Build a tag based on keywords or phrases in the dataset
Building a tag based on keywords or phrases can be helpful for specific brands, products or words that are not labelled as Intents in the platform.
How to add a keyword
A keyword or phrase can be added to a tag by typing it in to any search row. The gray word must then be selected in the list of options for the keyword to be added.
For example, the word “card” can be added to the tag issues_receiving_card by selecting the gray word below it.
Using specific phrases
Building a tag with specific key phrases can also help increase the accuracy of the capture rate. For example, the tag issues_receiving_card has a specific key phrase like “not received my card”. Including this key phrase in the tag can be helpful since it will tell the tag to capture documents that have those specific characters.
Using advanced search filters
Using advanced search filters to build tags with keywords is another way to increase the accuracy of the capture rate. For example, the tag issues_receiving_card has a phrase like “not received” in the first row and a keyword like “card” in the second row. Together, these search filters can capture documents that state, “I have not received my package yet, and you’ve already charged my card.”
This is not what the tag was intended to capture. The keyword “charged” needs to be negated in the third row to prevent the tag from capturing documents like it in the future.
Is the Lang.ai platform case-sensitive?
The Lang.ai platform is sensitive to spacing, punctuation, and apostrophe differences in keywords.