JP2008097617A - Hypertext inspection apparatus, method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically detect links to expired information. <P>SOLUTION: In a hypertext inspection apparatus, an information collecting means 11 takes the text of each document and link information included in each document from a hypertext database 21 and stores them in an information storage part 22. From the information storage part 22, a condition determining means 13 first extracts the link information containing date expressions in the descriptions of websites providing links and link information where documents with date expressions contained in their text are linked addresses. Next, the condition determining means compares the date expressions with the current date and extracts expired expressions from the documents indicated by the linked addresses in the link information extracted. Then the condition determining means calculates inappropriate scores of the link information on the basis of the result of comparing the date expressions with the current date and the result of extracting the expired expressions. The link information with the high inappropriate scores calculated is detected as information on links to expired information and is output from an output means 4. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a hypertext inspection apparatus, and more particularly to a hypertext inspection apparatus that detects an error location in a link relationship.

  In recent years, companies, organizations, and individuals have frequently released electronic information on Internet sites. Most of the information published on these sites is hypertext.

  Hypertext is a set of documents structured by hyperlinks (links), and has a structure in which documents are nodes and links are made between documents. A representative example of hypertext is WWW (World Wide Web). The WWW is a set of hypertext described in the HTML (Hyper Text Markup Language) format as in the document 101 of FIG. 2 (a), and links and anchor character strings are marked with <A> tags. In the document 101 of FIG. 2A, the identification information of the documents 102, 103, and 104 is specified in the href attribute of the <A> tag. Document identification information is usually called a URL or a web address on the WWW, but is simply called an address in the present invention. In addition, character strings such as “GX0011”, “GX0012”, and “GX0013” sandwiched between <A> tags are generally called anchor character strings. However, since an image file may be sandwiched between <A> tags, in the present invention, a character string or an image sandwiched between <A> tags is called a link source notation and is handled in the same manner.

  In addition to the href attribute, the <A> tag attribute described in the document 101 includes a target attribute and a style attribute. The target attribute is an attribute for designating a window for displaying the linked document. The style attribute is an attribute for designating the font size, color, emphasis expression, etc. when displaying the link source notation of the link. When the document 101 in FIG. 2 (a) is viewed with a browser, it is displayed as the document 101 in FIG. 2 (b). In the document 101 of FIG. 2B, the document 102 can be accessed via the link 201 by clicking the link source notation “GX0011”. Similarly, when the link source notation “GX0012” is clicked, the document 103 can be accessed via the link 202, and when the link source notation “GX0013” is clicked, the document 104 can be accessed via the link 203.

  Although the WWW has been described as a representative example of hypertext, the present invention is not limited to the WWW. Hypertext can be described using not only HTML but also XML (Extensible Markup Language), SGML (Standard Generalized Markup Language), and the like.

  Further, in the present invention, in order to avoid the confusion of the term “user”, a person who visits a company / organization / individual site and browses hypertext is “visitor”, and hypertext is managed using the present invention. Let's call a person an "administrator".

  Hypertext management is becoming more complex as the amount of information published to the Internet increases, and the number of link inconsistencies such as links with inappropriate link source notation and links with incorrect link destinations is increasing. . There are two types of link inconsistencies: physical inconsistencies and logical inconsistencies.

  A physical inconsistency is an inconsistency in which the link destination is physically inaccessible, such as there is no link destination text or the link destination server is down. With these physical inconsistencies, the server or client returns an error when accessing the document.

  A logical inconsistency is an inconsistency that causes a logical error even though it is physically accessible, such as a link to wrong product information or an expired campaign. In these logical inconsistencies, text exists at the link destination, and there is no abnormality in the linked server, so that no error occurs when the document is accessed. However, the impact is as great as physical inconsistencies, such as confusing visitors due to incorrect links, or the occurrence of visitors who apply for expired campaigns and the management struggling to respond. Examples of logical inconsistencies include: (1) Link misconfiguration, (2) Link to expired information, (3) Link source notation, (4) Link source notation style, (5 ) Ghost link, (6) Loop link, etc. Hereinafter, each example of logical inconsistency will be described in detail with reference to the drawings.

(1) Link tension error As shown in Fig. 3, the link tension error is an inconsistency when the content expected in the link source notation and the text content of the link destination are misaligned. In FIG. 3, the links 211, 212, 213, and 214 have the same link source notation “GX0011”. The links 211, 212, and 213 of the documents 111, 112, and 113 are all the document 116. Therefore, a visitor who has viewed the documents 111, 112, and 113 can access the document 116, which is introduction information of the GX0011, as expected by the link source notation “GX0011”. However, the link destination of the link 214 incorrectly specifies the document 117 that is the product introduction of GX0012. For this reason, a visitor who has viewed the document 114 is confused because the product introduction different from the information expected by the link source notation “GX0011” is shown.

  The links 211, 212, 213, and 215 are all linked to the document 116. However, the link source notation of the link 215 is erroneously described as “GX0012”. Therefore, a visitor who has viewed the document 115 is presented with a product introduction that is different from the information expected in the link source notation “GX0012”, which is confusing.

