In SAS Viya Text Analytics, sentiment analysis produces sentence-level sentiment scores. A product review contains 5 sentences: 3 positive, 1 neutral, 1 negative. What does the document-level sentiment score represent?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — the aggregate sentiment computed from all sentence-level scores, typically using a combination of positive/negative sentence counts and sentiment strength, is correct because SAS Viya Text Analytics combines sentence-level positive and negative counts, average polarity, and sentiment strength. You don't just take the first sentence, a length-weighted average, or the maximum score. The document score reflects the relative counts and magnitudes of sentiment across all sentences.
Full Explanation
The correct answer is the aggregate sentiment computed from all sentence-level scores, typically using a combination of positive/negative sentence counts and sentiment strength. SAS Viya Text Analytics aggregates sentence-level sentiment scores to the document level by combining positive and negative sentence counts, average polarity, and sentiment strength — not simply max or first-sentence sentiment. The exact aggregation method considers the relative counts and magnitudes of positive/negative sentiment across all sentences. The first-sentence option is wrong; leading sentence dominance is not how SAS aggregates. The length-weighted average option is wrong; length-weighted averaging is not the SAS sentiment aggregation approach. The maximum-score option is wrong; max sentiment would ignore negative sentences entirely.