NTerminal provides sentiment scores for each triggered keyword by leveraging neural network services from Amazon, Google, and IBM. This enables our customers to monitor different communities’ moods about specific digital assets, people, or companies.
NTerminal NLP events contain fields associated with different sentiment values. These values are often combine by our developers to give single sentiment classifications for platform users. While we are likely to change these fields as we improve our modules, the following examples should help you understand some of the different ways sentiment metrics might be structured.
Google and IBM are both structured the same way. The values range from -1 to 1, with 0 being neutral and -1 being extremely negative.
google_sentiment_document
field (overall sentiment of the document/post) and a google_sentiment_sentence
field (sentiment of the sentience with the identified keyword).ibm_sentiment_document
field for the overall sentiment and has a ibm_sentiment_term
field for the sentiment identified as being directly related to the identified keyword.AWS has a field aws_sentiment_document
for overall sentiment of the entire document/post with 4 sub-fields. These four sub-fields are from 0-1, with 1 being 100% of the document being identified as the corresponding sentiment. In the below example you can see the event is ~88.7% neutral and ~10% negative.
aws_sentiment_document
: { [-]AWS also has an aws_sentiment_phrase
field - a metric which gives the percentage of the context associated with the identified keyword that is negative, neurtal, positive, or mixed. The sub-fields are also structured from 0-1, with 1 being 100% of the sentiment regarding that keyword to be identified as the corresponding sentiment.
aws_sentiment_phrase
: { [-]