Pengembangan Metode Seleksi Fitur Berbasis Chi-Square dan Algoritma Exhaustive untuk Meningkatkan Performa Deteksi pada Jaringan Komputer

Nururrahmah, Aulia Teaku (2023) Pengembangan Metode Seleksi Fitur Berbasis Chi-Square dan Algoritma Exhaustive untuk Meningkatkan Performa Deteksi pada Jaringan Komputer. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 05111950010031-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (2MB) | Request a copy

Abstract

Mendeteksi serangan intrusi pada jaringan telah menarik perhatian banyak peneliti. Sejauh ini terdapat dua metode, yakni berbasis signature dan anomaly. Mendeteksi serangan yang berbasis anomaly membutuhkan Machine Learning dalam proses klasifikasinya. Masalah yang banyak diteliti adalah topik tentang bagaimana mereduksi jumlah fitur sebelum dataset diserahkan ke proses klasifikasi. Metode seleksi fitur sendiri dilakukan untuk menentukan fitur relevan dan tidak relevan. Kami mengusulkan metode seleksi fitur berbasis uji chi-square dengan pencarian Exhaustive. Uji Chi-Square digunakan untuk menghitung skor statistik menggunakan uji independence level. Nilai statistik pada masing-masing fitur akan ditentukan relevansinya menggunakan taraf signifikan dan tabel distribusi chi-square. Dari proses uji chi-square maka diperoleh daftar fitur relevan dengan kelas target. Proses selanjutnya adalah mencari kombinasi terbaik antar fitur dengan menggunakan Exhaustive Algorithm. Penelitian ini diuji coba pada empat dataset yakni KDD Cup 99, NSL KDD, Kyoto 2006+, dan UNSW-NB15. Metode klasifikasi yang dimanfaatkan pada penelitian ini adalah Support Vector Machine (SVM), Decision Tree (DT), dan Naïve Bayes (NB). Berdasarkan penelitian yang sudah dilakukan, metode yang diusulkan terbukti memiliki performa yang lebih baik dibanding saat tanpa menggunalan seleksi fitur apapun. Performa terbaik didapatkan pada uji dataset UNSW-NB15 dengan akurasi mencapai 98,71%. Metode yang diusulkan juga melampaui performa dari metode lain
===================================================================================================================================
Detecting intrusion attacks on networks has attracted the attention of many researchers. So far, there are two methods, namely signature-based and anomaly-based. Detecting anomaly-based attacks requires Machine Learning in the classification process. The feature selection method itself is used to determine relevant and irrelevant features. Irrelevant features will be removed from the feature list. We propose a feature selection method based on the chi-square test with an Exhaustive search. The Chi-Square test is used to calculate statistical scores using the independence level test. The statistical value of each feature will be determined by its relevance using the significant level and the chi-square distribution table. A list of features relevant to the target class is obtained from the chi-square test process. The next process is to find the best combination between features using the Exhaustive Algorithm. This research was tested on four data: KDD Cup 99, NSL KDD, Kyoto 2006+, and UNSW-NB15. The classification methods used in this research are Support Vector Machine (SVM), Decision Tree (DT), and Naïve Bayes (NB). Based on the research that has been done, the proposed method is proven to have better performance than without using any feature selection. The best performance was obtained in the UNSW-NB15 dataset test with an accuracy of 98.71%. The proposed method also surpasses the performance of other methods.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Chi-Square Test, Exhaustive Algorithm, Intrusion Detection System (IDS), Machine Learning, Network Infrastructure, Network Security ============= Chi-Square Test, Exhaustive Algorithm, Intrusion Detection System (IDS), Machine Learning, Network Infrastructure, Network Security
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. T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User:	AULIA TEAKU NURURRAHMAH
Date Deposited:	03 Aug 2023 06:22
Last Modified:	03 Aug 2023 06:22
URI:	http://repository.its.ac.id/id/eprint/102506

Actions (login required)

View Item