Wijayanti, Ratna Ayu (2018) Penerapan Algoritme Support Vector Machine Terhadap Klasifikasi Tingkat Risiko Pasien Gagal Ginjal. Sarjana thesis, Universitas Brawijaya.
Abstract
Gagal ginjal merupakan suatu kondisi bahwa ginjal tidak dapat menjalankan fungsinya secara tepat. Di seluruh dunia kasus mengenai gagal ginjal yang mengalami peningkatan setiap tahun adalah gagal ginjal kronik. Di Indonesia menurut data dari penetri (Persatuan Netrologi Indonesia) diperkirakan mencapai 70 ribu penderita penyakit gagal ginjal kronik. Untuk membantu mengetahui status fungsi ginjal seseorang dibuat suatu sistem yang dapat melakukan proses klasifikasi tingkat risiko pasien gagal ginjal menggunakan algoritme support vector machine (SVM) dan strategi one-againts-all. Alur dari penelitian yaitu menggunakan analisis korelasi untuk melihat hubungan antar fitur, melakukan normalisasi agar nilai data berada pada interval yang sama, perhitungan kernel RBF, melakukan proses training dengan sequential training, kemudian menggunakan one-againts-all untuk proses klasifikasi. Pengujian akhir dari penelitian ini menghasilkan nilai rata-rata akurasi sebesar 83,998% dan akurasi tertinggi sebesar 98,33% dengan menggunakan rasio perbandingan data 80%:20%, nilai parameter λ (lambda) = 1, γ (gamma) = 0,0001, σ kernel RBF = 2, C (Complexity)=0,0001 dan jumlah iterasi SVM=100. Berdasarkan hasil tersebut dapat disimpulkan bahwa algoritme SVM dan strategi one-againts-all dapat digunakan untuk klasifikasi tingkat risiko pasien gagal ginjal.
English Abstract
Kidney failure is a condition that the kidneys can not function properly. Worldwide cases of kidney failure are on the rise every year is chronic renal failure. In Indonesia the disease sufferers of chronic kidney failure are categorized as very high. According to data from the penetri (Union of Netrologi Indonesia) was estimated at 70 thousand kidney failure chronic disease sufferers. To help knowing the status of kidney function someone, we made an intelligent system using support vector machine (SVM) algorithm for classification of risk of kidney failure and using one-againts-all strategy. The flow of research those are using correlation analysis to look at the relationships between features, with normalization for data values are at the same interval, the calculation kernel RBF, do the training process with sequential training, then use one-againts-all for the process ofclassification. This study The final test result of this research obtained the average value of accuracy is 83,998% and the highest accuracy is 98,33% using the ratio of data 80%: 20%, with the parameter value of λ (lambda) = 1, γ (gamma) = 0,0001 , σ for kernel RBF = 2, C (Complexity) = 0,0001 and the number of iterations =100. Based on these results it can be concluded that the SVM algorithm and strategy one-againts-all can be used for classification of risk of kidney failure.
Item Type: | Thesis (Sarjana) |
---|---|
Identification Number: | SKR/FTIK/2018/70/051800930 |
Uncontrolled Keywords: | Klasifikasi, Gagal Ginjal, Support Vector Machine, One-Againts-All, Analisis Korelasi |
Subjects: | 000 Computer science, information and general works > 005 Computer programming, programs, data |
Divisions: | Fakultas Ilmu Komputer > Teknik Informatika |
Depositing User: | Yusuf Dwi N. |
Date Deposited: | 03 Jul 2018 01:55 |
Last Modified: | 27 Oct 2021 05:28 |
URI: | http://repository.ub.ac.id/id/eprint/11845 |
Preview |
Text
BAB VI.pdf Download (842kB) | Preview |
Preview |
Text
BAB VII.pdf Download (571kB) | Preview |
Preview |
Text
LAMPIRAN.pdf Download (781kB) | Preview |
Preview |
Text
BAGIAN DEPAN.pdf Download (858kB) | Preview |
Preview |
Text
DAFTAR PUSTAKA.pdf Download (483kB) | Preview |
Preview |
Text
BAB I.pdf Download (502kB) | Preview |
Preview |
Text
BAB II.pdf Download (977kB) | Preview |
Preview |
Text
BAB III.pdf Download (685kB) | Preview |
Preview |
Text
BAB IV.pdf Download (1MB) | Preview |
Preview |
Text
BAB V.pdf Download (951kB) | Preview |
Actions (login required)
![]() |
View Item |