臺灣博碩士論文加值系統

隨著WWW的普及，資訊的散佈非常迅速，對知識的累積有很正面的幫助，然而Web上的各種資訊也衍生出負面的問題，就是越來越多不當資訊充斥其間，例如色情圖片與粗暴文字。本論文主要分析如何以色情網站阻擋進行不當資訊防治，同時根據TANet現狀，提出整合網路代理伺服器及網域名稱系統過濾色情資訊的策略，達到可擴充大小(scalable)的網頁過濾架構與分流管制功能。由於網域名稱系統是利用雜湊表記載，詢問快速，加上網域名稱的快取機制，所以過濾系統能夠快速的作出決定，網頁存取延遲時間比傳統產品短，系統耗費資源也很低；本架構可輕易擴充，提供攔截站專業進行網頁分級，而非完全攔阻。另外被攔阻網頁、檔案可利用快取再利用，有效的降低頻寬的需求。
另外由於色情網站的變動性很大，每天都有新的色情網站，固此本系統需要建立精準及完整的管制名單，根據色情網站的特徵，本論文提出三種管制名單更新及建立的方法，第一種 ”Monitor DNS traffic analysis”，是利用網域名稱與色情關鍵字的關係，第二種 ”Pornographic network group links analysis”， 是利用色情網站互相連結的特性。經過評估兩種方法辦認色情網站的精準度皆可達到九成以上，處理速度也比現有方法快上許多，目前在我們已經收集十萬個色情網站的名單，另外我們提出了第三種 ”Proxy access log mining”，可補前兩種方法之不足。同時本系統也安裝於桃園區網中心進行管制。

With the popularity of WWW, fast information distribution greatly speed up the aggregation of knowledge. However, a variety of information on Web pages produces an unexpected problem-abusing information distribution, such as adult and violence. The article addresses how to find out adult contents on Internet and to avoid accessing them by a scalable simple approach. Considering the operation of TANet, we propose DNS-based with proxy approaches to filter abusing information. The approaches utilize existing domain name software as a special server. There is no extra load on the filtering system. Also, because of DNS’s cache mechanism, filtering system can make faster blocking decision.
Pornographic web sites appear and disappear everyday, so it is hard to keep the blocking list up-to-date. We use the pornographic web sites’ feature to propose three kinds of update blocking list methods. The first one is ”Monitor DNS traffic analysis”. Some web sites with keywords in their hostnames is the principle. Besides, the pornographic web sites always link each other. So, we proposed the second method - ”Pornographic network group links analysis”. These methods have higher precision over than 90%. And they have less processing time. Now, we have collected more than 100,000 web sites.
Finally, we implemented a prototype to demonstrate our approaches. The system was installed at National Central University in Taoyuan to test its effect. The experiment shows that our system effectively blocks the retrieval to abusing information. Because of the great effect, many other institutions joint to our system and inform us to adopt the system.

摘要ii
Abstractiii
目錄iv
圖目錄vi
表目錄vii
第1章 緒論1
第2章 不當資訊的防治策略5
2.1收集可疑的不當資訊5
2.2分析判斷不當資訊6
2.2-1影像分析7
2.2-2文字分析7
2.2-3 PICS (Platform for Internet Content Selection ) 標籤12
2.2-4影像分析與文字分析綜合比較13
2.3攔阻策略15
2.3-1黑名單阻擋 (Blocking List)16
2.4攔阻點的配置17
2.5過濾方法相關學術研究18
第3章 系統設計22
3.1過濾系統架構設計22
3.2問題討論28
3.3 可擴充性架構29
3.4黑名單更新設計29
第4章 系統實作及測試36
4.1過濾系統實作36
4.2過濾系統評估與討論39
4.3更新黑名單方法實作與評估42
4.3-1 DNS traffic monitor方法實作與結果評估42
4.3-2 Pornographic network groups similarity方法實作與結果評估45
4.4更新黑名單方法試結果討論48
第5章 結論及未來工作50
5.1結論50
5.2未來工作51
參考文獻52

