Patent Application
20050160360
This application is based on application number 2004-009337 filed in Japan, dated Jan. 16, 2004.
This invention pertains to a related Web contents search and presentation system that aims to simultaneously present to a user multiple Web pages whose contents are related.
The World Wide Web (WWW), implemented by browsers over the Internet, is practically as common a medium as radio or TV, which makes it possible to obtain news of foreign countries by browsing the contents of foreign news Web sites in countries around the world. Since the content of a news source can be presented in various ways depending on the viewpoint of the press or editors that create the content, and an incident or an accident may be portrayed from various points of view in different countries, audiences or readers may receive markedly different impressions regarding the same news item.
For example, in the case of a news item about a war between country A and country B, a report provided by country A might exaggerate the bad points of country B, and vice versa. The World Wide Web makes it possible for audiences or readers to grasp and compare how a given news item is reported by different news sources in different countries, which has the potential to profoundly change our lives or worldviews.
In order to compare multiple Web pages, however, related pages must be searched for across multiple Web pages and the contents of such related pages must be read by opening a number of Web pages or Web browser instances separately, scrolling each window or viewing pane, and so on. This cumbersome process makes it difficult for the user to compare multiple Web sites. Conventional studies such as “Finding Related Pages in the World Wide Web”, The 8th International World Wide Web Conference (WWW8), Toronto, Canada, May 1999, http://www8.org/w8-papers/4a-search-mining/finding/finding.html by Dean, J., and Henzinger, M., and “Evaluating Strategies for Similarity Search on the Web “The 11th International World Wide Web Conference (WWW2002), Honolulu, Hi., May 2002, http://www2002.org./CDROM/refereed/75/index.html by Haveliwala, T., Gionis, A., Klein, D., and Indyk, P. offer systems wherein similar Web pages are automatically searched.
However, these studies do not indicate methods for controlling the display and presentation of related pages from disparate sources in a unified manner. Furthermore, in order to compare Web sites with content in various languages, not only do multiple Web browsers have to be operated simultaneously, but also multiple languages have to be understood by the user, which makes it even more difficult for the user to browse multiple Web pages with ease.
In order to solve the above-mentioned problems and to make it extremely easy to browse and compare multiple Web pages in various languages, the present claimed invention, as defined in the claims, is arranged so as to find related Web pages among multiple Web sites in various languages designated by a user and to automatically present the contents of such related Web pages at the same time, in a single window, synchronously, under the control of the user, who performs simple operations such as navigating backward, forward, and clicking or scrolling.
The related contents synchronization and presentation system in accordance with the present claimed invention is characterized as comprising a site specifying portion that receives an identifier of a base site as a Web site that becomes the basis of the displayed presentation and specifies this base site and receives an identifier of a compared site as a Web site displayed in a language different from that of the base site and specifies the compared site, a base page specifying portion that specifies a base page as a Web page from the base site shown on a display screen, and a language converting portion.
The language converting portion has a function of consolidating the languages of words included in each site into a single language. The related contents synchronization and presentation system further includes an aggregated word information producing portion that has functions to produce aggregated word information as information based on the words included in a Web page or a paragraph and derived through the language converting portion in a manner that enables comparison with other aggregated word information, a related page specifying portion that specifies one or multiple related pages similar to the base page from among each of the compared pages based on a comparison between the aggregated word information of the base page and the aggregated word information of each compared page as a Web page in the compared site, and a display controlling portion that presents the related page together with the base page on the same display screen.
If the related page specifying portion further has a function of specifying a related paragraph that is similar to a paragraph in the base page, taken from the related pages and based on a comparison between the base paragraph aggregated word information as aggregated word information of a paragraph indicated and displayed in a predetermined position of the display window of the base page and the aggregated word information of each related paragraph in the related page, and the display controlling portion further has a function of displaying these related paragraphs in a predetermined position of the screen, then it will be possible for the user to grasp the similarity of the content of the target paragraphs, thereby facilitating the comparison and browsing of a set of related Web pages.
If the related page specifying portion further has a function of producing difference information between the aggregated word information of each paragraph in the related pages and the base paragraph aggregated word information, respectively, and the display controlling portion further has a function of displaying this difference information for each paragraph of the related pages shown on the display screen in a visually recognizable manner, it will be possible for the user to easily grasp the degree of similarity of each of these paragraphs to the base paragraph, thereby improving usability.
A preferable mode for calculating the degree of similarity for comparison of content represented in different languages is for the related page specifying portion to produce a feature vector whose element is a value obtained by multiplying the frequency of the appearance of each word constituting aggregated word information by a weighting coefficient concerning the part of speech of the words, to calculate a Euclidian distance between feature vectors produced from two elements of the aggregated word information being compared, and to judge the compared page to be similar if the Euclidian distance is less than or equal to a predetermined threshold value.
