SELEKSI ATRIBUT PADA ALGORITMA NEURAL NETWORK MENGGUNAKAN PARTICLE SWARM OPTIMIZATION UNTUK DIAGNOSIS PENYAKIT LIVER

Abstract

The liver is a vital human organ that has complex and diverse functions, one of which is to maintain the needs of the organs in the body, especially the brain. One of the diseases that attack the liver is hepatitis or liver. According to WHO (World Health Organization) data, nearly 1.2 million people per year, especially in Southeast Asia and Africa, die from liver disease. The problem that usually occurs is that it is difficult to recognize liver disease early on, even when the disease has spread. From these problems, the researchers diagnosed liver disease using data mining using the Neural Network Algorithm and Particle Swarm Optimization (PSO)-based Neural Network Algorithm which was taken from secondary data from the UCI Machine Learning Repository (University of California Invene). Based on the results of the research, the accuracy value of the Neural Network algorithm is 66.83%, while the accuracy value of the Neural Network Optimization algorithm using PSO is 72.37% so that the difference in the accuracy value is 5.54%. So it can be concluded that the application of particle swarm optimization techniques is able to select attributes on the Neural Network, resulting in a better level of accuracy in the diagnosis of liver disease than using the individual method of the Neural Network algorithm.

Keywords: Liver, Neural Network Algorithm, Particle Swarm Optimization (PSO)-based Neural Network Algorithm 

Hati  adalah  organ vital  manusia  yang memiliki   fungsi   kompleks   dan   beragam,   salah satunya  adalah  dengan  menjaga  kebutuhan  organ dalam  tubuh,  khususnya  otak. Salah satu penyakit yang menyerang hati adalah hepatitis atau liver. Menurut data WHO (World Health Organization) menunjukkan hampir 1,2 juta orang per tahun khususnya di Asia Tenggara dan Afrika mengalami kematian akibat terserang penyakit liver. Permasalahan yang biasanya terjadi adalah sulitnya mengenali penyakit liver sejak dini, bahkan ketika penyakit tersebut sudah menyebar. Dari permasalahan tersebut peneliti melakukan diagnosa penyakit liver dengan data mining menggunakan algoritma Neural Network dan Algoritma Neural Network dioptimasi dengan Particle Swarm Optimization (PSO) yang diambil dari data  sekunder Machine Learning Repository  UCI (Universitas California Invene). Berdasarkan hasil penelitian nilai akurasi algoritma Neural Network senilai 66,83%, sedangkan untuk nilai akurasi Optimasi algoritma Neural Network menggunakan PSO sebesar 72,37% dan tampak selisih nilai akurasi yaitu sebesar 5,54%. Sehingga dapat disimpulkan bahwa penerapan teknik optimasi particle swarm optimization mampu menyeleksi atribut pada Neural Network, sehingga menghasilkan tingkat akurasi diagnosis penyakit liver yang lebih baik dibanding dengan menggunakan metode individual algoritma Neural Network.

Kata kunci: Liver, Algoritma Neural Network, Algoritma Neural Network berbasis Particle Swarm Optimization (PSO)