hello rasa friends, i have a issue about how to use the machine learning model for split my data
the trouble is, i want to create a model with a external data of the nlu archive, like a supermarket data and predict the best option based on a budget
thanks a regards 
Hi Robert,
Iām not sure I understand your use case except that your training data is stored in another system.
When the data is updated, you can convert this data to the Rasa markdown format and then run a train. You can also write your own custom data importer as described in our docs here to pull your data directly from the location where it is stored.
Something else that might be of interest is our knowledge base examples which is another way to access an external source of information.
Greg
thanks for your reply, that solve mi issue about external data, but i need to solve how to get a specific parameter with a machine learning model for example a product based on a budget i guess i need to create a custom actions, or rasa have a method o a way to create that example
regards