JP2001331529A - Device and method for presenting news information - Google Patents

Abstract

PROBLEM TO BE SOLVED: To select, for example, useful information which is generally high in topicality from a large amount of news information, to reconstitute the useful information in proper form according to the purpose of a user, and to provide the information. SOLUTION: This device is equipped with a means 2 which divides news information distributed in the form of electronic mail, a magazine, etc., into article units, a means 3 which selects useful information out of the divided articles, and a means 4 which reconstitutes news information in format to be presented to the user by gathering the selected articles and outputs the information.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for presenting news information, and more specifically, for example, selects useful information having a general topic from a large amount of news information distributed by e-mail or the like. In addition, the present invention relates to a news information presentation device and a presentation method for reconstructing and presenting the information in an appropriate form according to the purpose of the user.

[0002]

2. Description of the Related Art With the recent spread of personal computers and the Internet, a large amount of information can be easily obtained on an individual level. As one of information collection methods using the Internet, there is an information distribution service using electronic mail. In this service, by registering a mail address in advance as a distribution source, news and various kinds of information can be automatically distributed by e-mail, and the latest information can be known while staying. Unlike WWW, there is no need to access a specific site each time, and high-quality information can be easily obtained.

[0003] However, when information is distributed by such a service, there is a problem in that the amount of information to be distributed is too large and it is difficult to read all of the information. That is, it is difficult for the user to find useful information for the user from the distributed information.

In such a case, as a typical technique for retrieving an article of interest to a user, a retrieval request is input to the system using a keyword or the like corresponding to the content that the user wants to retrieve. There is a technology in which articles that meet requirements are searched and presented.

As another technique, there is also a technique called content filtering in which a user registers a search request in advance in the form of a user profile, and only articles that match the search request are selected and presented to the user. is there.

[0006]

As described above, when news is distributed by e-mail or the like, a plurality of articles including a plurality of topics are distributed in a single distribution in many cases. In addition, emails may contain non-news parts, such as advertisements. Therefore, there is a problem that it is difficult to obtain information efficiently by simply retrieving an article including a certain topic by simply storing the mail.

Next, when searching for an article, the user must input a search request, and the user must have a clear search intention. In addition, for content filtering and the like, it is necessary to give a topic of interest to the system in advance in the form of a keyword or the like before distribution of an article. Therefore, if the user does not have a clear search intention of what information they want,
There is a problem in that when it is desired to know an article that is widely talked about regardless of the field of the article, such information cannot be extracted well.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to extract a general topical article from news information distributed within a certain period specified by a user and to provide the user with the information. An object of the present invention is to provide a news information presenting apparatus and a presenting method. Another problem of the present invention is that, when presenting highly topical information, by reconstructing a plurality of articles to be presented as news information, news information to be presented in accordance with a user's request or a restriction on a display screen, etc. A presentation device and a presentation method are provided.

[0009]

FIG. 1 is a block diagram showing the principle configuration of a news information presenting apparatus according to the present invention. FIG. 1 is a block diagram showing the principle configuration of a news information presenting apparatus that extracts, for example, generally highly topical information as useful information from a large amount of news information distributed by e-mail and presents it to a user. It is.

In FIG. 1, a news information presenting apparatus 1 of the present invention comprises a news information dividing means 2, a useful article selecting means 3,
And news information reconstructing means 4. The news information dividing means 2 divides news information distributed by e-mail into articles, for example.

The useful article selecting means 3 is for selecting an article useful for the user, for example, an article having high topicality, from the articles divided by the news information dividing means 2.

The news information reconstructing means 4 collects the articles selected by the useful article selecting means 3 and reconstructs and outputs the news information in a format to be presented to the user.

In the embodiment of the present invention, the news information dividing means 2 determines the article part in the news information by analyzing the format of the news information, and divides the news information into article units.

In the embodiment of the present invention, the useful article selecting means 3 may include an article grouping means, an article group selecting means, and a representative article selecting means. The article grouping means groups the divided articles into a plurality of groups in accordance with the similarity between the articles divided by the news information dividing means 2.

