Sentiment Analysis is a natural language processing (NLP) technique used to enable a computer system to analyze a piece of text and classify it as either POSITIVE, NEGATIVE, MIXED or NEUTRAL. The Bellomy Text Analytics toolkit employs a proprietary ensemble of analytic techniques to arrive at a sentiment classification.
- NGRAM Analysis comes first – Assembled from human analysis of 10M+ survey responses
- Small list of strongly-indicative sentiment words (very small list)
- Statistically modeled word pairings (2 and 3 word co-occurrences)
- Deep learning AI trained on over 2M comments