Fibrianda, Mercury Fluorida (2018) Analisis Perbandingan Akurasi Deteksi Serangan Pada Jaringan Komputer Dengan Metode Naïve Bayes Dan Support Vector Machine (SVM). Sarjana thesis, Universitas Brawijaya.
Abstract
Serangan Denial of Service (DoS) merupakan suatu tindakan untuk melumpuhkan server komputer pada jaringan internet sehingga komputer tidak dapat menjalankan fungsinya dengan benar. Untuk melakukan pendeteksian atau pencegahan berbagai potensi serangan telah dikembangkan Intrusion Detection System (IDS). IDS memiliki dua metode dalam melakukan pendeteksian yaitu Rule Based (Signature Based) dan Behavior Based. Dalam penelitian ini digunakan metode behavior based dimana dalam proses kerjanya membutuhkan sebuah dataset dan metode. Metode yang dapat digunakan salah satunya adalah teknik klasifikasi data mining. Tetapi tidak semua algoritma data mining memiliki kinerja yang baik dalam mengklasifikasi jenis serangan. Oleh karena itu, penelitian ini melakukan perbandingan beberapa metode yaitu Naïve Bayes, SVM Linear, SVM Polynomial, dan SVM Sigmoid. Dataset yang digunakan dalam penelitian ini adalah dataset dari ISCX testbed tanggal 14 Juni 2012. Penelitian ini menganalisis perbandingan metode yang dihasilkan dari proses klasifikasi berupa confusion matrix yang menghasilkan nilai accuracy, precision, recall, dan f1 score. Naive Bayes, SVM Linear, SVM Polynomial dan SVM Sigmoid menghasilkan persentase akurasi berturut-turut sebesar 85,055%, 99,995%, 99,999%, dan 99,995%. Persentase akurasi tertinggi diperoleh SVM Polynomial, sedangkan Naive Bayes menghasilkan persentase akurasi terendah.
English Abstract
Denial of Service (DoS) attacks is an action to cripple the server computers on the internet so that the computer cannot perform its function properly. To perform the detection or prevention of a wide range of potential attacks, the solution which has been developed is Intrusion Detection System (IDS). IDS has two methods in doing detection that is Rule Based (Signature Based) and Behavior Based. In this study we use the methods of behavior based where in the process of its works require a dataset and methods. The methods that can be used, one of which is the classification of data mining techniques. But not all data mining algorithm has good performance in classifying the type of attack. Therefore, this research did a comparison of several methods i.e. Naïve Bayes, SVM Linear, SVM Polynomial, and SVM Sigmoid. The dataset used in this research is the dataset of a testbed ISCX June 14th, 2012. This research analyzes the comparative method which are resulted from a process of classification such as confusion matrix which resulting into the value of accuracy, precision, recall, and the f1 score. Naive Bayes, SVM Linear, SVM Polynomial, and SVM Sigmoid produce consecutive accuracy with the percentage of 99.995%, 85.055%, 99.999%, and 99.995%. The highest percentage of accuracy obtained by SVM Polynomial, while the Naive Bayes generates the lowest accuracy percentage.
Item Type: | Thesis (Sarjana) |
---|---|
Identification Number: | SKR/FTIK/2018/99/051801086 |
Uncontrolled Keywords: | Serangan, DoS, IDS, Klasifikasi, Naïve Bayes, SVM |
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: | 07 Jun 2018 01:49 |
Last Modified: | 27 Oct 2021 05:59 |
URI: | http://repository.ub.ac.id/id/eprint/11476 |
Preview |
Text
BAB I.pdf Download (356kB) | Preview |
Preview |
Text
Bagian Depan.pdf Download (511kB) | Preview |
Preview |
Text
BAB VI.pdf Download (383kB) | Preview |
Preview |
Text
Daftar Pustaka.pdf Download (389kB) | Preview |
Preview |
Text
BAB III.pdf Download (898kB) | Preview |
Preview |
Text
BAB II.pdf Download (1MB) | Preview |
Preview |
Text
BAB IV.pdf Download (904kB) | Preview |
Preview |
Text
BAB V.pdf Download (896kB) | Preview |
Actions (login required)
![]() |
View Item |