[0015] The article group selecting means selects one or more groups from the plurality of groups according to the topic level indicating the degree of topicality of each of the plurality of groups grouped by the article grouping means. It is.

The representative article selection means selects one or more representative articles from each of the one or more groups selected by the article group selection means. In this case, the article group selecting means can also calculate the topic level based on the total number of articles belonging to the group and the number of information sources of articles belonging to the group.

In the embodiment of the present invention, the news information presenting apparatus 1 can further comprise a user instruction input means for receiving an instruction from the user regarding the presentation of news information, and the representative article selecting means is provided with a representative article selecting means. One or more representative articles can be selected from each of the groups in response to the instruction.

In this case, the representative article selecting means can also select a representative article based on an instruction from the user regarding the distribution date of news information. In the embodiment of the present invention, the news information reconstructing means 4 selects one or more groups in the output order corresponding to the topic level of the selected one or more groups and / or the distribution date of the representative article. May be output.

In this case, the news information reconstructing means 4 is 1
In addition to one or more representative articles for one or more groups, one or more representative articles for each of one or more groups may be further output as related articles, or a simplified form in which the contents of the representative articles are partially omitted. It is also possible to output a representative article and set a hyperlink to the original representative article in the simplified representative article. It is also possible to output figures and tables together.

In the embodiment of the present invention, the news information reconstructing means 4 outputs an output corresponding to the topic level of one or more groups selected by the article group selecting means and / or the distribution date of the representative article. In order, the content of one or more representative articles corresponding to each of the one or more groups can be converted into a format suitable for voice reading and output.

In the news information presenting method of the present invention, news information is divided into articles, useful articles are selected from the divided articles, and the selected articles are collected and presented to the user. A method of reconstructing and outputting the news information is used.

In the embodiment of the present invention, when selecting useful articles in this method, articles divided according to the similarity between the divided articles are divided into a plurality of groups. And one or more groups are selected from a plurality of groups in accordance with the topic level indicating the degree of topicality of each group, and one or more representatives are selected from each of the selected one or more groups. An article selection method can also be used.

As the computer-readable portable storage medium of the present invention, a step of dividing news information into articles, a step of selecting useful articles among the divided articles, and a step of collecting the selected articles A portable storage medium storing a program for causing a computer to execute the steps of reconstructing and outputting news information in a format presented to a user is used.

In the embodiment of the present invention, the step of selecting useful articles among the steps stored in the portable storage medium is performed in accordance with the similarity between each of the divided articles. Grouping the divided articles into a plurality of groups, selecting one or more groups from the plurality of groups according to the topic level indicating the degree of topicality of each of the groups, Selecting one or more representative articles from each of the above groups.

The program according to the present invention includes a procedure for dividing news information into article units, a procedure for selecting useful articles from the divided articles, and a form for collecting and presenting the selected articles to the user. And a procedure for causing the computer to execute the procedure of reconstructing and outputting the news information.

As described above, according to the present invention, news information distributed, for example, in the form of e-mail is divided into individual articles, articles whose contents are related are grouped, and representative articles in the group are It is reconstructed into news information and output.

[0027]

FIG. 2 is an overall explanatory diagram of a news information presentation system according to the present invention. In FIG. 1, a plurality of information sources 11 a to 11
News information is distributed from 1c via the network 12, and the distributed news information 13 is given to the news information presentation device 10. Here, the information source is not limited to, for example, an electronic mail that is distributed via the network 12, but may be a news information that is also obtained from a magazine or the like and is also given to the news information presenting apparatus 10. I do.

The news information presenting apparatus 10 includes an article dividing section 14 for dividing news information into individual articles, an article selecting section 15 for selecting articles divided in response to a request from a user, etc. The system includes an article reconstructing unit 16 for reconstructing an article selected in response to a request from a user in the form of news information, and a presenting unit 17 for presenting the reconstructed news information to a user.

FIG. 3 is a block diagram showing the configuration of the news information presenting apparatus of the present invention. 2, a news information presentation device 20 includes an article division / article information extraction unit 21, an article selection unit 15, an article reconstruction unit 16, a presentation unit 17, and an article information division corresponding to the article division unit 14 in FIG. Is provided.

