Intents Categorization for Chatbot Development Using Recurrent Neural Network (RNN) Learning

In the chatbot application, the intent is a set of sentences representing specific user intention while interacting with the chatbot. A set of sentences with the same purpose should be categorized precisely to provide appropriate answers. For humans, understanding the meaning of sentence variations with the same intention is straightforward. But for chatbots, it requires a special technique. This study aims to develop a conversation intent categorization model on smart chatbots with a deep learning algorithm. Hence, the variation of sentence expressions can be understood by chatbots more accurately. The result of using RNN is promising as it reaches 81% accuracy.