STUDI KINERJA METODE SELEKSI FITUR BERBASIS UNCERTAINTY PADA DETEKSI CHRONIC KIDNEY DISEASE DENGAN KLASIFIKASI SVM

Qolby, Lailly Syifa`ul (2021) STUDI KINERJA METODE SELEKSI FITUR BERBASIS UNCERTAINTY PADA DETEKSI CHRONIC KIDNEY DISEASE DENGAN KLASIFIKASI SVM. Masters thesis, Institut Teknologi Sepuluh Nopember.

	Text 05111750010009-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2023. Download (1MB) | Request a copy
	Text 05111750010009-Master_Thesis.pdf Restricted to Repository staff only Download (1MB) | Request a copy
	Text 05111750010009-Master_Thesis.pdf Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Chronic Kidney Disease (CKD) merupakan sebuah kelainan yang merusak
fungsi ginjal. Tanda awal penderita CKD sangat sulit untuk diketahui hingga
penderita kehilangan 25% dari fungsi ginjalnya. Oleh karena itu dibutuhkan
pendeteksian dini dan treatment yang efektif untuk mengurangi tingkat kematian
penderita CKD.
Dalam penelitian ini penulis melakukan diagnosa pada dataset CKD
menggunakan metode klasifikasi Support Vector Machine (SVM) untuk
mendapatkan hasil diagnosa yang akurat. Untuk memperoleh hasil klasifikasi yang
terbaik, maka penulis mengusulkan perbandingan hasil pada penerapan metode
seleksi fitur agar mendapatkan kandidat fitur terbaik dalam meningkatkan hasil
klasifikasi.
Proses pengujian dilakukan dengan membandingkan metode seleksi fitur
Symmetrical Uncertainty (SU) dan Multivariate Symmetrical Uncertainty (MSU)
serta metode SVM sebagai metode klasifikasi. Dengan menggunakan dataset CKD,
dilakukan beberapa skenario percobaan baik dengan menggunakan metode seleksi
fitur SU maupun MSU. Dari hasil ujicoba yang dilakukan menunjukkan bahwa
dengan menggunakan metode seleksi fitur MSU dengan split data 80% : 20%
menghasilkan jumlah fitur penting sebanyak 9 fitur dengan nilai akurasi 0.9,
sensitivy 0.8400, dan specification 1 serta jika dilihat pada grafik ROC grafik
metode MSU menunjukkan nilai true positive lebih tinggi daripada nilai false
positive. Sehingga klasifikasi dengan menggunakan metode seleksi fitur MSU lebih
baik daripada metode seleksi fitur SU.
=====================================================================================================
Chronic Kidney Disease (CKD) is a disorder that impairs kidney function. Early signs of CKD patients are very difficult to know until the patient loses 25% of their kidney function. Therefore, early detection and effective treatment are needed to reduce the mortality rate of CKD sufferers.
In this study, the authors diagnose the CKD dataset using the Support Vector Machine (SVM) classification method to obtain accurate diagnostic results. The authors propose a comparison of the result on the application of the feature selection method in order to get the best feature candidates in improving the classification result.
Testing process is carried out by comparing the Symmetrical Uncertainty (SU) and Multivariate Symmetrical Uncertainty (MSU) feature selection method as well as the SVM method as a classification method. By using the CKD dataset, several experimental scenarios were carried out using both the SU and MSU feature selection methods. From the results of the tests carried out, it shows that using the MSU feature selection method with 80% : 20% data split produces 9 important features with an accuracy value of 0.9, sensitivity 0.8400, specification 1 and when viewed on the ROC graph, the MSU method graph shows the true positive value is higher than the false positive value. So the classification using the MSU feature selection method is better than the SU feature selection method.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Chronic Kidney Disease, Seleksi Fitur, Support Vector Machine, Uncertainty,Chronic kidney disease, feature selection , support vector machine, uncertainty
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User:	Lailly Syifa'ul Qolby
Date Deposited:	20 Aug 2021 06:44
Last Modified:	20 Aug 2021 06:44
URI:	http://repository.its.ac.id/id/eprint/88100

Actions (login required)

View Item