Pengembangan Metode Information Retrieval dan Haversine Formula Untuk Rekomendasi Penentuan Klinik di Kabupaten Jember

Hizham, Fadhel Akhmad (2021) Pengembangan Metode Information Retrieval dan Haversine Formula Untuk Rekomendasi Penentuan Klinik di Kabupaten Jember. Masters thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 05111950010007-Master_Thesis.pdf] Text
05111950010007-Master_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2023.

Download (1MB) | Request a copy

Abstract

Klinik merupakan fasilitas tempat orang berobat dan memperoleh advis medis serta tempat mahasiswa kedokteran melakukan pengamatan terhadap kasus penyakit yang diderita para pasien. Saat ini, hadirnya virus Corona (COVID-19) membuat banyak klinik menampung pasien yang terpapar virus tersebut. Virus yang pertama kali teridentifikasi di kota Wuhan, Cina pada Desember 2019 lalu, telah menularkan ke banyak orang di seluruh dunia, termasuk Indonesia, tak terkecuali di Kabupaten Jember. Dari kasus tersebut, rekomendasi penentuan klinik sangat diperlukan karena kondisi yang sangat darurat dan kasus positif yang bertambah setiap harinya, namun tidak terbatas dengan penyakit-penyakit lainnya. Rekomendasi penentuan klinik pada penelitian ini menggunakan metode information retrieval, yaitu metode TF-IDF dan BM25 untuk menentukan rekomendasi klinik di Kabupaten Jember berdasarkan kata pencarian dari penggunanya dan diurutkan berdasarkan kemiripan (similarity) dari yang terbesar hingga yang terkecil. Sementara metode haversine formula digunakan untuk memilih klinik dengan jarak yang ditentukan oleh pengguna sebelumnya. Penentuan rekomendasi klinik yang menggunakan metode gabungan information retrieval (similarity) + haversine dilakukan dengan formulasi rata-rata peringkat antara metode haversine dengan metode gabungan, dan formulasi normalisasi nilai similarity maupun nilai haversine. Hasilnya, ada 7 klinik yang menempati peringkat terbaik untuk metode gabungan dengan formulasi rata-rata peringkat, dan ada 47 klinik yang menempati peringkat terbaik untuk metode gabungan dengan formulasi normalisasi. Hasil evaluasi dari metode information retrieval yaitu dengan nilai precision 0,57, recall 0,7, dan F1-Score 0,6. Nilai evaluasi yang cukup rendah disebabkan karena fasilitas pelayanan kesehatan dari masing-masing klinik sangat umum, sehingga untuk kata kunci penyakit secara khusus dianggap tidak relevan. Perbandingan jarak Haversine dengan garis lurus pada Google Maps Line Measure API menunjukkan selisih sebesar 645 meter, dengan persentase selisih 0,639%. Sementara perbandingan jarak Haversine dengan jarak nyata pada Google Maps API berselisih 3.994 meter, dengan persentase selisih 22,411%.
=========================================================================================================
Clinic is a facility where people seek treatment and obtain medical advice and a place where medical students observe cases of illness suffered by patients. Currently, the presence of the Corona virus (COVID-19) has made many clinics accommodate patients who have been exposed to the virus. The virus, which was first identified in the city of Wuhan, China in December 2019, has infected many people around the world, including Indonesia, including Jember Regency. From these cases, recommendations for clinical determination are needed because the conditions are very emergency and positive cases are increasing every day, but not limited to other diseases Recommendations for clinical determination in this study use the information retrieval method, namely the TF-IDF and BM25 methods to determine clinical recommendations in Jember Regency based on the search terms from users and sorted by similarity from the largest to the smallest. Meanwhile, the haversine formula method is used to select clinics with a distance determined by the previous user. Determination of clinical recommendations using the combined information retrieval (similarity) + haversine method is carried out by formulating the average rank between the haversine method and the combined method, and the formulation of normalization of similarity values and haversine values. As a result, there were 7 clinics that ranked best for the combined method with the average formulation of the ranking, and there were 47 clinics that ranked best for the combined method with the normalized formulation. The results of the evaluation of the information retrieval method are with a precision value of 0.57, recall 0.7, and F1-Score 0.6. The evaluation value is quite low because the health service facilities of each clinic are very general, so that the disease keyword is specifically considered irrelevant. Comparison of Haversine distance with a straight line on Google Maps Line Measure API shows a difference of 645 meters, with a percentage difference of 0.639%. Meanwhile, the comparison between the Haversine distance and the real distance on the Google Maps API is 3,994 meters apart, with a percentage difference of 22.411%.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Haversine Formula, information retrieval, Okapi BM25, sistem rekomendasi, TF-IDF ============================================================ Haversine Formula, information retrieval, Okapi BM25, recommendation system, TF-IDF
Subjects: Q Science > Q Science (General)
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User: Fadhel Akhmad Hizham
Date Deposited: 18 Aug 2021 12:49
Last Modified: 18 Aug 2021 12:49
URI: http://repository.its.ac.id/id/eprint/87172

Actions (login required)

View Item View Item