Creditability evaluation device for data on webs

Abstract

To provide viewers of a terminal device with information regarding credibility of data on the Web.SOLUTION: A device includes an input device 17 that accepts classification data for classifying comments, based on a comment field 14 written in data on the Web, into a supporting group, an opposing group, and a neutral group, a data aggregation unit 21 that aggregates percentages of the supporting group, the opposing group, and the neutral group based on the classification data collected via the input device, a graph creation unit 22 that creates a graph of the percentages of the supporting group, the opposing group, and the neutral group based on an output of the data aggregation unit 21, and a graph output unit 25 that displays an image created by the graph creation unit via the Internet 60 (Claim 1). With the present configuration, based on the comments written in the data on the Web, it is possible to graphically distinguish among those who support the contents of the data on the Web, those who oppose it, and those who have opinions in between support and opposition (the neutral group), and to display the graph via the Internet 60.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a device for evaluating the credibility of data on the Web, and in particular to a device for determining and presenting the credibility of a site based on comments written on a text-based site or comments posted on a video site. [Background technology]

[0002] The web includes websites (so-called homepages), search engines, video streaming services, and various social networking sites.

[0003] However, while some of this data on the web is accurate, it also contains false information, such as so-called fake news, and it can be difficult to determine its authenticity.

[0004] Traditionally, the authenticity of data on the web was left to the discretion of the user viewing it on their device.

[0005] Patent Document 1 below discloses a technology for acquiring and evaluating speech characteristics of speech data on the Web, and improving the speaking style of the speaker. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent Publication No. 2024-55546 DISCLOSURE OF THE INVENTION [Problem to be solved by the invention]

[0007] There are cases where it is relatively easy to determine whether data on the web is accurate, such as when multiple websites deliver similar news.

[0008] However, there are cases where it is difficult to determine whether information is true or false, or where information is thought to be false but its truthfulness cannot be determined conclusively, so there are cases where information that is not accurate is spread as so-called rumors.

[0009] In such cases, there is a risk of violating privacy, denying a person's character or abilities, and distorting international sentiment. For this reason, it is important to accurately assess the authenticity of data on the Web as much as possible.

[0010] Therefore, an object of the present invention is to make it possible to provide information regarding the authenticity of data on the Web to viewers of various terminal devices. [Means for solving the problem]

[0011] In order to achieve the above-mentioned object, the credibility evaluation device for data on the Web according to the present invention comprises an input device that accepts classification data that classifies comments written on the data on the Web into those who agree, those who disagree, and those in the middle, based on the comments; a data aggregation unit that tally the percentages of those who agree, those who disagree, and those in the middle based on the classification data collected through the input device; a graph creation unit that graphs the percentages of those who agree, those who disagree, and those in the middle based on the output of the data aggregation unit; and a graph output unit that displays the image created by the graph creation unit via the Internet (Claim 1).

[0012] With this configuration, based on the comments posted on the online data, it is possible to graph the comments separately into those who agree with the content of the online data (those who agree), those who disagree (those who disagree), and those who have opinions somewhere in between (those who agree and disagree), and then display the graph via the Internet.

[0013] Viewers of the web data to be evaluated can refer to the graph and make their own judgment about the credibility of the web data.

[0014] Viewers' judgments can be based on, for example, the percentage of those who support or oppose the content, or the percentage of those in the middle. When there are many middle-income groups, it is difficult to make a judgment, so it is possible to take appropriate measures, such as refraining from making hasty judgments and waiting for new information to be provided on the web, thereby preventing the spread of rumors.

[0015] It is relatively easy to determine whether a comment posted on the web belongs to the support group or the opposition group from the displayed content of the comment. If it is difficult to determine, it can be classified as belonging to the intermediate group.

[0016] The credibility evaluation device for data on the web according to the present invention may include an input device that accepts classification data that classifies comment posters into the support group, the opposition group, the middle group, and the self-sufficient group based on comments written on the data on the web and comments by comment posters who have linked to the data on the web; a data aggregation unit that tally up the percentages of the support group, the opposition group, the middle group, and the self-sufficient group based on the classification data collected via the input device; a graph creation unit that graphs the percentages of the support group, the opposition group, the middle group, and the self-sufficient group based on the output of the data aggregation unit; and a graph output unit that displays the image created by the graph creation unit via the Internet (Claim 2).

