Language Processing Pipeline

When you train a chatbot on a corpus, BotSharp produce a object which is fullfiled in several different steps - this is also refered to as the processing pipeline. The pipe consists of a tagger, named entity recoginzer and so on. Each pipeline component returns the processed data, which is passed over to the next component. A pipeline is linear and one-dirctional, so you should arrange the right order of components.

Intent Classification Pipeline

Named Entity Recognition Pipeline