The article division / article information extraction unit 21 reads the distributed news information 13, extracts the article part, divides it into individual articles, and obtains information such as the information source of each article, the article distribution date, and the like. The information necessary for calculating the similarity of is extracted, and the result is output to the article information storage unit 22.

The article selection section 15 is composed of a related article group generation section 25, a topic degree calculation section 26, and a representative article selection section 27, reads article information stored in the article information storage section 22, and is useful for the user. And outputs the result to the article reconstruction unit 16.

The related article group generation unit 25 retrieves, for example, articles distributed during a period designated by the user from the article information storage unit 22, calculates the similarity between the articles, and the contents are related. Collect related articles and generate related article groups.

The topic level calculation unit 26 calculates the topic level indicating the degree of topicality for each of the related article groups generated by the related article group generation unit 25.
This calculation will be described later.

The representative article selecting section 27 includes a topic degree calculating section 26
The related article group having high topicality is selected based on the degree of topic of each group calculated by the above, and a representative article and a related article are selected from the selected related article group. The selection of the representative article and the related article will be described later.

The article reconstructing section 16 collects the representative articles selected by the representative article selecting section 27 of the article selecting section 15 and related articles, determines the output order thereof, and, for example, matches the user's use. Articles are arranged according to an appropriate form and layout, and news information to be presented to users is generated. The presentation unit 17 presents the user with the news information generated by the article reconstruction unit 16.

FIG. 4 shows the article division / article information extraction unit 2 of FIG.
3 is a flowchart of a process by No. 1. When the process is started in the figure, first, news information is read in step S1, and format analysis of the news information is performed in step S2. Here, it is assumed that the format is already determined, and the format analysis method will be described later.

As a result of the format analysis, a portion corresponding to an article is extracted in step S3, and divided into individual articles in step S4. Then, in step S5, the distribution source (information source) of the article, the distribution date, the headline of the article, the text of the article, and the like are extracted for each article.
An article is represented by the information extracted in step 6, the result is stored in the article information storage unit 22 in FIG. 3, and the process is terminated.

FIG. 5 is a processing flowchart of the article division / article information extraction unit when an article is also extracted from information such as a magazine. Compared to FIG. 4, the only difference is that after news information is read in step S1, text information is extracted in step S10, and then format analysis of news information is performed in step S2.

When extracting article information from a magazine or the like, it is necessary to extract only text information, excluding information such as photographs and figures. Various techniques can be used for this processing. For example, information given in the form of image data can be handled by using a character recognition (OCR) technique. As described above, the format is determined in the format analysis, but the format conversion to a certain format can be executed in the text information extraction process.

FIG. 6 shows an example of the data structure of article information stored in the article information storage section 22. In the figure, article information for one article includes an article identifier (ID), an information source,
It consists of a distribution date, an article headline, an article body, chart data, supplementary information, and an article length.

The article ID is a unique value of the article, and is used to identify each article and to correspond to the original article. The information source is the name of the information source that distributed the article, for example, the name of a newspaper. The chart data is to store the data as needed when the article includes a chart or table. The supplementary information is not directly related to the article body, such as link information to a related article, but is data that plays a supplementary role, such as a URL address.

The format analysis in step S2 in FIGS. 4 and 5 is performed based on rules prepared in advance. Examples of rules are "" ... "and"-
Lines in which specific characters such as "-" appear consecutively are dividing lines that divide an article or an article headline from the main body of an article. "The section divided by the dividing line has only one line. Is the headline of the article. "," If the number of lines in the section divided by the dividing line is multiple lines, it is the body of the article. "," Lines beginning with "http: //" access related information. This is link information for performing the operation. "

FIG. 7 is an example of news information distributed by electronic mail. FIG. 8 shows a result of performing a format analysis on this example using the above-described rule. In FIG.
There are lines where the same symbol continues, such as "...". First, this is regarded as a separator line, and the text is divided. Next, the number of lines in each divided section is checked,
For example, a single line is recognized as an article heading, and a plurality of lines is determined as an article body. Then, from the top, the article heading and the article body are set as one section, which is taken out as one article, and the text of FIG. 7 is divided into two articles.