[0017] This evaluation device according to claim 2 requires more processing time than the evaluation device according to claim 1 because it refers to the comments linked to by the comment poster, but by classifying the self-perpetuating layer (those who agree when their opinion matches and disagree when their opinion does not match), it can obtain more detailed analysis results than the evaluation device according to claim 1. [Effects of the Invention]

[0018] According to the apparatus for evaluating the credibility of data on the Web according to the present invention, it is possible to provide information about the credibility of data on the Web to viewers of various terminal devices. [Brief explanation of the drawings]

[0019] [Figure 1] FIG. 1 is a block diagram illustrating an evaluation device according to a first embodiment. [Figure 2] FIG. 10 is a block diagram illustrating an evaluation device according to a second embodiment. [Figure 3] FIG. 10 is a block diagram illustrating an evaluation device according to a third embodiment. [Figure 4] FIG. 10 is a block diagram illustrating an evaluation device according to a fourth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0020] FIG. 1 is a diagram illustrating a credibility evaluation device 10 according to the first embodiment.

[0021] The evaluation device 10 according to the first embodiment is a technology that quickly evaluates and displays the credibility of online data posted on socially important topics, such as famous entertainers, famous athletes, political / international topics, etc.

[0022] Reference numeral 11 denotes a monitor device such as a PC terminal, 12 denotes web data displayed on the monitor device 11, and 14 denotes a comment field that displays comments posted regarding the web data 12.

[0023] The monitor device 11 may be a mobile terminal. In this embodiment, for example, N comments are posted in the comment field 14. The function displays, icons, etc. displayed on the monitor device 11 are built-in functions of the terminal device itself (not shown).

[0024] 60 is the Internet (Internet line), and 17 is an input device that accepts input data in which comments (for example, N comments) displayed in the comment section 14 are classified into the support group (positive), the opposition group (negative), and the neutral group.

[0025] Reference numeral 19 denotes classification data of the support group, the opposition group, and the neutral group, input via the input device 17. If a comment is neither supportive nor opposed, the comment can be processed as the neutral group.

[0026] Specifically, it is as follows: The support group is a comment posted with the intention of supporting the article (including video) of the web data 12 to be evaluated. Any comment intended to agree with the article (including video) of the web data 12 is acceptable.

[0027] The opposition group is a comment posted with the intention of opposing / opposing the article (including video) of the web data 12 that is being evaluated. Any comment with the intention of not agreeing with the article (including video) of the web data 12 is acceptable.

[0028] The middle layer is made up of comments that either agree, disagree, or are neutral about the web data12 articles (including videos) being evaluated. To speed up the evaluation process, the middle layer can be calculated by subtracting the number of agreeing comments and the number of disagreeing comments from the total number of comments.

[0029] Reference numeral 21 denotes a data collection unit that calculates the percentages of the supporters, opponents, and neutrals based on the input data collected via the input device 17.

[0030] A graph creation unit 22 creates a graph of the percentages of the supporters, opponents, and neutrals based on the output of the data aggregation unit 21. The type of graph is not important, and examples include pie charts, band charts, bar charts, and picture charts.

[0031] A graph output unit 25 displays the image created by the graph creation unit 22 via the Internet 60 .

[0032] The graph output unit 25 transmits the graph data via the Internet 60, for example, to a terminal device (not shown) of a viewer of the Web data 12. The graph data may be transmitted to the Web data 12 or to another data processing device on the Internet 60.

[0033] Each part will be explained below.

[0034] The web data 12 is, for example, a website (a so-called homepage) and includes text data, still image data, video data, etc. More specifically, it includes news articles, advertising articles, video articles, etc.

[0035] The more the website article becomes a hot topic in society, the more comments are posted in the comment section 14. Conversely, when the website article is not a hot topic in society, the fewer comments are posted.

[0036] The evaluation device 10 according to the first embodiment preferably evaluates the credibility of articles with a certain number of comments, for example, articles with 30 or more comments. This is because if articles with a small number of comments are evaluated, the so-called cost performance (cost-effectiveness) will be poor.

[0037] The comments displayed in the comment section 14 are divided into three types of classification data (19) in order to speed up the processing. It is easy to distinguish between approval and disapproval, and if the rest are processed as an intermediate layer, the evaluation can be done quickly.