  The two links 215 and 216 extending from the document 115 have the link source notation “GX0012”. However, since the respective link destinations are different from those of the documents 116 and 117, a visitor who has viewed the document 115 follows a different document regardless of the same link source notation, and is confused.

  In addition, although the link error to the product information was explained here as an example of the link error, other link errors between the English version document and the Japanese version document, or to pages that are not related at all Mismatches such as wrong links are also included in link mistakes.

(2) Link to expired information The link to expired information is inconsistent when there is a link to an expired campaign or closed service, as shown in FIG. In FIG. 4 (a), the campaign is conducted for a limited time from July 20, 2002 to August 31, 2002 in Document 125. In addition, links 221, 222, 223, and 224 are extended from the documents 121, 122, 123, and 124 to the document 125, which is a campaign page, with the same link source notation “free membership”. On the other hand, in FIG. 4B, since the campaign period is ended, the document 125 notifies the end. In the documents 121, 122, and 123, the link to the document 125 that is the campaign page is deleted. Nevertheless, in the document 124, since the link is forgotten to be deleted, the link 224 to the document 125 remains with the link source notation “free membership”. Therefore, even if a visitor who has viewed the document 124 expects “free membership” in the link source notation, the service cannot be received.

  In addition, although the link to the expiration campaign was explained here as an example of the link to the expiration information, the address of the document that was linked first is transferred, and another document with another content is transferred to the original address. Inconsistencies caused by being placed are also included in the link to the expiration information. Even if the period is not limited from the beginning, the link to the expiration information includes inconsistencies caused by termination of the link destination service for some reason or closing of the site. However, if the document has expired and is deleted, an error occurs during access, which is included in the physical inconsistency. Moreover, although an expired link may be considered as a kind of erroneous link, in the present invention, links that have expired in particular among the erroneous links are distinguished as expired links.

(3) Link source notation inconsistency Link source notation inconsistency is inconsistency when the link source notation is not uniform and there is fluctuation, as shown in FIG. In FIG. 5, links from the documents 131, 132, 133, and 134 to the document 135 are provided. The link source notation of the links 231, 232, and 233 are all “GX Series”. However, the link source notation of the link 234 is “gX Series”. For this reason, a visitor who has viewed the document 134 follows the link 234 by misunderstanding that there is a “gX Series” that is different from the “GX Series” in the link source notation.

  In addition, here, we explained the fluctuation of the capital letter and small letter of the link source notation as an example of inconsistency of the link source notation, but in addition, the fluctuation of the English / Katakana notation such as `` Top page '' `` Top page '', `` Fluctuations in katakana notation such as violin and violin, fluctuations in katakana and hiragana notation such as watermelon and watermelon, fluctuations in notation with ambiguous similar expressions such as event information and seminar information, "Series" Spelling mistakes such as “Selies” are also included in inconsistencies in the link source.

(4) Link source notation style inconsistency Link source notation style inconsistency, as shown in Fig. 6, because the link style attribute and target attribute are different, the link looks different, or the link It is inconsistency when the effect when clicking is different. In FIG. 6 (a), four links are specified in the document 141, and in three of them, the target attribute "_blank" is specified so that when the link is clicked, the link destination page is displayed in the pop-up window. . Therefore, as shown in FIG. 6B, a visitor browsing the document 141 with a browser can browse the linked documents of the links 241, 242, and 243 in a pop-up window one after another while the document 141 remains open. In particular, it is often convenient to display the link destination page in this pop-up window in a document such as a collection of links that often browses the link destination and returns to browse another link destination. On the other hand, since the target attribute is not specified for the link 244, the document is switched when the link is clicked. For this reason, when the visitor clicks on the link 244, the document is switched, and it becomes necessary to search for a link for returning to the document 141 or to use the back button of the browser.

  Note that here, the target attribute is inconsistent as an example of inconsistency in the style of the link source notation. However, because the style attribute is inconsistent, the color of some links may be different. Inconsistencies such as the presence / absence of emphasis on links in parts are also included in the inconsistency in the style of link source notation.

(5) Ghost link As shown in FIG. 7, the ghost link is an inconsistency when the link is specified in the HTML notation of the document, but the existence of the link is not noticed when viewed with a browser. In FIG. 7 (a), there is an <A> tag between the character string “GX series inventory status” representing the headline and the <TABLE> tag representing the table, and HIDDEN_URL is specified as the link destination. However, there is a link between the heading and the table as shown in Fig. 7 (b) when browsing with a browser because nothing is inserted between the <A> tag as a link source notation. I don't know. Such links are easy to follow on a crawler, but are difficult for administrators to check. If HIDDEN_URL points to a confidential file such as a customer list, confidential information can be easily acquired by a crawler, but there is a possibility that a human may not notice the leakage.

  In this example, the link source notation is not specified as an example of a ghost link. However, a transparent image is specified in the link source notation, or a very small image or character is specified. Inconsistencies in cases where visual confirmation in the browser is difficult, such as when images and characters of the same color as the background are specified, are also included in the ghost link. Even if it can be seen, if the link style of the link source notation is the same color as the main text and there is no emphasis expression, and the link and the main text cannot be distinguished, it is difficult to visually check with the browser. Include in the link.

