logofirst
logofirst GitHub

Intent Classification

If you want to identify the intent of an audio recording, such as an interview or customer service call, use this template to classify intent by selecting choices and listening to an audio clip.

Interactive Template Preview

Labeling Configuration

<View>
  <Labels name="labels" toName="audio">
    <Label value="Segment" />
  </Labels>
  <AudioPlus name="audio" value="$audio"/>
  <Choices name="intent" toName="audio" perRegion="true" required="true">
    <Choice value="Question" />
    <Choice value="Request" />
    <Choice value="Satisfied" />
    <Choice value="Interested" />
    <Choice value="Unsatisfied" />
  </Choices>
</View>

About the labeling configuration

All labeling configurations must be wrapped in View tags.

Use the Labels control tag to label specific segments of the audio clip:

<Labels name="labels" toName="audio">
    <Label value="Segment" />
</Labels>

Use the AudioPlus object tag to specify the audio data and display an audio wave that can be segmented:

<AudioPlus name="audio" value="$audio"/>

Use the Choices control tag to classify the intent for each segmented region of the audio clip:

<Choices name="intent" toName="audio" perRegion="true" required="true">
  <Choice value="Question" />
  <Choice value="Request" />
  <Choice value="Satisfied" />
  <Choice value="Interested" />
  <Choice value="Unsatisfied" />
</Choices>

Because of the perRegion="true" argument, each choice applies to a different segment labeled as a segment. The required="true" argument ensures that each labeled audio segment has a choice selected before the annotation can be submited.