English  |  正體中文  |  简体中文  |  Items with full text/Total items : 78728/78728 (100%)
Visitors : 34156060      Online Users : 2456
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU 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://ir.lib.ncu.edu.tw/handle/987654321/13354

    Title: 以統計分析探討文件分類程序對期刊論文分類效果之影響;The Study of the Effects of Text Categorization Processes on Journal Papers Classification by Statistical Analysis
    Authors: 賴昆佑;Kun-You Lai
    Contributors: 資訊管理研究所
    Keywords: 統計檢定;分類器;期刊論文分類;文件分類;text categorization;journal papers classification;classifiers;hypothesis test
    Date: 2007-06-28
    Issue Date: 2009-09-22 15:29:47 (UTC+8)
    Publisher: 國立中央大學圖書館
    Abstract: 期刊論文提供專業領域知識,然資訊超載造成檢索時間成本浪費,應用文件分類技術可讓使用者迅速取得相關領域之期刊論文。文件分類程序包含「前處理」、「文件特徵建構」、「分類方法應用」與「分類結果評估」等四個階段。針對期刊論文之分類效果,本研究以統計假設檢定探討期刊論文分類程序中,特徵權重方法、文章欄位差異與應用不同分類器對分類效果之影響,並與本研究設計之抽樣分配分類器進行比較。由實驗模擬與統計假設檢定分析顯示,第一,以特徵比例作為特徵權重方法分類效果顯著優於特徵頻率。第二,文章欄位以「摘要」之分類效果最佳,優於標題與關鍵字,後兩者則無顯著差異。第三,期刊論文分類以支持向量機分類效果最佳,其次為貝式機率分類器、決策樹以及抽樣分配分類器。第四,應用文件分類技術將期刊論文分類之方法可行。另外針對抽樣分配分類器部分,亦提出分析結果與建議,以提升未來研究所需。 Journal papers provide professional domain knowledge. Nevertheless, emerging of information overloading causes considerable cost of time. Application of text categorization technology could help users to retrieve domain journal papers efficiently. Four phases of text categorization process are “text pre-processing”, “document feature construction”, “applying classification methods” and “evaluation”. This research probes for the effectiveness of: feature weighting, fields of articles and classifiers during the process of journal papers categorization, and also applied sampling distribution classifier within the process. The hypothesis test analysis shows that: 1st, feature ratio performs well significantly than feature frequency. 2nd, fields of abstract are more effective than titles and keywords of journal papers, and there are no difference between the latter two. 3rd, Support vector machines are most effective, then naïve-bayes, decision trees and sampling distribution classifier in order. And 4th, text categorization of journal papers is feasible. Additionally, analysis and recommendation of sampling distribution classifier are also proposed for the future study.
    Appears in Collections:[資訊管理研究所] 博碩士論文

    Files in This Item:

    File SizeFormat

    All items in NCUIR are protected by copyright, with all rights reserved.

    社群 sharing

    ::: Copyright National Central University. | 國立中央大學圖書館版權所有 | 收藏本站 | 設為首頁 | 最佳瀏覽畫面: 1024*768 | 建站日期:8-24-2009 :::
    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - 隱私權政策聲明