A label is a structured summary of an intent or concept expressed within a verbatim. A verbatim is typically summarised by multiple labels - i.e. a label isn't a mutually exclusive classification of the verbatim.
Labels are initially created by users, and then can either continue to be applied (or ‘pinned’) by users or automatically predicted by Re:infer.
As an example, in a dataset monitoring the customer experience we might create a label called ‘Incorrect Invoice Notification’, which describes when a customer is informing the business that they’ve received what they believe is an incorrect invoice.
Labels can be organised in a hierarchical structure to help you organise and train new concepts more quickly. In both the examples above hierarchical labels have been used, e.g. ‘CS > Agents’ and ‘Margin Call > Full Agreement’.
See the below screenshots for different examples of verbatims with different applied labels.
An email in the operations team of a financial services company
Labels are shaded by the confidence that Re:infer has in the predicted labels. The more opaque the label, the higher Re:infer’s confidence is that the label applies.
For datasets with sentiment analysis enabled, every label has an associated positive or negative sentiment indicated by a green or red colour (see below).
A response to an NPS survey. Two labels have been applied to this Verbatim one with positive sentiment the other with negative sentiment