Web : Web 14 2 8
Web : Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web i
ii Web (K 0,K 1, K n ) (A 1,A 2, A n ) Web 1 Web Web Web Web Web AND AND (K n,a n,k n 1 ),(K n 1,A n 1,K n 2 ),,(K 1,A 1,K 0 ) n AND Web Web Web Web Web FastMap Web Web Web Web Web Web Web Web Web Web Web
WebSign: Guidance of Peripheral Web Information Space by User Activities and Background Retrievals Shimpei Ikeda Abstract Since the Internet had spread in recent years, Web space has been the important information resource. However, since Web space is flooded with a variety of information, it is not easy to discover useful information on Web space. Although the link navigation is a useful way to explore infomation, the information on the structural and contents-circumference of the page visited nowis not visible. If the information on the circumference space of the page visited now can be acquired, by a map, a traffic-control sign, etc.like the real world, more efficient information search can be performed in Web space. Although many researches on visualization of Web space are done, research which visualizes Web space with the viewpoint from an inside, like the trafficcontro sign in the real world, is not done. In this paper, we propose a system which retrieves the circumference space usingthe visited history as a user s activity in the user s background, and displays the retrieval result in the form of the WebSign which used the trafficcontrol sign metaphor for Web space. The role of the WebSign is considered as what becomes a bookmark and a pair. Bookmark is a guidance to the page visited before, but the WebSign can also display guidance to a non-visited Web page.with the WebSign, a user can recognize circumference space. Then the WebSign support to arrive at the information which a user wants efficiently.we describe the method of generating circumference space, and the method of generating a Web sign in the following texts. We generate the circumference space to the visited history which include the Web page visited now, at the time of link navigation, in order to generate the WebSign. First, to characterize each Web page, we create the feature vector based on the frequency of appearance of each word in a Web page. And we extract the feature keyword and anchor keyword of each Web page as a keyword to generate the circumference space of the Web page visited now. iii
iv Next, we acquire all the sauce of the Web page of the 1 step of link place of the Web page visited now. Next, we rank to the sauce of a Web page, which ranking is as high as the feature vector similar to the average feature vector. We define the Web page group of a ranking higher rank as the structural circumference space of the Web page visited now. Next, we perform AND retrieval using the feature keyword and anchor keyword which were extracted from each Web page. We define the Web page group of the ranking higher rank in each retrieval resultin the result of AND retrieval, as the contents-circumference space of the Web page visited now. By the above operation, we can generate the circumference space of the Web page visted now. In this paper, we project the Web page which generates structural circumference space, the Web page which generates contents-circumference space, and the Web page visited now on the space of a low dimension using FastMap, and, we constitute the WebSign using the projection result. The WebSign is displayed, when a user notifies the demand of a Web sign to a system, or when a user visits the fixed number of pages. If the node on the WebSign is chosen, it will navigate to the Web page corresponding to the node. Moreover, a user can also do the usual link navigation without using the WebSign. A prototype is mounted based on the above-mentioned composition method, and the validity is checked.
