Klasifikasi Penyakit Ginjal (Kronis Atau Non-Kronis) Dengan Metode Random Forest Classifier

Basri, La Ode Muhammad Dzulvicar (2021) Klasifikasi Penyakit Ginjal (Kronis Atau Non-Kronis) Dengan Metode Random Forest Classifier. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
07311640000044-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 April 2023.

Download (4MB) | Request a copy

Abstract

Penyakit ginjal merupakan keadaan dimana fungsi ginjal menurun secara berkala yang dapat mengakibatka ketidakmampuan ginjal dalam menjalankan tugasnya. Dalam kondisi kronis, penyakit ginjal akan menyebabkan penurunan fungsi ginjal dalam jangka waktu tertentu. Untuk itu, adanya suatu sistem pendukung keputusan yang mampu untuk mengklasifikasikan jenis penyakit ginjal (kronis atau nonkronis) yang dialami oleh pasien menjadi sangat penting. Dalam penelitian ini, didefinisikan metode yang mampu mengklasifikasikan jenis penyakit ginjal (kronis atau non-kronis) sehingga penyakit ginjal kronis dapat terdeteksi sedini mungkin. Metode random forest dapat digunakan sebagai metode untuk mengklasifikasikan jenis penyakit ginjal (kronis atau non-kronis) dengan menggunakan dataset pasien penyakit ginjal. Dataset yang digunakan dalam penelitian ini merupakan dataset pasien pengidap penyakit ginjal yang ditemukan pada situs penyedia data untuk penelitian, yaitu UCI Repository. Dataset ini dibuat oleh L.Jerlin Rubini dari Universitas Alagappa menggunakan data yang diperoleh dari Dr. P. Soundarapandian dari Rumah Sakit Apollo, Tamil Nadu, India. Dataset penyakit ginjal kronis terdiri dari 400 sampel data, dimana 250 data ditandai sebagai “CKD” (menderita penyakit ginjal kronis) dan 150 lainnya ditandai sebagai “not-CKD” (tidak menderita penyakit ginjal kronis). Ada bebeberapa pre-processing data yang dilakukan sebelum proses klasifikasi, antara lain yaitu encode data kategorik, perhitungan nilai missing values, normalisasi data numerik, menghapus outlier pada data numerik, dan seleksi fitur. Setelah melewati beberapa pre-processing data, dihasilkan 270 sampel data dan 15 fitur untuk proses klasifikasi akhir. Evaluasi proses klasifikasi ditampilkan dalam nilai accuracy, precision, sensitivity, specivicity, dan f1 \ score. Nilai accuracy, precision, sensitivity, specivicity, dan f1 score yang dihasilkan sistem masing masing sebesar 98.5%, 96.4%, 100%, 97.5%, dan 98.1%. ==================================================================================================== Kidney disease is a condition in which kidney function decreases periodically which can result in the inability of the kidneys to carry out their duties. In chronic conditions, kidney disease will cause decreased kidney function over a period of time. For this reason, the existence of a decision support system that can classify the type of kidney disease (chronic or non-chronic) experienced by a patient is very important. In this study, a method is defined that can classify the type of kidney disease (chronic or non-chronic) so that chronic kidney disease can be detected as early as possible. The Random Forest method can be used as a method for classifying the type of kidney disease (chronic or non-chronic) by using a dataset of kidney disease patients. The dataset used in this study is a dataset of patients with kidney disease found on the data provider site for the study, namely the UCI Repository. This dataset was prepared by L. Jerlin Rubini from the University of Alagappa using data obtained from Dr. P. Soundarapandian from Apollo Hospital, Tamil Nadu, India. The chronic kidney disease dataset consisted of 400 data samples, of which 250 were marked as “CKD” (suffering from chronic kidney disease) and another 150 marked as “not-CKD” (without chronic kidney disease). There are several data pre-processing that is done before the classification process, including encoding categorical data, determining missing values, normalizing numeric data, removing outlier from numeric data, and selecting features. After going through some data pre-processing, 270 data samples and 15 features were generated for the final classification process. The classification process evaluation is displayed in accuracy, precision, sensitivity, specificity, and f1 score. The accuracy, precision, sensitivity, specificity, and f1 score values generated by the system are 98.5%, 96.4%, 100%, 97.5, and 98.1%, respectively.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Penyakit ginjal kronis, random forest classifier, K-Nearest Neighbours imputer, metode normalisasi min-max, Boxplot, Recursive Features Elimination with Cross Validation Chronic kidney disease, random forest classifier, K-Nearest Neighbours imputer, min-max method, Inter-quartile Range, Recursive Features Elimination with Cross Validation
Subjects: R Medicine > RZ Other systems of medicine
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User: La Ode Muhammad Dzulvicar Basri
Date Deposited: 06 Mar 2021 00:42
Last Modified: 06 Mar 2021 00:42
URI: https://repository.its.ac.id/id/eprint/83516

Actions (login required)

View Item View Item