Identifikasi Penyakit dengan Gejala Awal Demam Menggunakan K-Nearest Neighbor (K-NN)
DOI:
https://doi.org/10.24002/jbi.v4i1.329Abstract
Abstract. K-Nearest Neighbor (K-NN) is a method that uses a supervised algorithm where the results from the new sample test are classified based on the majority of the category on K-NN. K-Nearest Neighbor method (K-NN) is one of the clinical decision making method known as Clinical Decision Support System (CDSS). This Research employs the data of patients who have fever symptoms, in order to be classified into 10 possible diseases. The Research objects are 82 data and 72 data are used for training while 10 data are used for testing. Value K=3, will be used for the best results in the disease grouping, with the accuracy value result of classification is 97,2%. It is shown that the K-NN method is part of the CDSS because the value of accuracy that can be tolerated for grouping diseases reaches more than 97%.
Keywords: Classification of disease, fever symptoms, K-NN.
Â
Abstrak. K-Nearest Neighbor (K-NN) adalah suatu metode yang menggunakan algoritma supervised dimana hasil dari sampel uji yang baru diklasifikasikan berdasarkan mayoritas dari kategori pada K-NN. Metode K-NN merupakan salah satu dari metode pengambilan keputusan klinik atau Clinical Decision Support System (CDSS). Penelitian ini menggunakan data pasien dengan gejala awal demam untuk mengelompokkan penyakit yang terdiri dari 10 penyakit. Obyek penelitian menggunakan data sebanyak 82 dengan 72 data digunakan untuk training dan 10 data digunakan untuk testing. Hasil terbaik pengelompokan penyakit menggunakan nilai K=3 dengan nilai akurasi hasil pengelompokkan sebesar 97,2%. Hal ini menunjukkan bahwa metode K-NN merupakan bagian dari CDSS karena nilai akurasi yang dapat ditoleransi untuk pengelompokan penyakit harus mempunyai nilai akurasi diatas 97%.Â
Kata kunci: Gejala awal demam, K-NN, penyakit.
Published
Issue
Section
License
Copyright of this journal is assigned to Jurnal Buana Informatika as the journal publisher by the knowledge of author, whilst the moral right of the publication belongs to author. Every printed and electronic publications are open access for educational purposes, research, and library. The editorial board is not responsible for copyright violation to the other than them aims mentioned before. The reproduction of any part of this journal (printed or online) will be allowed only with a written permission from Jurnal Buana Informatika.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.