FIG. 9 is another example of news information distributed by electronic mail. The text in FIG.
There is a line in which the same symbol character continues as “−”. First, this is regarded as a separator and the text is divided. The determination of the article heading and the article body is the same as in FIG. 7. “http: The line starting with "/" is regarded as link information for accessing related information, is separated from the article body, and is stored as supplementary information in FIG.

FIGS. 10 and 11 show the results of extracting the headline of each article and the body of the article from the text shown in FIGS. 7 and 9 and storing them in the article information storage unit 22. FIG. Here, the distribution date and the information source are described in the header portion of an e-mail, for example, and can be easily extracted. Note that the article numbers 05 and 06 for the two mails in FIG. 9 are unnecessary information and are not stored in the article information storage unit 22.

FIG. 12 is an overall flowchart of the process performed by the article selection unit 15 of FIG. When the process is disclosed in the figure, first, in step S11, for example, a target period input from a user is received. The user wants to present, for example, an article with high topicality within this target period as a useful article.

Subsequently, in step S12, a related article group is generated by the related article group generation unit 25 in FIG. In the process of generating a related article group, the calculation of the similarity between two articles of all combinations is generally performed for all articles within a target period designated by a user as described later. This is performed, and articles having a high degree of similarity are collected as one related article group.

Subsequently, in step S13, the topic level of each related article group is calculated by the topic level calculation unit 26 in FIG. The calculation of the topic level will be described later. Finally, in step S14, the representative article selecting unit 27 selects a useful article, for example, a representative article to be described later, from among the related article groups, for example, for a related article group having a high topic level. finish.

FIG. 13 is a flowchart of the process performed by the related article group generation unit 25. When the process is disclosed in the figure, first, in step S16, the article information of all the articles distributed within the target period designated by the user is read from the article information storage unit 22, and in step S17, all the articles are read. Is calculated for any two articles.

The similarity S between _two articles A ₁ and A ₂
Is calculated using the following formula.

[0051]

(Equation 1)

Where w _i and w _j are articles A ₁ and A, respectively.
₂ , W _ij are the longest matching partial character strings in these keywords, | w _i |, w _j |, |
W _ij | is the length of each keyword and the longest matching partial character string, and n (A ₁ ) and n (A ₂ ) are the number of keywords extracted from each article.

The calculation of the degree of similarity will be further described with reference to FIGS. FIG. 14 is an explanatory diagram of a method for extracting keywords from two articles. Here, it is assumed that keywords are extracted from the headlines of two articles. From the headline of the article, we divide the headline by considering characters that are hard to be words that directly indicate the content of the article, such as hiragana, parentheses, punctuation, etc. I do.

Six keywords are extracted from the heading of the upper article, and five keywords are extracted from the heading of the lower article. Therefore, the number of keywords n (A ₁ ) for the article A ₁ above is 6, and the number of keywords for the article A ₂ below is n (A ₂ ) 5.

FIG. 15 is an explanatory diagram of the calculation of S _ij by equation (2) using the longest matching partial character string between two keywords. In the figure, if the upper “TFT liquid crystal panel” is w _i and the lower “liquid crystal panel technology” is w _j , the “liquid crystal panel” is the longest matching partial character string, and its length is 5,
Also, w _j has a shorter keyword length, and its length is 7
Since it is, the value of the keyword among the similarity S _ij for the two keywords 5/7, that is, 0.714.

In the calculation of the similarity S corresponding to FIG. 14, the value of S _{ij is} similarly calculated between each of the upper keywords and each of the lower keywords, and the sum thereof is calculated as S is calculated as a value divided by n (A ₂ ) = 5 extracted from the article headline.

