Improving NLU model

It's a common situation that you can not foresee all possible phrases, intents, and entities in your conversation script. Thus, it would be best if you improve your NLU model and script all the time. You should use Profiler in DashaStudio for better debugging and maintenance.

You need to validate your model on errors and fix them regularly because it can make mistakes. Do it after a batch of new conversations. To improve the model, extend training examples with new correct data. You should fill data.json file with additional data for intents and entities.

There are actions you can do to improve NLU model:

1. Add the correctly classified phrases

If the model classified unseen phrases correctly, you should add them to includes training examples of corresponding intent. It will extend the domain of this intent and increase the classification precision of these phrases and other as yet unseen phrases.

2. Fix false-positives

False-positive classification error means that intent was recognized but really shouldn't have. These phrases don't have this meaning. To fix this error, you should add them to excludes training examples of corresponding intent. Then the algorithm will know that and make these phrases less similar to phrases from the includes list.

3. Fix false-negatives

False-negative classification error means that intent was NOT recognized but really should have. Usually, the reason is that there are not enough similar phrases for this intent. To fix this error, you should add them to includes training examples of corresponding intent. It will extend the domain of this intent and increase the classification precision of these phrases and other as yet unseen phrases.

4. Create new intents

Take into account that the interlocutor can have some other intention that you didn't use previously. When you validate conversations, you can notice that there are many similar phrases of new meaning. Then you need to create new intent in the includes section and create a transition with it in your script.

Found a mistake? Email us, and we'll send you a free t-shirt!

Enroll in beta

Request invite to our private Beta program for developers to join the waitlist. No spam, we promise.