Rancang Bangun Chatbot Monitoring Menstruasi Berbasis Web Automation Dalam Deteksi Dini Polycistic Ovarium Syndrome (Pcos)

Novitasari, Tiara Hillda (2025) Rancang Bangun Chatbot Monitoring Menstruasi Berbasis Web Automation Dalam Deteksi Dini Polycistic Ovarium Syndrome (Pcos). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023211010-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Polycystic Ovarium Syndrome (PCOS) merupakan gangguan hormonal pada wanita yang mempengaruhi 8-13% wanita usia subur, dengan 70% penderitanya tidak terdiagnosis seumur hidup. Ketidakteraturan siklus menstruasi sebagai gejala utama PCOS sering diabaikan, terutama pada wanita usia 18-24 tahun yang memiliki tingkat ketidakteraturan siklus dua kali lipat dibandingkan usia 35-39 tahun. Kurangnya kesadaran dan edukasi mengenai kesehatan reproduksi menjadi penyebab utama keterlambatan diagnosis. Penelitian ini mengembangkan sistem chatbot WhatsApp berbasis web automation menggunakan Selenium WebDriver untuk deteksi dini risiko PCOS dan monitoring menstruasi, yang diintegrasikan dengan website edukasi kesehatan reproduksi wanita. Sistem diimplementasikan dalam lingkungan Virtual Machine Linux Ubuntu tanpa memerlukan backend server, API eksternal, atau layanan Cloud. Menggunakan algoritma Decision Tree C4.5 untuk klasifikasi risiko PCOS berdasarkan 17 parameter self-assessment. Chatbot mampu melakukan registrasi pengguna, self-assessment gejala PCOS, klasifikasi risiko, serta prediksi dan reminder menstruasi otomatis. Pengujian algoritma Decision Tree C4.5 pada 119 data sekunder menunjukkan peningkatan akurasi dari penelitian sebelumnya, yaitu dari 94.59% menjadi 100% dengan optimasi rasio training-testing 75:25. Struktur Decision Tree mengidentifikasi parameter “keteraturan siklus menstruasi” sebagai root node dengan threshold ≤2.5, mengonfirmasi bahwa parameter ini merupakan indikator paling signifikan untuk klasifikasi PCOS, sesuai dengan kriteria Rotterdam. Validasi pada 94 data primer menghasilkan akurasi 100%, dengan 58 partisipan terklasifikasi tidak berisiko dan 36 berisiko PCOS. Dari 36 subjek berisiko, 28 telah terdiagnosis PCOS oleh dokter spesialis, memvalidasi keakuratan sistem. Evaluasi user experience menggunakan System Usability Scale (SUS) terhadap 36 subjek menghasilkan skor rata-rata 88.61 (kategori “excellent”, grade A).
=====================================================================================================================================
Polycystic Ovary Syndrome (PCOS) is a hormonal disorder in women that affects 8-13% of women of reproductive age, with 70% of sufferers remaining undiagnosed throughout their lives. Menstrual cycle irregularities, as a primary symptom of PCOS, are often overlooked, especially in women aged 18-24, who have a rate of cycle irregularity that is double that of women aged 35-39. A lack of awareness and education regarding reproductive health is the main cause of delayed diagnosis. This research developed a WhatsApp chatbot system based on web automation using Selenium WebDriver for early detection of PCOS risk and menstrual monitoring, integrated with a website for women's reproductive health education. The system was implemented in a Virtual Machine Linux Ubuntu environment without the need for a backend server, external APIs, or cloud services. It uses the Decision Tree C4.5 algorithm for PCOS risk classification based on 17 self-assessment parameters. The chatbot is capable of user registration, self-assessment of PCOS symptoms, risk classification, as well as automatic prediction and reminders for menstruation. Testing of the Decision Tree C4.5 algorithm on 119 secondary data showed an increase in accuracy from previous research, from 94.59% to 100% with an optimized training-testing ratio of 75:25. The structure of the Decision Tree identified the parameter `menstrual cycle regularity` as the root node with a threshold of ≤2.5, confirming that this parameter is the most significant indicator for PCOS classification, in accordance with Rotterdam criteria. Validation on 94 primary data resulted in 100% accuracy, with 58 participants classified as not at risk and 36 at risk for PCOS. Of the 36 atrisk subjects, 28 had been diagnosed with PCOS by a specialist doctor, validating the system's accuracy. User experience evaluation using the System Usability Scale (SUS) on 36 subjects resulted in an average score of 88.61 (in the `excellent` category, grade A).

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Chatbot, deteksi dini, PCOS, Selenium WebDriver, Web Automation, Chatbot, early detection, PCOS, Selenium WebDriver, Web Automation
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. R Medicine > RG Gynecology and obstetrics T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.888 Web sites--Design. Web site development.
Divisions:	Faculty of Electrical Technology > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Hillda Tiara Novitasari
Date Deposited:	04 Aug 2025 12:17
Last Modified:	04 Aug 2025 12:17
URI:	http://repository.its.ac.id/id/eprint/126077

Actions (login required)

View Item