Implementing Long Short Term Memory (LSTM) in Chatbots for Multi Usaha Raya

Ilham Dwi Raharjo (Universitas Dian Nuswantoro)
Egia Rosi Subhiyakto (Universitas Dian Nuswantoro)

Advance Sustainable Science, Engineering and Technology (ASSET)
Vol. 6 No. 4 (2024): August-October

The furniture industry is an important sector in Indonesia that supports the economy and provides quality furniture. An in-depth understanding of the furniture business is essential for industry players to improve operational efficiency and customer satisfaction. This research aims to develop a chatbot for Multi Usaha Raya furniture company to improve customer service and operational efficiency. In its development, the Machine Learning Model Development Life Cycle (MDLC) and deep learning approach using the Flask platform are employed. LSTM, a type of recurrent neural network (RNN) architecture capable of handling long-term dependencies, is utilized in this chatbot model. The model training results show an accuracy of 99%, validation accuracy of 96%, loss of 0.1%, and validation loss of 0.2% after 200 epochs, demonstrating the effectiveness of the LSTM algorithm for developing a chatbot in this company.