[0038] The input device 17 only needs to be able to input the type of comment. For example, since it can be handled by inputting numbers such as 1, 2, 3 (or alphabetic input, etc.), a keyboard, numeric keypad, touch panel, etc. can be used. If text is to be added, it is preferable to use a keyboard, in order to improve the speed and performance of input.

[0039] A plurality of input devices 17 may be used in order to increase the processing speed of the evaluation device 10. When a plurality of input devices 17 are used, it is desirable to perform teamwork processing appropriately, such as marking comments that have been processed as processed, and leaving comments that are difficult to process to a person in charge for classification processing, for example.

[0040] The data counting unit 21 counts the percentage of the discrimination process via the input device 17. When multiple input devices 17 are used, there may be multiple discrimination processes via the input devices 17. In this case, it is possible to perform a percentage calculation process based on each discrimination process.

[0041] The graph creation unit 22 creates an appropriate graph based on the percentage aggregation via the data aggregation unit 21. The type of graph is not important, but it is not necessary to limit the creation to one type. This is because viewers of the online data 12 are diverse and may have different preferences for (easily viewed) types of graphs.

[0042] The graph output unit 25 can transmit graph data to, for example, the terminal device of the viewer of the web data 12, but it can also forcibly send it to the viewer's terminal device and display it automatically, or it can be configured to transmit it to the viewer's terminal device in response to a request signal from the viewer of the web data 12.

[0043] The graph output unit 25 may transmit the data to an appropriate site via the Internet 60, and a viewer of the web data 12 may view the site. For example, this may be a site that compiles reviews of various types of web data, and the viewer may refer to the review summary site.

[0044] FIG. 2 is a diagram illustrating a credibility evaluation device 30 according to the second embodiment.

[0045] The evaluation device 30 according to the second embodiment is a technology for performing a more accurate credibility evaluation of web data posted on socially important topics.

[0046] The same components as those in the evaluation device 10 shown in FIG. 1 are denoted by the same reference numerals, and redundant explanations will be omitted.

[0047] The evaluation device 30 according to the second embodiment classifies the commenters into the supporters, opponents, neutrals, and self-sufficient groups based on the comments posted about the online data and the comments of the commenters who have linked to the online data.

[0048] Comment posters are distinguished based on the comments displayed in the comment field 14 and the comments in the link destination display section 38, which displays the link destination of the comment. 37 is an input device that accepts data distinguishing between the supporters, opponents, neutrals, and self-sufficient people. Multiple input devices 37 can be used.

[0049] Since the evaluation is based on two or more comments, the content of the comment posts can be distinguished with high accuracy.

[0050] Basically, we place importance on the approval or disapproval (opposition) of comments displayed in the comment section 14, but the intention behind the comment posting is often not so simple.

[0051] For example, members of the self-perpetuating group will post comments in favor of data on the web if it matches their own opinion, and in opposition if it does not match their own opinion, regardless of the content of the data.

[0052] When evaluating a comment displayed in the comment field 14, by referring to two or more linked comments, it is possible to accurately input an evaluation as to whether or not the comment is self-consistent. It is desirable that the linked comments are posted on the same type of article as the web data being evaluated. This is because it is easy to distinguish whether or not the comment is self-consistent.

[0053] When the number of comments from the self-perpetuating group is large, there is a high possibility that the article on the web data in question has had a significant social impact.

[0054] The self-perpetuating layer posts comments on the web data article, either in favor or against it. Therefore, the number of comments from the middle layer is roughly the total number of comments minus the number of favorable comments and the number of unfavorable (opposed) comments.

[0055] When there are a large number of comments from the middle class, it is often the case that the web data article being evaluated is a major social topic and it is not easy to evaluate it.

[0056] The evaluation device according to the present invention does not definitively indicate the credibility of an article, but the number of comments from the self-perpetuating layer and the number of comments from the intermediate layer serve as reference data for article readers to make their own judgments.

[0057] FIG. 3 is a diagram illustrating a credibility evaluation device 40 according to the third embodiment. The same components as those of the first and second evaluation devices (10, 30) are designated by the same reference numerals, and redundant explanations will be omitted.

