English  |  正體中文  |  简体中文  |  Items with full text/Total items : 49287/83828 (59%)
Visitors : 7149330      Online Users : 70
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    Please use this identifier to cite or link to this item: http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/76342


    Title: 協同式動態網路釣魚知識擷取與融合之研究
    Other Titles: A Study of Collaborative Dynamic Phishing Knowledge Acquisition and Fusion
    Authors: 林順傑
    Contributors: 淡江大學資訊工程學系
    Keywords: EMCUD;Dynamic knowledge;Collaborative;Incremental machine learning;Phishing;Version Space
    Date: 2011
    Issue Date: 2012-05-07 14:54:48 (UTC+8)
    Abstract: 知識擷取是建立知識庫系統中的一個主要瓶頸。大多數之前提出的知識擷取方法, 均從專家那邊萃取出靜態知識,但這些方法因為缺乏足夠的資訊,固並未討論到主動發 覺包括新形成的知識。因此,如何蒐集到足夠的資訊,並主動學習並協助專家建立新形 成的知識,驗證並擴展舊有的知識庫,變成一個重要的議題。在去年度,我們提出一個 架構在Dynamic EMCUD 上的協同式知識擷取架構,利用觀察知識庫各個低信賴程度的 隱含規則被推論的行為,包括頻率以及趨勢變化並藉此用來學習可能的新演化物件,然 後再引導專家根據這些推論行為的趨勢來萃取便是這些物件的動態知識。在這個研究計 畫中,我們將提出一個新的協同式動態知識學習與融合方法來協助專家,利用多變量資 訊獲得(Multivariate Information Gain; MIG)來衡量一組特徵屬性子集鑑別力的好壞,應 用在版本空間(Version Space)這個漸進式機器學習演算法上,搭配容易解釋的特徵屬性 空間來協助專家察覺到新形成的知識重要特徵,並進行舊有資料庫的更新。第一年中, 我們將設計最核心的MIG 模式,利用資料探勘的技術,來估計真實的MIG 數值,在不 損失精確度的原則下,改善計算理論MIG 值的效率問題。此外,也針對網路釣魚這個 特定領域,運用訪談、機器學習與知識擷取系統等不同技巧,進行網路釣魚知識庫的建 立。在第二年中,著手進行協同式網路釣魚動態知識擷取與融合架構的協同式策略,並 設計漸進式版本空間機器學習與融合方法。處理更一般的候選集以及更明確的候選集之 概念,來解決原先增加新案例時,要批次更新最佳特徵屬性集的缺點。另外,我們也將 設計一個協同式網路釣魚偵防與教育訓練應用雛形來實際驗證在本計畫中所提的機器 學習方法,並可直接應用於現有的網路釣魚防護系統上,加強知識學習的能力。
    Knowledge acquisition is known to be a critical bottleneck of building knowledge based systems. Many knowledge acquisition methodologies have been proposed to systematically elicit rules of static substantive knowledge from domain experts in the past twenty years. However, none of these methods discusses the issue of discovering dynamic substantive knowledge due to the lack of sufficient information. Hence, how to collect sufficient information and to learn the new evolving knowledge to assist experts in evaluating and extending the original knowledge base becomes increasingly important in the knowledge acquisition field. Last year, we have proposed a collaborative knowledge framework based upon Dynamic EMCUD to monitor the frequent inference behaviors and the trend of weak embedded rules with lower certainty degree and learn the candidates of new evolved objects and then guide the experts to extract the dynamic knowledge of these objects according the trend of inference behaviors. In this research project, we will propose a new iteratively collaborative dynamic knowledge learning and fusion method to assist experts to extract the embedded rules of new evolved knowledge. We focus on designing a Multivariate Information Gain (MIG) on the incremental version space learning algorithm with easy-to-explain feature synthesis to help evaluate the discrimination quality of a subset feature set. Hence, the new evolved features can be easily learned and can be used to update the original knowledge base. In the first year, we will design the core MIG method to improve the efficiency of calculating the MIG value by utilizing the data mining skill to estimate the MIG value without loss the degree of precision. Besides, we will construct the phishing related knowledge base via interviewing, machine learning and knowledge acquisition systems in network security domain. In the second year, we will design the collaborative heuristics in the collaborative phishing dynamic knowledge learning and fusion framework. We will propose an incremental version space learning and fusion based upon the more-general candidate set and more-specific candidate set to manage the potential feature subset to resolve the weakness of the batch update the best feature subset when new case is considered. Moreover, we will also design a collaborative phishing detection and prevention with training education application prototype to evaluate the usability of our proposed incremental machine learning method. In summary, our method can also apply on the current phishing protection systems to improve their knowledge learning ability.
    Appears in Collections:[資訊工程學系暨研究所] 研究報告

    Files in This Item:

    There are no files associated with this item.

    All items in 機構典藏 are protected by copyright, with all rights reserved.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - Feedback