SISTEM FAQ KONSULTASI DOKTER GIGI MENGGUNAKAN ALGORITMA WINNOWING DAN SYNONYM REPLACEMENT

Abstract

ABSTRAK

Kesehatan gigi amat penting bagi manusia, namun tidak semua orang masa kini memiliki waktu untuk pergi ke dokter gigi dalam rangka pencegahan penyakit gigi yang lebih kronis, akibatnya banyak orang mendatangi dokter gigi ketika keadaan giginya sudah rusak parah. Untuk menangani permasalahan ini, sangatlah tidak efisien untuk membangun aplikasi konsultasi konvensional. Mengingat jumlah pasien yang tidak sedikit dan kemungkinan terus bertambah kedepannya. Metode yang digunakan dalam penelitian ini adalah deteksi kemiripan dengan bantuan metode Winnowing dan didukung oleh Synonym Replacement. Dengan penyusunan FAQ dalam penerapannya menjadikan konsultasi akan semakin mudah. Uji coba dilakukan dengan menguji beberapa jenis Synonym Replacement dan diketahui melalui ujicoba Synonym Replacement Versi 4 yang terbaik untuk membantu pendeteksian dengan peningkatan angka deteksi sebesar 40% dibandingkan tanpa Synonym Replacement, serta dengan merubah konstanta algoritma Winnowing dan diketahui pula bahwa konstanta prime = 7, n-gram dan n-window = 3 adalah yang terbaik untuk deteksi kalimat, dengan angka deteksi mencapai 58%.

Kata Kunci: Algoritma Winnowing, Synonym Replacement, dan Frequently Ask Question

ABSTRACT

Tooth’s health was very important for human being, unfortunately nowadays people have no enough time to go to the dentist in order to prevents more cronical teeth problems, some of them went to the dentist when the condition was too late for prevention. At this rate, it is very not efficient to build the consulting application that require an answer manually. Considering the number of patient that keep grows by time. In this paper, we would represent our research in building an application such as the answer from problem stated above using Winnowing Algorithm and Synonym Replacement. This application using FAQ to compile and collect the same question in order to ease the process of consultation on the other hand.In this research number of tests are performed, in the test of several versions of Synonym Replacement found that Version 4 were the best in case similarity detection, compared to non-included Synonym Replacement similarity detection was approximately 40% higher, on the other side constant values of Winnowing tests found that prime = 7, n-gram and n-window = 3 constants were the best combination with similarity detection reaching 58%.

Keywords: Winnowing Algorithm, Synonym Replacement, and Frequently Ask Question