Evaluasi Klasifikasi Menggunakan Metode Logistic Regression untuk Memprediksi Penyakit Gagal Ginjal Kronis

Tualeka, SIti Halimah (2021) Evaluasi Klasifikasi Menggunakan Metode Logistic Regression untuk Memprediksi Penyakit Gagal Ginjal Kronis. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
07311640000048-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 April 2023.

Download (6MB) | Request a copy

Abstract

Ginjal merupakan salah satu organ tubuh manusia yang vital. Ginjal berperan penting dalam metabolism tubuh seperti fungsi untuk menyaring darah, menyaring dan membuang limbah, memantau dan mengendalikan keseimbangan air dalam tubuh, mengatur tekanan darah dan tingkatan garam dalam tubuh, mengatur sel darah merah dan mengatur keseimbangan asam-basa (pH) darah. Penyakit ginjal sering disertai penyakit lain yang mendasarinya seperti diabetes militus, hipertensi, dan lain-lain. Menurut World Health Organization (WHO), penyakit gagal ginjal kronis merupakan penyakit dunia dengan angka kematian mencapai 850.000 jiwa per tahun. Menurut data dari IRR (Indonesian Renal Registry) dari 249 renal unit yang melapor, tercatat 30.554 pasien aktif menjalani dialisis pada tahun 2015, sebagian besar adalah pasien dengan gagal ginjal kronik. Gejala gangguan ginjal stadium awal ringan, sehingga sulit didiagnosis hanya dengan pemeriksaan klinis. Dengan menggunakan teknologi yang berkembang saat ini, proses diagnosis dapat dilakukan dengan mudah. Diagnosis dapat dilakukan dengan menggunakan teknik machine learning seperti klasifikasi. Klasifikasi menggunakan solver yang ada dalam Logistic Regression (LogR). Dengan dilakukan proses imputasi menggunakan KNN dan seleksi fitur menggunanakan RFECV. Hasil klasifikasi menggunakan dataset penyakit ginjal kronis menghasilkan nilai paling tinggi dicapai dengan saat k = 3, dengan semua solver mendapatkan ukuran akurasi (1.00), presisi (1.00), spesifitas (1.00), recall atau sensitifitas (1.00) dan F-score (1.00). Oleh karena itu, kami menyimpulkan bahwa pada saat nilai k pada imputasi KNN adalah 3, CKD dapat dideteksi dengan kelimah solver yaitu newton-cg, lbfgs, liblinear, sag, dan saga dengan menggunakan 9 fitur. ================================================================================================ Kidney is a vital organ of the human body. The kidneys play an important role in body metabolism, such as functions to filter blood, filter and remove waste, monitor and control water balance in the body, regulate blood pressure and salt levels in the body, regulate red blood cells and regulate blood acid-base (pH) balance. Kidney disease is often accompanied by other underlying diseases such as diabetes mellitus, hypertension, and others. According to the World Health Organization (WHO), chronic kidney failure is a world disease with a death rate of 850,000 people per year. According to data from the IRR (Indonesian Renal Registry) of the 249 reporting units, there were 30,554 active patients undergoing dialysis in 2015, most of whom were patients with chronic renal failure. Symptoms of early stage renal impairment are mild, making it difficult to diagnose by clinical examination alone. By using technology that is currently developing, the diagnosis process can be done easily. Diagnosis can be made using machine learning techniques such as classification. Classification using a solver in the Logistic Regression (LogR). With the imputation process using KNN and Feature Selection using RFECV. The results of the classification using the chronic kidney disease dataset produced the highest value achieved when k = 3, with all solvers obtaining measures of accuracy (1.00), precision (1.00), specificity (1.00), recall or sensitivity (1.00) and F-score (1.00). Therefore, we conclude that when the k value of the KNN imputation is 3, CKD can be detected by the five solver (newton-cg, lbfgs, liblinear, sag, and saga) using 9 features.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Kidney, Chronic Kidney Deseasem, Classification, Logistic Regression, solver
Subjects: Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression
R Medicine > RZ Other systems of medicine
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User: Siti Halimah Tualeka Ac
Date Deposited: 08 Mar 2021 05:04
Last Modified: 08 Mar 2021 05:04
URI: https://repository.its.ac.id/id/eprint/83777

Actions (login required)

View Item View Item