Returning to FIG. 13, when the similarity in the case where any two articles are combined is calculated for all the articles read in step S17, in step S18 the similarity is high according to the calculation result. The article group is put together as a related article group, and the process ends. In the related article group generation process, a grouping is performed such that a threshold having a similarity is provided and two articles having a similarity greater than or equal to the threshold are put into the same group. This threshold value is a relative value, and for example, preprocessing such as preparing data in advance and finding an appropriate value in advance is required.

FIG. 16 is a processing flowchart of the topic level calculation unit 26 of FIG. When the process is started in the figure, first, in step S20, information of articles belonging to each related article group is read, and in step S21, the total number of articles and the number of information sources are calculated for each related article group. In step S22, the topic level is calculated for each related article group based on these calculation results, and the process ends.

The topic level T is calculated by the following equation.

[0060]

(Equation 2)

Where m is the total number of articles belonging to the group,
n is the number of information sources that have delivered articles belonging to the group, and m _i (i = 1, 2,..., m) is the number of articles for each information source.

Generally, one group is considered to be a group of articles corresponding to one topic, and therefore, the greater the total number m of articles belonging to the group, the higher the topic level. Also, the more sources that distribute those articles,
For example, it is a topic reported by many newspapers, and it is considered that the topic level is high. In equation (3), a weight is particularly given to the number of information sources, and mn ² is multiplied before the radical.

Even if the total number of articles and the number of information sources are the same, when articles are distributed equally from each information source, the number of articles output from each information source is different from that when the number of articles from each information source is different. It is considered highly topical. For example, the total number of articles m =
6, when the number of information sources is n = 3, and two articles appear from each of the three information sources (m ₁ = m ₂ = m ₃ = 2), the value of equation (3) is 18 For example, two from the first information source (m ₁ = 2) and three from the second information source (m
₂ = 3), when one (m ₃ = 1) article is distributed from the third information source, the value is outside 1 and the article is distributed equally.

[0064]

[Outside 1]

In this case, the topic level is higher. FIG. 17 is a flowchart of the process by the representative article selecting unit 27 in FIG. When the process is disclosed in FIG.
At 24, a related article group having a high topic level is selected as a main related article group. In this selection, for example, a threshold value of the topic level is determined in advance, and a related article group whose topic level exceeds the threshold value is selected as a main related article group. This threshold is also relative. Alternatively, related article groups are arranged in descending order of topic level, and a certain number of related article groups at the top are selected as main related article groups.

Subsequently, in step S25, a representative article is selected for each of the selected main related article groups, and the process ends. The representative article is used to select a representative article from a group of articles belonging to the main related article group, for example, in response to a user request. For example, if the user is mainly requesting the latest information, the article with the latest delivery date is required, and if the user is requesting progress information on how the situation has progressed on a certain topic, the delivery date is relatively old. The length of the article body is the longest if the article, for example, one week ago, seeks more information about a topic, for example, or the length of the article body if it seeks concise information Is selected as the representative article. The number of representative articles is not limited to one for each main related article group, and a plurality of articles are selected as representative articles as needed.

FIG. 18 is an overall flowchart of the process of the article reconstructing unit 16 of FIG. When the process is started in the figure, first, in step S27, it is determined in the output order of the article group selected by the representative article selecting unit 27 of the article selecting unit 15. The output order may be, for example, an order in which the degree of topic with respect to each main related article group is high, or an output order in which the distribution date of each representative article is new.

Next, the article information of the article group selected in step S28 is read from the article information storage unit 22, and the article information read in step S29 is processed as necessary according to, for example, a request from the user. The result is arranged according to the output order and the layout according to the use of the user, output to the presentation unit 17, and the process ends.

FIG. 19 is a detailed flowchart of the article information arranging process in step S29 of FIG. When the process is started in the figure, first, the headline and the article text of the article to be arranged are extracted from the article information read in step S28 in step S31, and the arrangement of the headline is performed in step S32. In S33, it is determined whether or not the article body is also displayed.
At 36, it is determined whether or not there is another article to be arranged. If there is another article, the processing after step S31 is repeated, and if there is no other article, the processing is terminated as it is.

