in the message.
The department categories are predefined. Explain the NLP tasks by distinguishing the routing category from information extracted within the text.
Choose TWO appropriate mappings between the required outputs and NLP tasks.
Routing intent and extracting product names or dates are distinct outputs. Sentiment and language identification do not directly establish the request category.
Detailed explanation
This assigns the message to the single category requested in this scenario. Its output differs from extracting product names or dates.
Products and dates are different outputs from billing, delivery, or cancellation intent. Even when entities are used as classification inputs, a routing decision is still needed.
Language identification returns the language. Messages in the same language can express different intents.
Sentiment concerns attitudes rather than the requested action. Complaints about billing or delivery can also be negative.
This identifies text spans and their types. Check language and product-name coverage; extracted entities alone do not determine the department.
Check official documentation
Compare the following points with the official documentation.
Review the Comprehend classification, entity, and sentiment documentation. For an invented message such as a request to change a product delivery date, list the intent category, product, date expressions, sentiment, and language in separate fields.
Official references
- Official documentation 1
- Official documentation 2
- Official documentation 3
- Official documentation 4
- Official documentation 5
- Official documentation 6
What to confirm
Distinguish routing categories from extracted spans without inferring intent solely from sentiment or language. An invented conceptual example does not establish measured model accuracy.
Key points
- Intent classification
- Named entities
- Sentiment distinction
- Classification outputs
Notes
- Environment: Current AIF-C01 Task 1.1 and related AWS primary-source review using invented examples; no model training or inference is performed.
- Review the Comprehend classification, entity, and sentiment documentation. For an invented message such as a request to change a product delivery date, list the intent category, product, date expressions, sentiment, and language in separate fields.
- Distinguish routing categories from extracted spans without inferring intent solely from sentiment or language. An invented conceptual example does not establish measured model accuracy.
Foundation review
Distinguish the outputs
Intent classification assigns a category to a message; entity recognition identifies spans inside it. A pipeline may use both, but producing one output does not establish that the other requirement has been met.
Read the single-category assumption
This scenario requests one primary category. Classification is not universally restricted to one category: multi-label approaches also exist. Define the business requirement and label meanings first.
Keep intent separate from sentiment
A delivery complaint and a cancellation request can both express negative sentiment. Treat intent, sentiment, language, and extracted entities as separate outputs and evaluate each against suitable verified examples.