Sentiment

Article summary

Did you find this summary helpful?

Thank you for your feedback

The emotional state of the conversations occurring in the text is evaluated as "sentiment scores".

Parameters:

name	description	default
Address	The address of the Sentiment service that will be used to analyze the sentiment of the text.	http://core-text-sentiment
IgnoreSslErrors	ignore any certificate errors if Sentiment address contains https	false
Language	The information about the language of the text to be analyzed is transmitted through this parameter.	en-US
DialogBased	Evaluate the text to be sent either as a plain text or by converting it into a dialogue format.	false
PreventEmptySentiment	If true, the sentiment node will refrain from sending any event in cases where detected sentiment is neutral.	false

Inputs

Events:

name	description	known nodes that generate this event
*	all events containing the 'Text' attribute	SR Http

Outputs

Audio:
none

Events:

name	description
sentiment	A sentiment event is raised once for each separate input event with a Text field.

Remarks:
This node can work in one of two modes:

Utterance Based :
Analyses each incoming event independently.

Dialog Based :
Collects input events from multiple channels/speakers and treats them as a dialog. Output contains dialog information.

Project Structure
A minimal project that utilizes Sentiment node can be built as such:

Note that the project above is `Utterance Based`. Which means each speech segment will be evaluated independently of each other.

A Dialog Based Sentiment project can be built as such:

Supported flow types: Stream, Batch

Was this article helpful?