Mobile Application berbasis Android dalam Lingkup Polycystic Ovary Syndrome (PCOS) dengan Chatbot Natural Language Processing

Padantya, Athallah Belva (2026) Mobile Application berbasis Android dalam Lingkup Polycystic Ovary Syndrome (PCOS) dengan Chatbot Natural Language Processing. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023221054-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Polycystic Ovary Syndrome (PCOS) merupakan gangguan hormonal dan metabolik pada wanita yang sering terlambat proses diagnosisnya karena tingkat kesadaran yang rendah serta keterbatasan akses layanan kesehatan, khususnya di daerah terpencil. Beberapa alat deteksi PCOS masih belum sepenuhnya memberikan pengalaman pengguna secara optimal, baik karena sistem kuesioner dengan jawaban biner, aplikasi yang kurang terarah, dan fitur yang tersebar di berbagai platform. Penelitian ini mengembangkan mobile application berbasis Android yang mampu mengintegrasikan beberapa fitur relevan dalam 1 aplikasi. Penelitian ini membangun full stack application dengan Flutter, FastAPI, dan MangoDB. Deteksi dini PCOS dikembangkan menggunakan chatbot berbasis Natural Language Processing (NLP) dengan GPT-4o untuk ekstraksi jawaban dan respon dialog, IndoBERT untuk klasifikasi biner, dan decision tree untuk prediksi akhir risiko PCOS. Dalam proses pengembangan sistem IndoBERT, dibangun dataset Bahasa Indonesia sebanyak 1,764 data yang seimbang antar kelasnya. Hasil pengujian yaitu menunjukkan model IndoBERT mendapatkan nilai accuracy 96.23%, precision 95.52%, recall 96.97%, dan F1-Score 96.24%. Sedangkan, model Decision Tree mendapatkan nilai accuracy, precision, recall, dan F1-Score masing-masing sebesar 100% yang dikategorikan sebagai nilai sempurna. Hasil pengujian aplikasi menggunakan evaluasi System Usability Scale (SUS) terhadap 35 subjek wanita memperoleh skor 82.5% yang tergolong kategori Acceptable, grade B, dan Good. Penilaian kedua untuk aplikasi ini berupa evaluasi Black Box yang mendapatkan skor keberhasilan 97.8%. Penilaian terakhir, kecepatan sistem mendapatkan hasil berkisar antara 400 hingga 2,000 ms pada setiap prosesnya, tetapi ada peningkatan pada proses chatbot yang membutuhkan waktu berkisar antara 5,000 hingga 7,000 ms. Berdasarkan hal tersebut, aplikasi yang dikembangkan dalam penelitian ini telah berhasil menyediakan media informasi kesehatan wanita dan deteksi dini PCOS yang lebih natural, terintegrasi, dan memiliki potensi untuk memperluas akses edukasi dan layanan kesehatan wanita.
================================================================================================================================
Polycystic Ovary Syndrome (PCOS) is a hormonal and metabolic disorder in women that is often diagnosed late due to low awareness and limited access to healthcare services, especially in remote areas. Several existing PCOS detection tools have not yet provided an optimal user experience, as they rely on binary questionnaires, feature poorly organized application interfaces, and have their functions separated across multiple platforms. This study develops an Android-based mobile application that integrates several relevant features into a single application. The application was developed as a full-stack system using Flutter, FastAPI, and MongoDB. Early PCOS detection was implemented through a chatbot based on Natural Language Processing (NLP), utilizing GPT-4o for answer extraction and conversational responses, IndoBERT for binary classification, and a decision tree for final PCOS risk prediction. During the development of the IndoBERT model, a balanced Indonesian-language dataset consisting of 1,764 samples was constructed. The evaluation results showed that the IndoBERT model achieved an accuracy of 96.23%, a precision of 95.52%, a recall of 96.97%, and an F1-score of 96.24%. Meanwhile, the Decision Tree model achieved perfect performance, with accuracy, precision, recall, and F1-score values of 100%. The application usability evaluation using the System Usability Scale (SUS), involving 35 participants, obtained a score of 82.5%, which falls into the Acceptable category, Grade B, and Good rating. In addition, Black Box testing achieved a success rate of 97.8%. For the final assessment, the system's response time ranged from 400 to 2,000 ms for each process, but there was an increase in response time for the chatbot process, which took 5,000 to 7,000 ms. Based on these results, the application developed in this study has successfully provided a platform for women's health education and early PCOS detection that is more natural and integrated, with the potential to expand access to women's health education and healthcare services.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Android Mobile Application, Decision Tree, IndoBERT, Natural Language Processing (NLP), Polycystic Ovary Syndrome (PCOS)
Subjects:	R Medicine > R Medicine (General) > R858 Deep Learning R Medicine > RC Internal medicine > RC78 Diagnosis, Radioscopic--Examinations, questions, etc. R Medicine > RG Gynecology and obstetrics T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.6 Management information systems T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Electrical Technology > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Athallah Belva Padantya
Date Deposited:	31 Jul 2026 01:49
Last Modified:	31 Jul 2026 01:49
URI:	http://repository.its.ac.id/id/eprint/138161

Actions (login required)

View Item