Concrete aggregated word information is handled as a data structure comprising one or multiple sets for each subject word, that is a word wherein the value obtained by multiplying the frequency of the appearance of the word by a weighting coefficient pertaining to its part of speech is over a predetermined threshold value, and a single content word or group of content words whose degree of cooccurrence with the subject word exceeds a predetermined threshold value, wherein the content word(s) group correlates with the subject word.
When handling word information, it is preferable that the related page specifying portion only compares aggregated subject words appearing in both the base and related page, and only compares aggregated content words that appear in both the base and related page.
Further, in order to present difference information in a way that enhances the user experience by utilizing the comparison of subject and content words, it is preferable that the related page specifying portion consider a paragraph whose subject words are similar by not less than a predetermined degree and whose content words are similar by not less than a predetermined degree as being a globally related paragraph, a paragraph whose subject word alone is similar by not less than the predetermined degree as being a content different paragraph, and a paragraph whose content word alone is similar by not less than the predetermined degree as being a subject different paragraph, by comparison with the aggregated word information of the base page, and that the display controlling portion further has a function of displaying the globally related paragraph, the content different paragraph and the subject different paragraph in a visually and thereby easily recognizable manner.
The related page specifying portion may extract and display multiple related pages, however, it is preferable to specify a compared page that is the most related to the base page as the related page. As a concrete mode for display representation, it is preferable that the display controlling portion presents the base page and the related page in the same display window. In cases where the language of the base site is different from that of the compared site, related words are easily viewable if the display controlling portion presents them in both the base page and the related page in a mode that makes their relationship clear and indicates their difference from other words.
In cases where the system is comprised of multiple, networked computers, it is preferable to have a central information processing unit or server that maintains and operates the site specifying portion, base page specifying portion, language converting portion, aggregated word information producing portion and related page specifying portion, and client stations that operate the display controlling portion of the system.
In addition, a related page synchronization and presentation method in accordance with the present claimed invention is characterized as comprising a site specifying step that receives an identifier of a base site as a Web site that becomes the basis of the displayed presentation and specifies this base site and receives an identifier of a compared site as a Web site displayed in a language different from that of the base site and specifies the compared site, a base page specifying step that specifies a base page as a Web page from the base site and shown on a display screen, a language converting step that has a function of consolidating the languages of the words included in each site into a single language, an aggregated word information producing step that produces aggregated word information whose elements consist of all or some of the words included in a Web page or a paragraph obtained through the language converting step in a manner that enables comparison with other aggregated word information, and a related page specifying step that specifies one or multiple related pages similar to the base page from among each compared pages based on a comparison between the aggregated word information of the base page and the aggregated word information of each compared page as a Web page in the compared site.
Furthermore, a program in accordance with the present claimed invention is characterized by a computer equipped with functions of a site specifying portion that receives an identifier of a base site as a Web site that becomes the basis of the presentation and specifies this base site and receives an identifier of a compared site as a Web site displayed in a language different from that of the base site and specifies the compared site, a base page specifying portion that specifies a base page as a Web page from the base site shown on a display screen, a language converting portion that functions to consolidate the languages of the words included in each site into a single language, an aggregated word information producing portion that functions to produce aggregated word information whose elements consist of all or some of the words included in a Web page or a paragraph obtained through the language converting portion in a manner that enables comparison with other aggregated word information, and a related page specifying portion that specifies one or multiple related pages similar to the base page from among each of the compared pages based on a comparison between the aggregated word information of the base page and the aggregated word information of each compared page as a Web page in the compared site.
The objects and features of the present invention, which are believed to be novel, are set forth with particularity in the appended claims. The present invention, both as to its organization and manner of operation, together with further objects and advantages, may best be understood by reference to the following description, taken in connection with the accompanying drawings.
Reference will now be made in detail to the preferred embodiments of the invention, examples of which are illustrated in the accompanying drawings. While the invention will be described in conjunction with the preferred embodiments, it will be understood that they are not intended to limit the invention to these embodiments. On the contrary, the intention is intended to cover alternatives, modifications and equivalents, which may be included within the spirit and scope of the invention as defined by the appended claims.
Furthermore, in the following detailed description of the present invention, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be obvious to one of ordinary skill in the art that the present invention may be practiced without these specific details. In other instances, well known methods, procedures, components, and circuits have not been described in detail as not to unnecessarily obscure aspects of the present invention.
