基於多元化部落格網頁之自動化擷取部落格主要文章; Automatic Extraction of Blog Post from Diverse Blog Pages

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/48420

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/48420

Title:	基於多元化部落格網頁之自動化擷取部落格主要文章;Automatic Extraction of Blog Post from Diverse Blog Pages
Authors:	陳志銘;Jhih-ming Chen
Contributors:	資訊工程研究所
Keywords:	最大加總子序列;序列標記;資訊檢索;部落格;blog post extraction;sequence labeling;maximum subsequence
Date:	2011-07-22
Issue Date:	2012-01-05 14:54:20 (UTC+8)
Abstract:	近年來，部落格為主的相關研究蓬勃發展，例如：意見檢索、情緒分析。因此，擷取部落格的主要文章即是一個不可或缺的步驟。在此篇論文中，我們將探討如何從各式各樣的部落格網頁精確且自動化的擷取部落格的主要文章。許多先前的研究著重於擷取新聞網頁的主要文章，若將其應用於部落格網頁並無顯著的效果，這是由於部落格網頁風格五花八門且文章內容包含多種格式，致使擷取部落格主文變得較為複雜。針對此問題，我們結合MSS [24] 和CETR [34] 這兩篇論文的研究並加以修改調整，提出兩個部落格主文擷取的方法。第一個方法為PTR Scoring，結合了Post-to-Tag Ratio和Maximum Scoring Subsequence，是一個非監督式演算法。第二個方法為CRF Scoring，透過Conditional Random Fields此機率模型並利用Maximum Scoring Subsequence提升擷取的準確率。實驗結果顯示CRF Scoring的F-Measure可達到91.9%，是本篇論文中準確率最高的擷取方法。本篇論文所提出之方法可應用於PDA、手機…等螢幕較小的裝置，以及提升部落格搜尋引擎的效能，並提供後續相關研究之參考與幫助。 With the rapid development of the blogosphere, blog post extraction is an essential task for researches on blogosphere. However, very little attention has been given specifically to blog post extraction. In this paper, we address the issue of extracting blog posts from diverse blog pages, which aims at automatically and precisely finding the location of each blog post. Most of previous researches focused on extracting main content from news pages, but the problem becomes more complex when one turns to blog pages, since some blog posts may employ a variety of content formats concurrently and miscellaneous information could negatively affect the accuracy of extraction. Our research is based on the combination of MSS [24] and CETR [34] to develop algorithms that are suitable for blog pages. The 1st method that we propose is PTR Scoring, which combines Post-to-Tag Ratio with maximum scoring subsequence. The 2nd method is CRF Scoring, which applies Conditional Random Field to train models and use maximum scoring subsequence to improve the accuracy of extraction. The experimental results show that CRF Scoring achieves the best F-Measure at 91.9% among existing methods.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	473	View/Open

社群 sharing

Loading...