[1]http://dir.salon.com/sex/world/2000/06/19/nasa/index.html
[2]http://ir.csie.ncku.edu.tw/Project/researchAchievement3.htm
[3]http://www.w3.org/PICS/
[4]Calamaris, http://cord.de/tools/squid/calamaris/Welcome.html.en
[5]Dr. Pornographic Image GateLocker
[6]IMira Screening, http://www.ulead.com.tw/es/imscreening/runme.htm
[7]Microsystems Software, ”CyberPatrol”, http://www.microsys.com/cyber/default.htm
[8]NoPorn, http://www.noporn.com.tw/
[9]Recreational Software Advisory Council, http://www.rsac.org
[10]SafeSurf, http://www.safesurf.com/.
[11]Solid Oak Software, ”CyberSitter”, http://www.solidoak.com/cybersit.htm
[12]Squid, http://www.squid-cache.org
[13]SquidGuard, http://www.squidguard.org/
[14]PORNsweeper, http://www.mimesweeper.com/products/msw/pornsweeper/
[15]Marchiori A., Brodley C., Dy J., Pavlopoulou C., Kak A., Broderick L., Aisen A.M., ”CBIR for medical images - an evaluation trial ,” IEEE Workshop on Content-Based Access of Image and Video Libraries, pp.89-93, 2001
[16]I. Anagnostopoulos, G. Kouzas, C. Anagnostopoulos, D. Vergados, I. Papaleonidopoulos, A. Generalis, V. Loumos and E. Kayafas, ”Automatic web site classification in a large repository under information filtering and retrieval techniques,” Electrotechnical Conference of 11th Mediterranean, pp. 279 -283, 2002.
[17]Wayne B. Salamonsen, Roland Yeo, ”PICS-Aware Proxy System Versus Proxy Server Filters,” Proceedings of INET’97 on Internet Society
[18]Chen Ding, Chi-Hung Chi, Jing Deng, Chun-Lei Dong, ”Centralized content-based Web filtering and blocking: how far can it go?, ”Proceedings of IEEE SMC 99’ Conference, Volume: 17, pp. 48 -57,2002
[19]Feng Jiao, Wen Gao, Lijuan Duan and Guoqin Cui, ”Detecting adult image using multiple features, ”Proceedings of International Conferences on Info-tech and Info-net, Volume: 3, pp378-383, 2001.
[20]Pieper J., Srinivasan S., Dom B., ” Streaming-media knowledge discovery,” Computer, Vol. 34, pp.68-74, 2001
[21]Rose J., Shah M., ”Content-based image retrieval using gradient projections ,” Proceedings of Southeastcon ''98 on IEEE, pp. 118-121, 1998
[22]J.M. Balkin, Beth Simone Noveck, Kermit Roosevelt, ”Filtering the Internet:A Best Practices Model,” 1999
[23]Hoashi K., Inoue, N., Hashimoto, K., ”Data collection for evaluating automatic filtering of hazardous WWW information ,”Internet Workshop on IEEE, pp. 157 -164, 1999
[24]Ki-Wook Kim, Ki-Byoung Kim, Hyoung-Joo Kim, ” VIRON: an annotation-based video information retrieval system,” Proceedings of Computer Software and Applications Conference, pp.298-303, 1996
[25]Andrei Popescu''s , ”Implementation of term weighting in a simple IR system,”,
[26]D. Smith, R. Harvey, Yi Chan and Bangham J.A, ”Classifying Web pages by content,” IEE European Workshop on Distributed Imaging, 1999.
[27]Patrick S. Chen, ”An Automatic System for Collecting Crime Information on the Internet , ” Journal of Information, Law and Technology
[28]Lee P.Y, Hui S.C., Fong A.C.M., ”Neural networks for web content filtering ,” IEEE Intelligent Systems, Volume: 17, pp. 48 -57,2002
[29]林承宇, ”網際網路上「有害資訊內容」之探討─以我國法律管制可行性為中心,” 國立政治大學廣電研究所, 民89。
[30]終止童妓協會Web547, http://www.web547.org.tw