Implementasi Large Language Model Dalam Pengembangan Chatbot Informasi Layanan Untuk Pasien Di Klinik Pratama Mitra Medicare

Bumbungan, Maria Teresia Elvara (2025) Implementasi Large Language Model Dalam Pengembangan Chatbot Informasi Layanan Untuk Pasien Di Klinik Pratama Mitra Medicare. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027211042-Undergraduate-Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB)

Abstract

Klinik Pratama Mitra Medicare merupakan fasilitas pelayanan kesehatan yang berupaya meningkatkan efisiensi dan kualitas layanan melalui adopsi teknologi digital. Salah satu inovasi yang dikembangkan dalam penelitian ini adalah chatbot berbasis Large Language Model (LLM), yang terfokus menggunakan model IndoBERT, yang bertujuan untuk memberikan informasi layanan kesehatan secara otomatis, cepat, dan akurat kepada pasien. Chatbot ini dirancang untuk menjawab pertanyaan seputar jadwal dokter, sistem antrean, serta informasi layanan tambahan seperti asuransi dan jenis layanan kesehatan yang tersedia, yang terintegrasi langsung dengan sistem manajemen klinik berbasis database MySQL.
Metode pengembangan mencakup pelatihan model IndoBERT menggunakan pendekatan data pair untuk meningkatkan pemahaman semantik chatbot terhadap pertanyaan pengguna. Evaluasi sistem dilakukan menggunakan metrik BERTScore, yaitu precision, recall, dan F1-Score, dan accuracy yang memberikan pengukuran kesamaan semantik antara jawaban chatbot dan jawaban referensi. Hasil evaluasi menunjukkan bahwa pendekatan data pair menghasilkan F1-Score sebesar 0,8277 dan akurasi 95,24% dalam 7 epoch, lebih tinggi dibanding pendekatan tanpa data pair yang hanya mencapai akurasi 60,79% dan F1-Score 0,8173. Selain itu, performa model IndoBERT juga dibandingkan dengan XLM-RoBERTa, di mana IndoBERT menunjukkan efisiensi pelatihan yang lebih baik meskipun sedikit lebih rendah dari sisi F1-Score (0,8335).
Kepuasan pengguna juga menjadi indikator penting, dengan rata-rata skor 3,65 dari skala 4 atau setara dengan tingkat kepuasan sebesar 91,25%. Pengujian menunjukkan bahwa sistem mampu merespons pertanyaan dengan baik, serta mendukung alur pelayanan secara real-time antara pengguna dan admin. Dengan demikian, chatbot IndoBERT tidak hanya meningkatkan aksesibilitas informasi bagi pasien, tetapi juga mengurangi beban kerja administratif staf klinik. Implementasi ini menunjukkan potensi besar dalam mendukung digitalisasi layanan kesehatan yang responsif, efisien, dan berorientasi pada kebutuhan pasien.
=======================================================================================================================================
Pratama Mitra Medicare is a healthcare facility that strives to enhance service efficiency and quality through the adoption of digital technology. One of the innovations developed in this study is a chatbot based on a Large Language Model (LLM), specifically utilizing the IndoBERT model, aimed at providing automated, fast, and accurate health service information to patients. This chatbot is designed to answer questions related to doctor schedules, the queuing system, as well as additional service information such as insurance and available healthcare services, all of which are directly integrated with the clinic's MySQL-based management system. The development method involves training the IndoBERT model using a data pair approach to improve the chatbot's semantic understanding of user queries. System evaluation is conducted using BERTScore metrics precision, recall, and F1-Score along with accuracy, which measure the semantic similarity between the chatbot’s response and the reference answer. The evaluation results show that the data pair approach achieved an F1-Score of 0.8277 and an accuracy of 95.24% over 7 epochs, significantly higher than the non-pair approach, which only reached 60.79% accuracy and 0.8173 F1-Score. In addition, IndoBERT’s performance was compared with XLM-RoBERTa, where IndoBERT demonstrated better training efficiency, although slightly lower in F1-Score (0.8335). User satisfaction also served as a key indicator, with an average score of 3.65 out of 4, equivalent to a satisfaction level of 91.25%. Testing showed that the system was able to respond well to questions and support real-time service flow between users and administrators. Therefore, the IndoBERT chatbot not only enhances information accessibility for patients but also reduces the administrative workload of clinic staff. This implementation demonstrates significant potential in supporting responsive, efficient, and patient-centered digital healthcare services.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Chatbot, Pratama Mitra Medicare Clinic, IndoBERT, LLM, Clinic, BERTScore, Queue System
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Maria Teresia Elvara Bumbungan
Date Deposited:	09 Jul 2025 07:29
Last Modified:	09 Jul 2025 07:30
URI:	http://repository.its.ac.id/id/eprint/119429

Actions (login required)

View Item