(implemented in version 12.0 - Dec 2023)
Labelling data is a time-consuming and critical part of the classification process. Creating a relevant, highly accurate training dataset for a categorisation model can be challenging, however AI applied in data labelling has the potential to provide tremendous value both in direct monetary terms and in terms of strategic advantages.
Data labelling function in VOC Mine is a technology that uses unsupervised models to provide automation in the creation of training-data, without the need for manual training or topic identification.
The process of using data labelling in Mine is the following:
1. Create/open an existing a query in Mine with using a survey or external data source containing unstructured text-cases. (for more information on how to use Mine, please click here)
2. Once you are in your Query interface, select "Data Labelling" (last tab on the right).
3. You will be able to visualise suggestions of Topics. From here, you can select the topics you want to include. You can create custom topics on the bottom right corner of the interface by clicking on '+'.
4. Configure the accuracy threshold by moving the slider left (0 - low accuracy) or right (1 - high accuracy).
5. Click 'Validate Labels' in order to view the results in Chart view or Table view
Chart View: In chart view, you get an overview of your topics and their respective occurrence vs probability threshold. (You want to aim for higher occurrence on the higher probability threshold Otherwise, it may mean that your topic is not relevant to the dataset).
Table View: Table view gives you a full view of the text cases, organised by labelled topic and sorted by probability score.
6. Export your labelled data by clicking on Export (top right) and then on 'Export data labelling'.