Question: Explain how dividing a review-analysis system into a text tagger and a sentiment scorer can improve prediction quality. Use a different example from everyday product reviews.

Sample answer: A first stage can label words or phrases that carry useful meaning, such as product names, opinion words, and intensifiers. A second stage can then use those labels to focus on the most informative parts of the review. For example, in the sentence 'The camera is remarkably clear but the battery is weak,' the tagger might mark 'remarkably' as an intensifier and 'weak' as a negative opinion word. The scorer can give more weight to 'weak' when deciding overall sentiment and treat neutral words like 'camera' as less important. This division helps the system reduce noise and make a more accurate prediction.

Key points:

• The first component adds labels that identify important words or phrases.
• The second component uses those labels to make the sentiment decision.
• Highlighting opinion-bearing words helps the scorer weight them more heavily.
• Separating the tasks can improve accuracy by reducing the impact of uninformative words.

Rubric: Full credit: response explains the role of each stage and how the added labels help the scorer focus on informative words. Partial credit: response describes a two-stage pipeline but does not explain why the extra labels help. No credit: response does not describe a pipeline or the benefit of separating the work.