[0058] The evaluation device 40 according to the third embodiment outputs graph diagram data to a viewer of data on the Web via the graph diagram output unit 25 when a request signal is received from the viewer requesting the presentation of an evaluation of the credibility.

[0059] A request signal receiving unit 42 receives a request signal from a viewer of the data on the web, requesting the presentation of an evaluation of the credibility. A graph output unit 25 outputs graph data to the terminal device (PC, mobile, etc.) of the viewer of the data on the web when an output command signal is received from the request signal receiving unit.

[0060] With this configuration, graph data is output and presented to the viewer's terminal device only when an evaluation request is made by the viewer of the web data, so the monitor display on the viewer's terminal device can be simplified and made easy to view.

[0061] FIG. 4 is a diagram illustrating an evaluation device 50 according to the fourth embodiment. The same components as those in the evaluation devices according to the first to third embodiments are denoted by the same reference numerals, and redundant explanations will be omitted.

[0062] The evaluation device 50 according to the fourth embodiment is configured to automatically classify the types of comments written in the comment section 14 (e.g., supporters, opponents, neutrals, etc.) via a type determination unit 52, without using an input device such as a keyboard (17, 37).

[0063] The type determination unit 52 is configured so that the terminal device automatically determines the meaning of the comment (for, against, neutral, etc.) according to the wording of the comment written in the comment field 14.

[0064] The type determination unit 52 determines the meaning of the comment (for, against, neutral, etc.) according to predetermined words / expressions, and sends it to the subsequent stage.

[0065] The function of the type determination unit 52 is to determine that a posted comment is in the support group if it contains words such as "I agree," "I agree (agree)," "Great," "Perfect," "I agree," "I think it's reasonable," or "I agree in principle," in other words, a written string / expression intended to indicate support.

[0066] A posted comment is judged to be from the opposition group if it contains words such as "I cannot agree (agree / disagree)," "I can only agree (agree / disagree)," "I am against," "lacks credibility," "no room for acceptance," or "rejected," in other words, characters / expressions intended to be opposed.

[0067] The number of comments in the middle tier can be calculated by subtracting the number of comments in favor and the number of comments against from the total number of comments. However, it is also possible to determine the number of characters / expressions that are neither intended to support nor oppose, such as "I can't say," "insufficient evidence," or "insufficient information," as being in the middle tier.

[0068] When determining the comment type from a character string or expression, as in the type determination unit 52, a foreign language (e.g., English) may be more suitable than Japanese. Therefore, a configuration may be adopted in which posted comments are translated into a foreign language (e.g., English) between the comment field 14 and the type determination unit 52, and the comment is then left to the subsequent type determination unit 52 to make a judgment.

[0069] According to the evaluation device 50 of the fourth embodiment, the type of the posted comment can be determined without using the input device (17, 37), which ensures an improvement in processing speed. [Explanation of symbols]

[0070] 10, 30, 40, 50 evaluation device 11 Monitor device 12 Web-based data 14 Comments 60 Internet 17, 37 Input devices 19 Segment Data 21 Data Collection Department 22 Graph Creation Department 25 Graph output section 38 Link display section 42 Request signal receiver 52 Type determination unit

Claims

1. A device for evaluating the authenticity of data on the web, comprising: Based on comments posted about the data on the web, an input device that receives input data in which the comments are classified into those in favor, those in opposition, and those in the middle; Based on the input data collected via the input device, a data collection unit that collects the percentages of the supporters, opponents, and neutrals; a graph creation unit that creates a graph of the percentages of the supporters, opponents, and neutrals based on the output of the data aggregation unit; and a graph output unit that displays the image created by the graph creation unit via the Internet.

2. A device for evaluating the authenticity of data on the web, comprising: Comments posted about the data on the web, Based on the comments of commenters who have posted links to the web data, an input device that receives classification data that classifies the posters of the comments into supporters, opponents, neutrals, and self-sufficient groups; Based on the discrimination data collected via the input device, a data collection unit that collects the percentages of the supporters, opponents, middle-income groups, and self-sufficient groups; a graph creation unit that creates a graph of the percentages of the supporters, opponents, neutrals, and self-performers based on the output of the data aggregation unit; and a graph output unit that displays the image created by the graph creation unit via the Internet.

Images