碩士 / 國立中興大學 / 資訊管理學系所 / 102 / With the advances in information technology, the age of Big Data has coming; and feature selection, which is how to find critical information among big data, has become
crucial topic. Therefore, in this research we propose a method of Modified Artificial Fish Swarm algorithms combines Support Vector Machine (SVM) for feature selection.
We also use this proposed model in Botnet detection and find the critical feature of Botnet virus.
The main idea of AFSA is simulate the behaviors of fish swarm and its swarm intelligence to solve optimization problems, such as schedule management and function
optimization. Although the prior research shows AFSA has great performance in function optimization, but AFSA still has a lots defect. Therefore, this research propose a modified AFSA, which is MAFSA, its combine the mechanism of endocrine and give appropriate search space for each fish.
To verify the effectiveness of this model, out experimental used famous datasets in machine learning, University of California, Irvine (UCI). The experimental result shows MAFSA has better classification rate and also find better the optimal feature subset.
Furthermore, this research also simulates a LAN environment which was infected by Botnet virus. The packet data of network flow in this LAN has been collected; and we use the propose model to find the critical feature of Botnet virus.
| Identifer | oai:union.ndltd.org:TW/102NCHU5396023 |
| Date | January 2014 |
| Creators | Sih-Yanng Cheen, 陳斯揚 |
| Contributors | 林冠成 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 58 |
Page generated in 0.008 seconds