Klasifikasi Penyakit Typhoid Dengan Metode kNN Menggunakan Data Rekam Medis

Andrianto, Boby (2019) Klasifikasi Penyakit Typhoid Dengan Metode kNN Menggunakan Data Rekam Medis. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 07111750067026_Master_Tesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Penyakit demam typhoid adalah penyakit yang disebabkan oleh bakteri bakteri Salmonella. Diagnosis penyakit demam typhoid sulit dilakukan di daerah endemis seperti Indonesia. Penyebab sulitnya diagnosis adalah gejala klinis yang mirip dengan penyakit demam lainnya. Kesalahan diagnosis dapat menyebabkan resistensi antibiotik. Menurut WHO, Resistensi antibiotik sudah berada pada tingkat bahaya yang tinggi. Fakta bahwa diagnosis penyakit demam typhoid di beberapa unit pelayanan kesehatan primer di Indonesia masih menggunakan metode konvensional. Metode konvensional masih bergantung pada pengalaman petugas medis tanpa menggunakan data klinis yang kompleks. Akibatnya kesalahan diagnosis penyakit semakin besar karena kesalahan manusia.
Penelitian ini bertujuan untuk membantu petugas medis melakuakan diagnosis penyakit. Penelitian ini mengusulkan sistem pendukung keputusan menggunakan pendekatan data mining berdasarkan data rekam medis elektronik. Metode klasifikasi supervised k-nearest neighbor (kNN) and naïve bayes (NB) dipilih karena termasuk klasifikasi terbaik dalam dunia medis. Dalam penelitian ini membandingkan klasifikasi kNN dan NB dan melakukan optimasi. Dari hasil penelitian data rekam medis elektronik Rumahsakit Islam Jemursari dengan pembagian data 20% data testing dan 80% data latih. Hasil dari penelitian didapatkan metode kNN dengan optimasi k=1 adalah metode terbaik dengan akurasi 91% diatas NB.
=======================================================================================================================================
Typhoid fever is a disease caused by Salmonella bacteria. Typhoid fever difficult to be diagnosed at the endemic area such as Indonesia. The difficulties are causing by clinical indication which similar to the other fever. Misdiagnose could make antibiotic resistance. According to WHO, antibiotic resistance already on a high danger level. The fact that the diagnosis of typhoid fever in some primary service units in Indonesia still uses conventional methods. Conventional methods still depend on the experience of medical personnel without using complex clinical data. As a result, mistakes in diagnosing the disease are higher because of human error. The objective of this research is to support medical personnel to conduct the diagnose process. This research proposes a decision support system using a data mining approach based on electronic medical record data. The supervised k- nearest neighbor (kNN) and naïve bayes (NB) classification methods were chosen because they are among the best classifications in the medical world. This research is comparing kNN and NB classification and conduct the optimization. Analysis of electronic medical data at Jemursari Islamic hospital uses 20% testing data and 80% trained data, results from kNN using k=1 optimization is the best method with 91% accuracy above NB.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Classification, kNN, Typhoid
Subjects:	Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Divisions:	Faculty of Electrical Technology > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Boby BO Andrianto
Date Deposited:	05 Aug 2026 07:44
Last Modified:	05 Aug 2026 07:44
URI:	http://repository.its.ac.id/id/eprint/65710

Actions (login required)

View Item