If the article text is also to be displayed in step S33, the article text is processed in step S34, for example, in response to a request from the user, and the processed article text is arranged in step S35, and step S36 is performed. The following processing is performed. As for the processing of the article body, necessary processing is performed according to whether the request from the user is to publish the whole text, the beginning of the article, or the summary.

FIG. 20 shows an example of a request input screen for receiving a request from a user for presenting news information. In FIG. 2, the request from the user is given to the article selection unit 15, but in FIG. 3, this request is sent to step S1 in FIG.
1 is input to the related article group generation unit 25 for input of the target period, is also input to the representative article selection unit 27 for the above-described representative article selection, and is also transmitted through the article selection unit 15. It is also possible to provide to the article reconstructing unit 16, and of course, directly to the article reconstructing unit 16 without going through the article selecting unit 15.

In the example of FIG. 20, the target period is from October 1 to 1
The number of topics to be displayed, that is, the number of related article groups is set to three, and two representative articles should be selected from new ones for each topic. Is required to put the article text in a summarized form.

FIGS. 21 to 24 show examples of article restructuring.
In the first example of FIG. 21, only one representative article is selected from each main related article group, and the headline of the article and the text of the article are listed. The information source of the article and the distribution date information are also output.

The second example of FIG. 22 is an output example when a plurality of representative articles are selected from one main related article group. For example, one optimal article that meets the user's request is output as a representative article, and the other representative articles are output as related articles. Although not shown, when a representative article includes a figure or a table, the figure data can be read from the article information storage unit 22 and output together with the article text.

When the content of an article is output, the entire text of the article may be output, or the user may request a simple output, or the amount of information that can be output on the display screen is limited. It is also possible to output only a part of the article body. In this case, for example, it is conceivable to output a predetermined amount from the beginning of the article body, or to create and output an article summary using the article headline and the article body. The following literature describes the generation of such a summary sentence.

Japanese Patent Application Laid-Open No. 10-207891 Nakao
Yoshio "Document Summarization Apparatus and Method" FIG. 23 is a third example of article reconstruction when outputting only a part of the article text or generating and outputting a summary sentence. In such a case, by setting a hyperlink in the headline of each article as shown as an underline in FIG. 23, it is also possible to access the contents of the full text of the original article. This can be realized by a known technique such as HTML (Hyper Text Markup Language).

FIG. 24 shows a fourth example of article reconstruction.
One representative article on the left for each main related article group,
The headlines of related articles as other representative articles are displayed on the right side, only a part of the article text is output, and hyperlinks are set for the article headings.

As shown in this example, methods for outputting related article information, outputting only part of the article text, outputting charts, and setting hyperlinks can be arbitrarily combined. Of course it is possible.

Next, as a process in the article reconstructing unit 16, it is possible to perform reconstructing in a format suitable for reading out aloud. FIG. 25 shows an example of an article reconstructed result in a format suitable for such reading aloud. In the case of voice reading, it is desirable that the output content be as simple as possible. For example, as shown in FIG. 25, a method of arranging only the headings of articles may be considered.

Finally, the presentation unit 17 shown in FIG. 3 actually presents the news information reconstructed by the article reconstructing unit 16 to the user. Can be displayed on a display or printed using a printer. In addition to the display on the display screen, it is also conceivable to use voice output of a text-to-speech output text using voice synthesis technology. .

FIG. 26 is an explanatory diagram of a different embodiment. In this different embodiment, the related article group generation unit 2 of FIG.
When determining the main related article group from the article group generated by the above, without calculating the topic level,
The main relevant article group is determined using another source. For example, a representative article previously selected from each related article group in response to a request from the user, a heading of each one, and an article heading of another weekly magazine are compared by calculation of similarity, and the similarity is large. The related article group to which the representative article belongs is designated as the main related article group.

Here, among the article headings of another weekly magazine,
Only the comparison with the first headline is performed, but the comparison with the last headline also designates the representative article at the bottom, that is, the related article group to which "Electronic Show '99 opens" as the main related article group. Will be.