A. Configuration:
As shown in
The server computer P2 is a multi-purpose computer having, for example, a server function, and comprised, as shown in
Explaining from a functional point of view, as shown in
Each program portion or unit will be briefly explained in the following. The site specifying portion 1 receives an identifier of a base site as a Web site that becomes the basis of the display, and specifies this base site, and receives an identifier of a compared site as a Web site displayed in a language different from that of the base site, and specifies the compared site. The base page specifying portion 2 specifies a base page as a Web page from the base site, with said Web page shown on the display. The language converting portion 4 has functions to consolidate the languages of the words included in each site into a single language.
The aggregated word information producing portion 3 has functions to produce aggregated word information that represents all or some of the words included in a Web page or a paragraph provided by the language converting portion 4, in a manner that facilitates comparison with other aggregated word information. The related page specifying portion 5 specifies one or a number of related Web pages that are appropriately similar to the base page, said pages being selected from each of the compared pages based on a comparison between aggregated word information of the base page and aggregated word information for each of the compared pages of the compared site. The display controlling portion 6 displays the related page on the same display screen together with the base page.
B. Performance: Each portion will be described in combination with an explanation of the performance of this system. (1): A user designates URLs (Uniform Resource Locators, identifiers of Web addresses) of a base site and a site to be compared. An initial page of each site is then displayed as shown in
The above-mentioned sequence of the user's operation is received by an operation receiving portion 7 of the client computer P1, and the site specifying portion 1 and the base page specifying portion 2 of the server computer P2 specify each Web site, as well as the base page, based on the designated URLs. Next, the aggregated word information producing portion 3 verifies whether recent information (for example, content less than 1 hour old) on the designated sites has already been registered in a database D1.
If such information is not registered, the aggregated word information producing portion 3 obtains it, as shown in (2) as described below, obtaining predetermined Web page information from each Internet site. Then each Web page is analyzed and the results are registered in the database D1. If the information is already registered, an operation described in (3) below is conducted.
(2): In this system, word dictionary software is used to compare sites whose languages are different. Since current word dictionary software cannot convert all language content completely, the areas to be compared between two sites are limited to a certain size beforehand, and related pages are found. The areas to be compared are termed compared areas. For example, a news site is classified into each category. The compared areas are determined by finding similar shared categories.
More concretely, as shown in
Next, Topic Structures (aggregated word information, hereinafter called TS) are produced for a Web page in the compared area (hereinafter called a compared page) and the base page, by page and by paragraph. TSs consist of a topic (a word) ti, i ∈(1, . . . , n) where ti is comprised of a set consisting of subject words si and a set of content words Ci. In addition, Ci consists of multiple content words Cim, m ∈(1, . . . , k).
In other words, a TS is defined as follows.
TS={tl, . . . ti, . . . , tn}
ti=(si, Ci)
Ci={cil, . . . , cim}
An above-mentioned subject word is a word whose density is high in a given page (or paragraph) and in this system, subject words are extracted according to a frequency of word appearance measure, in order to obtain relationships of similarity across different pages (or paragraphs). Target words are limited to nouns only. Specifically, for a word t to be considered a subject word candidate, the requirements are as follows.
tf(t)×weight(t)>α
An above-mentioned content word is a word whose degree of cooccurrence with the subject word is high. In this system, a word cooccurrence dictionary D2 is created beforehand and the cooccurrence degree is obtained by the use of this cooccurrence dictionary D2, which has a database structure. If a subject word in a page (or paragraph) is {sl, . . . , si, . . . , sn}, Ci={cil, . . . , Cij} as a group of the content words is obtained for each subject word si. cij is a word whose cooccurrence degree with si is not less than a certain threshold value (β). Content words, like subject words, are also exclusively targeted at nouns. As mentioned above, subject and content words in a page (or paragraph) are determined. Then, the TS of a page (or paragraph) becomes {tl=(sl, Cl), . . . , tn=(sn, Cn)}. TSs are stored in the above-mentioned database D1.
(3): Next, the related page specifying portion 5 produces a feature vector for subject words and a feature vector for content words from the TSs stored in the above database, compares the base page and each compared page by the use of respective Euclidian distances, and obtains content difference information based on TSs at a page level and TSs at a paragraph level. A paragraph that is to be a basis for comparison at a paragraph level is, for example, a paragraph (hereinafter called as a base paragraph) located in the center portion of the base page shown in the left portion of the display screen. Prior to this, the dictionary software (language converting portion 4) converts compared site's language subject and content words into corresponding base site's language subject and content words. In cases where multiple base site's language versions correspond to a single compared site's language word, multiple base site's language versions become candidates for the particular subject word.
