東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁 到查詢結果 [ author_sort:"li, li." ]

切換: 標籤 | MARC模式 | ISBD

Applying neural network based approa...

Li, Li.

FindBook

Google Book

Amazon

博客來

Applying neural network based approaches to host based intrusion detection: Soft signatures.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Applying neural network based approaches to host based intrusion detection: Soft signatures./
作者:	Li, Li.
面頁冊數:	77 p.
附註:	Source: Masters Abstracts International, Volume: 43-06, page: 2281.
Contained By:	Masters Abstracts International43-06.
標題:	Computer Science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=MR00894
ISBN:	9780494008942

Applying neural network based approaches to host based intrusion detection: Soft signatures.
Li, Li.

Applying neural network based approaches to host based intrusion detection: Soft signatures. - 77 p.

Source: Masters Abstracts International, Volume: 43-06, page: 2281.

Thesis (M.C.Sc.)--Dalhousie University (Canada), 2005.

In this work, a novel host based intrusion detection system is developed by employing Self-Organizing Feature Maps (SOM) to represent system behaviors, using Back Propagation (BP) to model intrusive patterns and Mixture of Expert (MoE) algorithms to achieve robustness and flexibility. This model is trained and tested by off-line Basic Security Module audit files from DARPA 1998 Intrusion Detection Evaluation dataset. The results show that by analyzing the system call sequences at the session level, the system using SOMs presents powerful capability to discover intrusive patterns. Multilayer Perceptron (MLP) and Back Propagation algorithms are applied to post-analyze these intrusive patterns and enhance the detection accuracy, where the MoE model, which works on top of the SOM plus MLP architecture, is employed to gate between different experts to automate the process.

ISBN: 9780494008942Subjects--Topical Terms:

626642
Computer Science.

Applying neural network based approaches to host based intrusion detection: Soft signatures.
LDR:01758nmm 2200265 4500 001 1827251
005 20061222091133.5
008 130610s2005 eng d
020 $a 9780494008942
035 $a (UnM)AAIMR00894
035 $a AAIMR00894
040 $a UnM $c UnM
100 1 $a Li, Li. $3 1036040
245 1 0 $a Applying neural network based approaches to host based intrusion detection: Soft signatures.
300 $a 77 p.
500 $a Source: Masters Abstracts International, Volume: 43-06, page: 2281.
500 $a Adviser: Nur Zincir-Heywood.
502 $a Thesis (M.C.Sc.)--Dalhousie University (Canada), 2005.
520 $a In this work, a novel host based intrusion detection system is developed by employing Self-Organizing Feature Maps (SOM) to represent system behaviors, using Back Propagation (BP) to model intrusive patterns and Mixture of Expert (MoE) algorithms to achieve robustness and flexibility. This model is trained and tested by off-line Basic Security Module audit files from DARPA 1998 Intrusion Detection Evaluation dataset. The results show that by analyzing the system call sequences at the session level, the system using SOMs presents powerful capability to discover intrusive patterns. Multilayer Perceptron (MLP) and Back Propagation algorithms are applied to post-analyze these intrusive patterns and enhance the detection accuracy, where the MoE model, which works on top of the SOM plus MLP architecture, is employed to gate between different experts to automate the process.
590 $a School code: 0328.
650 4 $a Computer Science. $3 626642
690 $a 0984
710 2 0 $a Dalhousie University (Canada). $3 1017625
773 0 $t Masters Abstracts International $g 43-06.
790 1 0 $a Zincir-Heywood, Nur, $e advisor
790 $a 0328
791 $a M.C.Sc.
792 $a 2005
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=MR00894