Automated Thai-FAQ Chatbot using RNN-LSTM

摘要

In the e-commerce model that has online customer service, such as email or live chat, customers mostly use live chat because it is fast and comfortable. Thus, a company needs to hire and pay for admins. However, this incurs the problem that admins need to spend an extensive amount of time for writing an answer and customers have to wait for the answers. Several chatbots are available, but they require users to set up key phrases manually. In this article, we propose and develop a Frequently Asked Questions (FAQs) Chatbot which automatically responds to customers by using a Recurrent Neural Network (RNN) in the form of Long Short-Term Memory (LSTM) for text classification. The experimental results have shown that chatbot could recognize 86.36% of the questions and answer with 93.2% accuracy.