Optimalisasi Layanan Informasi Pulau Buru Melalui Chatbot Berbasis NLP dan NLU

Sirya Sabina, Bin Tahir and Imam Husni, Al Amin Optimalisasi Layanan Informasi Pulau Buru Melalui Chatbot Berbasis NLP dan NLU. Jurnal Riset Sistem Informasi Dan Teknik Informatika (JURASIK).

PDF Download (4MB)

Abstract

Buru Island is one of the prominent islands in the Maluku Archipelago, and the limited understanding of the local community about Buru Island has prompted the development of a Chatbot as an innovative information source. This Chatbot is designed utilizing Natural Language Processing (NLP) and Natural Language Understanding (NLU) through the RASA Framework. Its primary objective is to provide comprehensive information about Buru Island, encompassing tourist destinations, cultural heritage, traditions, culinary aspects, and various other public information. The testing process involved 96 diverse questions, and the results indicated an intent classification accuracy of 99.61%. Evaluation metrics such as micro average, macro average, and weighted average produced positive outcomes, with precision, recall, and f1-score reaching optimal values, ensuring the Chatbot's optimal accuracy level. The Chatbot technology not only enhances efficiency and speed in delivering information to users but also has the potential to operate without administrative supervision, significantly reducing operational costs. Consequently, this Chatbot serves as an effective solution to improve public understanding of Buru Island. The main advantage of this Chatbot lies in its ability to provide comprehensive information, supporting the potential of tourism and culture on Buru Island. Thus, this information service can enhance public understanding, promote tourist destinations, and contribute positively to the overall development of Buru Island

Item Type:	Article
Subjects:	Q Science > Q Science (General)
Faculty / Institution:	Fakultas Teknologi Informasi
Depositing User:	Fakultas Ekonomi
Date Deposited:	13 Mar 2024 04:41
Last Modified:	13 Mar 2024 04:41
URI:	https://eprints.unisbank.ac.id/id/eprint/9960

Actions (login required)

View Item