User permissions required: ‘View Sources’ AND ‘Review and label’
Reviewing unreviewed verbatims and accepting or rejecting Re:infer’s predicted labels further trains the model and its accuracy.
To review verbatims you should:
- Navigate to Explore
- Use the left-hand filter bar to view only unreviewed verbatims
- Filter by label or search for key terms or phrases using the search bar
- Select whether you want to see recent, random (Shuffle) or verbatims with the model is unsure how to label (Teach) from the dropdown menu in the top right-hand corner of the page
- Accept or reject the labels the model has predicted apply to the verbatims
Make sure to apply all of the relevant labels in your taxonomy to each verbatim. When you review a verbatim, not only do you teach the model which labels apply, but also which labels don’t. If you don’t apply all relevant labels, you send a negative training signal to the model, which will affect its performance.
Unreviewed verbatims with Re:infer applied labels
The opacity of a label indicates the confidence of Re:infer’s prediction of that label, with higher opacity indicating higher confidence.
Accepting or rejecting labels
Hovering your cursor over the label opens a modal showing the confidence with which the model has predicted the label and, if sentiment analysis is enabled, the net sentiment.
Accept/reject a label
- Clicking on the label, or the sentiment indicator (if sentiment analysis is enabled) pins the label to the verbatim, i.e. it confirms the model’s prediction of that label
- If you want to change the sentiment of the predicted label, click the face image that appears when you hover over the verbatim
- Clicking the ‘X’ on the label dismisses it as being incorrect