I am trying to build a chatbot using RASA The process to build the chatbot can summarized in two step:
First identify the intent and the entities in the user's question
Second, based on a history of intents and their corresponding actions, augment the history of the data.
Use the augmented dataset to train a model (which can be an ML model) in order to use it to identify the actions from the intents.
My problem is to understand how the data augmentation is done in this case. So far and after reading the code for a whole afternoon, I concluded that one of the first steps in the process is the generation of a graph which represents the dataset.