Web : Web 1 1 2 3 2.1... 3 2.1.1... 3 2.1.2... 3 2.1.3... 3 2.1.4 FastMap............... 3 2.2... 5 2.2.1 BookMap............... 5 2.2.2 A Metro Map Metaphor.... 5 3 Web 7 4 9 4.1... 9 4.2... 9 4.3... 9 4.4... 10 4.5... 12 5 Web 14 5.1... 15 5.2... 15 6 17 6.1... 17 6.1.1... 17 6.1.2... 17 6.1.3... 17 6.2... 18
6.3... 20 6.4... 21 6.4.1 Web... 21 6.4.2... 22 7 25 26 27
1 Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web Web 1
Web Web Web Web Web [1] Web 2
2 2.1 2.1.1 Web 2.1.2 Web Web Web Web Web 2.1.3 [2] 1997 2 19 version 1.0 JUMAN version 2.0 2.1.4 FastMap FastMap[3] C.Faloutsos FastMap O a O b pivot objects O a O b O a O b d 2 b,i = d2 a,i + d2 a,b 2x id a,b 3
x i x i = d2 a,i + d2 a,b d2 b,i 2d a,b x i O i k k Oi dai dbi Oa E Ob xi dab 1: FastMap FastMap pivot objects pivot objects O a O b (O a,o b ) 2 (O a,o b ) O(N 2 ) C.Faloutsos pivot objects 1. pivot objecto b 2. O b O a 3. O a O b 4. O a O b pivot objects 4
2.2 2.2.1 BookMap BookMap[4] Mountaz Hascoet Web Web Web Mountaz Hascoet 2: BookMap 2.2.2 A Metro Map Metaphor A Metro Map Metaphor[5] Elmer S Web Guided Tour Guided Tour Guided Tour Guided Tour Guided Tour Guided Tour Web Web Web Elmer S 5
3: A Metro Map Metaphor 6
3 Web Web 4 Web 1. 2. Web Web 3. Web 4. Web 5. Web 6. Web Web Web Web 1. 2. Web 4 5 Web 7
Web Web 4: Web 8
4 Web Web Web n (P 0,P 1,,P n )(n 1) n Web P n Web P 0 Web Web 4.1 Web Web Web Web tf 4.2 Web Web Web Web Web Web Web Web Web Web (K 0,K 1, K n ) 4.3 Web Web Web 9
Web Web Web (A 1,A 2, A n ) An An-1 An-2 A1 Pn Pn-1 Pn-2 1 0 Kn Kn-1 Kn-2 K1 K0 5: 4.4 Web Web Web Web Web Web 1 Web Web Web m (n m 1) Web 10
= 6: 1 Web Web Web Web Web Web Web Web Web m 2 1 A B C D 7: 11
Web sim( x, y) = x y x y 4.5 Web Web Web Web AND Web Web AND AND 8 (K n,a n,k n 1 ),(K n 1,A n 1,K n 2 ),,(K 1,A 1,K 0 ) n AND Web Web Web Web Pn Kn An AND Pn-1 Kn-1 Kn-1 An-1 AND Pn-2 Kn-2 K1 A1 AND 0 K0 8: 12
(K i,a i,k i 1 )(n j 1) K i Web,A i,k i 1 Web (K i,a i,k i 1 ) AND Web P i P i 1 Web Web n Ki Ai Ki-1 Ai-1 Ki-2 Pi Pi-1 Pi-2 9: 13
5 Web Web Web Web Web Web Web Web Web Web Web Web Web 10 kakaku.com 4 Web / [ ] 10: Web 14
5.1 Web Web FastMap Web Web Web Web FastMap FastMap Web 11 [ ] / 11: Web 5.2 Web Web Web Web Web Web Web Web Web 12 Web 15
12: Web 16
6 6.1 FastMap Web 2 Web Web 1 2 Web Web 6.1.1 FastMap 2 2 pivot objects Web 2 2 2 pivot objects Web FastMap 2 Web 6.1.2 FastMap Web Web 1 1 2 Web pivot objects Web FastMap 2 Web 6.1.3 1 2 Web Web 17
6.2 Web 13 14 Web Web Web 13: Web 18
Web 14: Web 19
6.3 Web Web Web Web asahi.com Web Web Web HP Web Web Web FastMap Web Web Web Web Web FastMap FastMap Web Web Web Web Web Web FastMap Web Web Web Web Web Web 20
Web Web Web 15: Web 6.4 6.4.1 Web Web Web Web Web Web Web 21
Web Web Web Web Web 16: Web 6.4.2 Web Web Web Web Web Web web 22
4 P n =( Web ) Web Web Web FastMap Web Web Web 23
Pn Pn-1 Pn-2 1 0 Kn An Kn-1 An-1 Kn-2 An-2 K1 A1 K0 17: 24
7 Web Web Web Web Web Web FastMap Web Web 25
26
[1] : Web Proc.of.DBWeb2001 Vol.2001, No.17 pp.201-208 (2001). [2] http://chasen.aist-nara.ac.jp/index.html.ja [3] C.Faloutsos and K.I.Lin : FastMap: A fast algorithm for indexing, datamining and visualization of traditional and multimedia datasets, Proc.ACM SIGMOD June, pp.163-174 (1995). [4] Mountaz Hascoet : A User Interface Combining Navigation Aids Proc.of.the eleventh ACM on Hypertext and hypermedia, pp.224-225 (2000). [5] Elmer S. Sandvad, Kaj Gronbak, Lennert Sloth and Jorgen Lindskov Knudsen : A Metro Map Metaphor for Guided Tours on the Web: the Webvise Guided Tour System, www10 Conference Proceedings, pp.326-333 (2001). 27