PRODUCTION SCENARIO
A review reads, 'The appointment booking was easy, but the technician arrived late.' The team needs a separate positive or negative label for the booking and for the technician.
Which capability fits best?
Answering here is anonymous. Nothing is saved unless you sign in.
Show answer and explanation
Answer: Sentiment analysis with opinion mining
Opinion mining is 'aspect-based sentiment analysis', giving 'more granular information about the opinions related to words (such as the attributes of products or services)', so booking and the technician each get a label. Document-level sentiment returns one label for the review. The review is a single sentence, so sentence-level sentiment also returns one mixed label. Custom text classification requires a labelled training set and returns document-level classes; even a multi-label project labels the whole review rather than attaching an opinion to each target.