As a final explanation of the embodiment, loading of a program used in the present invention into a computer will be described. For example, the news information presenting apparatus of FIG. 3 is generally realized by a personal computer placed at the user's hand, for example, but may also be realized by a computer installed on the side of a business operator providing a news information presenting service. it can. For example, the computer executes the programs of claims 14 and 15 in the claims of the present invention, and the programs shown in the flowcharts of FIGS. 4, 5, 12, 13, and 16 to 19. Accordingly, news information can be presented.

FIG. 27 is an explanatory diagram of loading of such a program into a computer. In the figure, a computer 51 for realizing a news information presentation device is shown.
Has a main body 52 and a memory 53. The memory 53 is a random access memory (RAM), a hard disk,
It is a memory such as a magnetic disk, and a program is stored in such a memory, and when the program is executed by the main body 52, news information can be presented to the user.

Such a program is provided from the program provider to the computer 5 via the network 54.
1 or a commercially available portable storage medium 5
5 and the portable storage medium 55 is loaded into the main body 52, so that the news information of the present invention can be presented. As the portable storage medium 55, various types of storage media that are commercially available and can be distributed, such as a floppy disk, a CD-ROM, an optical disk, and a magneto-optical disk, can be used.

[0086]

As described above in detail, according to the present invention, the distributed news information is divided into individual articles, and the articles related to each other are grouped into related article groups to reconstruct the news information. Since the information is presented, the user can efficiently obtain useful information. The topic level can be calculated for a group of articles with related contents, and only those with a high topic level can be presented to the user. If the user does not have a clear search intention, Regardless of this, useful information can be obtained when it is desired to know information that has been widely talked about in general, greatly contributing to the improvement of the practicality of the news information presenting apparatus.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the principle configuration of a news information presenting apparatus according to the present invention.

FIG. 2 is an explanatory diagram of a news information presentation method according to the present invention.

FIG. 3 is a block diagram illustrating a configuration of a news information presentation device.

FIG. 4 is a processing flowchart (part 1) of an article division / article information extraction unit.

FIG. 5 is a processing flowchart (part 2) of an article division / article information extraction unit.

FIG. 6 is an example of a data structure of article information stored in an article information storage unit.

FIG. 7 is a first example of news information distributed by e-mail.

8 is a diagram showing a result of format analysis on the news information of FIG. 7;

FIG. 9 is a second example of news information distributed by e-mail.

10 is an example (part 1) of data stored in the article information storage unit for the news information of FIGS. 7 and 9; FIG.

FIG. 11 is an example (part 2) of data stored in the article information storage unit for the news information of FIGS. 7 and 9;

FIG. 12 is an overall flowchart of a process in an article selection unit.

FIG. 13 is a processing flowchart of a related article group generation unit.

FIG. 14 is a diagram illustrating extraction of a keyword for calculating a similarity;

FIG. 15 is a diagram illustrating the similarity between keywords between two keywords.

FIG. 16 is a processing flowchart of a topic level calculation unit.

FIG. 17 is a processing flowchart of a representative article selecting unit.

FIG. 18 is an overall flowchart of a process performed by an article reconstruction unit.

FIG. 19 is a detailed flowchart of article arrangement by an article reconstructing unit.

FIG. 20 is a diagram illustrating an example of a screen on which a request from a user is input.

FIG. 21 is a first example of article reconstruction.

FIG. 22 is a second example of article reconstruction.

FIG. 23 is a third example of article reconstruction.

FIG. 24 is a fourth example of article reconstruction.

FIG. 25 is an example of speech-reading news information.

FIG. 26 is an explanatory diagram of an embodiment in which a main related article group is designated using another information source.