A similarity degree of a given subject word, Sims, and a similarity degree of a given content word, Simc, between the two pages (or paragraphs), are determined as follows.
Sims=[(fb(sl)−fc(sl))2+. . . +(fb(sn)−fc(sn))2]1/2
Simc=[(fb(cl)−fc(cl))2+. . . +(fb(cm)−fc(cm))2]1/2
Fb(sn)=tf(sn)×weight(sn)
Global similarity: In cases where the similarity degrees of the subject and content words are less than or equal to a certain threshold value, the two pages (paragraphs) become similar in a global sense, that is, become related pages (or paragraphs) because the subject words of one page (or paragraph) are similar to the other and the content words of one page (or paragraph) are similar to the other.
Content difference: In cases where the similarity degree of a subject word is less than or equal to a certain threshold value and the similarity degree of a content word is greater than the threshold value, the two pages are similar regarding the subject word but different regarding the content word, so the status as a page (or paragraph) with different content is clearly recognized.
Subject difference: In cases where the similarity degree of a subject word is greater than a certain threshold value and the similarity degree of a content word is less than or equal to a certain threshold value, the two pages are different regarding the subject word but similar regarding the content word, so the status as a page (or paragraph) with a different subject is clearly recognized.
A compared page whose Euclidian distance from the base page is the smallest is set as the most related page among the set of compared pages judged to be globally similar to the base page. A paragraph that is judged to be globally similar to the base paragraph and whose Euclidian distance from the base paragraph is the smallest is set as the most related paragraph among the paragraphs in the related page.
(4): The display controlling portion 6 displays the specified related page in the same window as that of the base page. At this time, the display controlling portion 6 controls the display mode so as to simultaneously display the base paragraph and the related paragraph, for example, at nearly the same horizontal level on the display screen (refer to
As a result, when the user scrolls the base page, the base paragraph is changed. Then, the related page is automatically scrolled synchronously so as to display a related paragraph corresponding to the changed base paragraph, displaying both in a manner as noted above that maintains a horizontal congruence (refer to
Web pages usually consist of multiple paragraphs and it often happens that these multiple paragraphs are displayed in a single window. In this system, two Web pages are presented simultaneously and navigated synchronously by the operation of a user, who may have difficulty grasping the differences between similar and different paragraphs in the compared page. Therefore, in this embodiment, a globally related paragraph, a content different paragraph, and a subject different paragraph selected by the related page specifying portion 5 can all be presented in visually differentiated styles. Several methods are conceivable for carrying out this visualization, and here, such differentiation is presented by means of separate colors. This makes it possible for the user to clearly recognize which paragraph is a related paragraph or which paragraph is a paragraph with differences in the compared page.
More concretely, as the pattern diagram shown in
Furthermore, since two pages in different languages can be simultaneously presented by the system, it may be difficult for the user to find specifically correlated words. To meet this need, when the user selects a word in the base page, it is displayed in a different mode from other words so that correlated words in the base page and in the related page are clearly highlighted in the display, as shown in
In addition, although not shown in these drawings, in cases where there is no paragraph related to the base paragraph in the compared page, this system searches for related paragraphs in other compared pages. When a paragraph similar to the base paragraph is found among other compared pages, an icon indicating partially similarity is displayed in the paragraph in the base page. When the user clicks this particular icon, the system displays the relevant page, including the related paragraph, in another window, as a partially similar page.
C Summary: As mentioned above, in accordance with this embodiment, simply navigating to and within a Web site written in one language (such as Japanese) enables the presentation and operation of a related site written in another language (such as English), synchronously and automatically, and enables differences in the information contained in each paragraph in the related pages to be grasped visually, thereby greatly facilitating the browsing and comparison of multiple pages presented in a variety of languages.
The present claimed invention is not limited to this embodiment. For example, a method for presenting difference information is offered that makes use of differently colored paragraphs in the above embodiment, however, simply presenting and aligning related paragraphs according to a coloring scheme merely shows that they contain related information but does not make clear the nature or details of the relationship between the paragraphs. Thus, another visual presentation method for this type of information may be adopted, one that clearly delineates and explicates the particulars of these relationships. In addition, multiple sites in multiple languages may be presented comparatively, all referring to a single base site, and language conversion may be conducted when registered in database D1.
Those skilled in the art will appreciate that various adaptations and modifications of the just-described preferred embodiment can be configured without departing from the scope and spirit of the invention. Therefore, it is to be understood that, within the scope of the amended claims, the invention may be practiced other than as specifically described herein.
| Number | Date | Country | Kind |
|---|---|---|---|
| P2004-009337 | Jan 2004 | JP | national |
