Article Per Year (5 Year)
p-Index From 2020 - 2025
0
P-Index
This Author published in this journals
All Journal Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Sindy Erika Br Ginting
Fakultas Ilmu Komputer, Universitas Brawijaya
Computer Science & IT Control & Systems Engineering Education Electrical & Electronics Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Voting Based Extreme Learning Machine dalam Klasifikasi Computer Network Intrusion Detection Sindy Erika Br Ginting; Agus Wahyu Widodo; Putra Pandu Adikara
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 6 (2018): Juni 2018
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1258.29 KB)

Abstract

Intrusion Detection System (IDS) is useful software or system to detect intrusion on computer networks. It works by utilizing artificial intelligence to identify anomalies or signatures from the activity on computer networks. To refine more the IDS, it requires the development of intrusion classification algorithms with high accuracy. Voting based Extreme Learning Machine (ELM) is a new scheme algorithm which updates the Extreme Learning Machine (ELM) in improving ELM classification performance and is known more reliable for many data. In this study, the performance of the V-ELM has been evaluated on the Knowledge Discovery and Data Mining (KDD) Cup 99 dataset to support IDS development. This study showed that V-ELM was produced bad performance when using some data from KDD Cup 99. It was using 1000 training data and 250 testing data from KDD Cup 99 datasets. The data was divided into 3 variants are 40 classes, 5 classes, and 2 classes attack. The parameters which tested are the values of hidden neurons (L), independent training (K), and sensitivity of each intrusion class. This study found that the best accuracy result on independent training (K) was 3 and 100 hidden neurons in 2 attack class data with an accuracy of 72%. The lowest accuracy was obtained on hidden neurons was 100 and independent training (K) was 11 in 40 attack classes with an accuracy of 12%. This result showed that good classification capability in 2 classes and bad classification capability in 40 classes.
Co-Authors Agus Wahyu Widodo Putra Pandu Adikara