FIG. 27 is a diagram illustrating loading of a program into a computer according to the present invention.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 news information presenting device 2 news information dividing means 3 useful article selecting means 4 article reconstructing means 13 distributed news information 14 article dividing section 15 article selecting section 16 article reconstructing section 17 presenting section 21 article dividing / article information extracting section 22 Article information storage unit 25 Related article group generation unit 26 Topic degree calculation unit 27 Representative article selection unit

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 17/30 340 G06F 17/30 340A

Claims (16)

[Claims] 1. A news information presenting apparatus for extracting useful information from news information and presenting it to a user, comprising: news information dividing means for dividing news information into article units; News comprising: useful article selecting means for selecting useful articles; and news information reconstructing means for reconstructing and outputting the selected articles to news information in a format to be presented to a user. Information presentation device. 2. The news information presentation according to claim 1, wherein the news information dividing means determines an article portion in the news information by analyzing a format of the news information and divides the news information into article units. apparatus. 3. The article grouping means for grouping articles divided according to the similarity between each of the divided articles into a plurality of groups, the useful article selecting means comprising: An article group selecting means for selecting one or more groups from a plurality of groups in accordance with the degree of topic indicating each degree of topicality, and selecting one or more representative articles from each of the one or more groups 2. The news information presenting device according to claim 1, further comprising: a representative article selecting unit that performs the setting. 4. The news according to claim 3, wherein the article group selecting means obtains the topic level based on a total number of articles belonging to the group and a total number of information sources of articles belonging to the group. Information presentation device. 5. The news information presenting apparatus further comprises a user instruction input means for receiving an instruction from a user for presenting news information, and wherein the representative article selecting means responds to an instruction from a user. 4. The news information presenting apparatus according to claim 3, wherein one or more representative articles are selected from each group. 6. The news information presenting apparatus according to claim 5, wherein said representative article selecting means selects a representative article based on an instruction from a user regarding a distribution date of news information. 7. The news information reconstructing means may output the one or more groups to each of the one or more groups in an output order corresponding to a topic level of the one or more selected groups and / or a distribution date of the representative article. 4. The news information presentation device according to claim 3, wherein one or more representative articles are output. 8. The news information reconstructing means according to claim 1,
One or more representative articles for each of the one or more groups, plus one for each of the one or more groups.
8. The news information presenting apparatus according to claim 7, wherein one or more representative articles are further output as related articles. 9. The news information reconstructing means outputs a simplified representative article in which the content of the representative article is partially omitted, and sets a hyperlink to the original representative article in the simplified representative article. 9. The news information presentation device according to claim 7, wherein 10. The news information reconstructing means outputs, in addition to the representative article or simplified representative article, a chart in an article of a group to which the representative article or simplified representative article belongs. Item 7, 8, or 9
Described news information presentation device. 11. The news information restructuring means may output the one or more groups to each of the one or more groups in an output order corresponding to a topic level of the one or more selected groups and / or a distribution date of the representative article. 4. The news information presentation device according to claim 3, wherein the contents of one or more corresponding representative articles are converted into a format suitable for voice reading and output. 12. A news information presenting method for extracting useful information from news information and presenting it to a user, wherein the news information is divided into articles, and useful articles are selected from the divided articles. A news information presentation method, wherein the selected articles are collected, reconstructed into news information in a format to be presented to a user, and output. 13. In the selection of useful articles, the articles divided according to the similarity between the divided articles are grouped into a plurality of groups, and the topicality of each of the groups is determined. 13. The method according to claim 12, wherein one or more groups are selected from a plurality of groups corresponding to the topic level indicating the degree, and one or more representative articles are selected from each of the one or more groups. News information presentation method. 14. A storage medium used by a computer for extracting useful information from news information and presenting it to a user, wherein the news information is divided into articles, A computer readout storing a program for causing a computer to execute a step of selecting useful articles among them, and a step of collecting and collecting the selected articles into news information in a format to be presented to a user and outputting the news information; Possible portable storage medium. 15. The method according to claim 15, wherein the step of selecting useful articles comprises: grouping the divided articles into a plurality of groups according to a similarity between each of the divided articles; A step of selecting one or more groups from a plurality of groups corresponding to a topic level indicating each topicality; and a step of selecting one or more representative articles from each of the one or more groups. 15. The computer-readable portable storage medium according to claim 14, wherein: 16. A program used in a computer for extracting useful information from news information and presenting it to a user, wherein a procedure for dividing news information into articles is provided. And a program for causing a computer to execute a procedure of selecting useful articles in the above, and a procedure of collecting the selected articles, reconstructing and outputting news information in a format to be presented to a user.