The sentiment classifier analyzes text to identify and extract opinions. Analysis is at the sentence level. Positive, neutral, and negative sentiments are each rated from 0 - 1 (total value of all three scores is 1.)
Note: If advertisers want to target positive content, it's recommended that they look for neutral or positive scores, as most content scores in the neutral range.
Syntax | "sentiments": [ { "sentiment": "neu", "score": <0-1> }, { "sentiment": "neg", "score": <0-1> }, { "sentiment": "pos", "score": <0-1> } ] |
Example | "sentiments": [ { "sentiment": "neu", "score": 0.53 }, { "sentiment": "neg", "score": 0.41 }, { "sentiment": "pos", "score": 0.06 } ] |
Data Type | Array of objects |
Required | No |
Page Element Analyzed | Text |
Used For | Targeting/anti-targeting |
More Information | /wiki/spaces/VDOC/pages/920781072 |