(6) Loop link As shown in FIG. 8, the loop link is an inconsistency when returning to the original page when a certain information is requested and the link is followed. In FIG. 8, a link 261 is set from a document 161 to a document 162 with a link source notation “notice of present”. Further, a link 262 is provided from the document 162 to the document 163 with the link source notation “digital camera present”. Further, a link 263 is set from the document 163 to the document 161 with the link source notation “present here”. For example, it is assumed that a visitor who has viewed the document 161 follows the link 261 with an interest in the “presentation notification”. Then, since there is a link 262 whose link source notation is “digital camera present” in the document 162, the document 163 is accessed with the expectation that there is detailed information on the present beyond that. Further, the document 163 also follows the link 263 so as to obtain the desired information because there is a link 263 whose link source notation is “here is the present”. However, the end of the link 263 returns to the document 161, and eventually it is impossible to know where to go. Thus, if there is a loop link, the visitor will be lost between documents without obtaining the information he wants.

  As a first conventional technique for inspecting hypertext, there is a link checker for hypertext on the Internet, which is introduced in Non-Patent Document 1 described later. This is a tool that automatically circulates hypertext on the Internet and records a log when an error occurs. There are two types of link checkers: an on-line diagnosis by designating an inspection target address and an off-line diagnosis by designating a specific folder on a hard disk.

  The second prior art includes an invention disclosed in Patent Document 1 described later. According to this method, a hypertext address to be managed is stored in a database, and a browser is automatically connected to the address periodically. As a result, it is possible to check whether a document exists at the stored hypertext address and to detect physical inconsistencies such as dead links. Further, according to the present invention, keywords and images that serve as clues for specifying each document in the database are registered in the system in advance. As a result, when a dead link is detected, a lost page can be searched using a search engine and correction candidates can be presented.

  Also, as a conventional technique for general document inspection, there is a document proofing system such as Microsoft's Word autocorrect function. These document proofing systems can detect inappropriate expressions such as mistakes in sending kana and repetition of the particle “no” and output correction candidates.

JP 2001-273185 A Linkcheck "LinkScan" manufactured by Elsop, USA, [October 9, 2002 search], Internet <URL: http: /www.elsop.com/linkscan/>

  The first problem is that only the physical inconsistency can be detected by the first and second prior arts, and the logical inconsistency cannot be detected. The reason is that in the first and second prior arts, when connecting to a hypertext address, whether or not there is a mismatch is determined only by whether or not an error is returned from the server. At present, the only way to detect logical inconsistency that does not cause an error on the server is to rely on manual confirmation on the browser.

  The second problem is that the first and second prior arts can only present correction candidates for physical inconsistencies and cannot present correction candidates for logical inconsistencies. The reason is the same as the first problem.

  The third problem is that it is too expensive to manually check on the browser. The reason is that a large-scale site such as a company is composed of thousands to tens of thousands of hypertexts, and there are tens of thousands to hundreds of thousands of links between documents. It is not practical to check these links one by one in terms of time and cost. In addition, in the visual confirmation on the browser, checking of ghost links and the like is likely to leak.

  A first object of the present invention is to enable detection of logical inconsistencies, particularly links to expired information.

  A second object of the present invention is to enable a manager to present correction candidates for links to expiration information.

  The third object of the present invention is to greatly reduce the cost of inconsistency checking.

  A hypertext inspection apparatus according to the present invention includes a hypertext database that stores a set of documents structured by hyperlinks, takes out each document from the hypertext database, and sets link source addresses and link destinations as link information included in each document. Information collecting means for retrieving the address and link source notation, information storage means for storing the text and link information of the document retrieved by the information collecting means, link information including date representation in the link source notation, and text A link information having a document containing a date expression as a link destination address is extracted from the information storage means, the date expression is compared with the current date and time, and the document indicated by the link destination address of the extracted link information The expired expression is extracted from the result, and the comparison result between the date expression and the current date and time And calculating the inappropriate score of the link information based on the extraction result of the expired expression, detecting the calculated link information having a high inappropriate score as link information to the expired information, and outputting the detection result from the output means. Output condition determining means.

  According to the present invention, a link to expiration information can be automatically detected.

"First embodiment"
Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 1, the first embodiment of the present invention includes a data processing device 1 that operates under program control, a storage device 2 that stores information, an input device 3 such as a keyboard, a display device, and a printing device. And the like.

  The data processing device 1 includes information collection means 11, candidate calculation means 12, condition determination means 13, and correction reflection means 14.

  The information collecting means 11 takes out each document from the hypertext database 21 stored in the storage device 2, takes out the link information, and stores it in the information storage unit 22. Here, the link information includes a link source address, a link destination address, a link source notation, a target attribute, a style attribute, and the like. In addition to the link information, the information storage unit 22 may record the text of the document, the update date, the acquisition date and time, and the status at the time of acquisition (such as error or success).

  The condition determination means 13 groups the links stored in the information storage unit 22 for each item of link information, and extracts a specific link out of the group from the information storage unit 22 as a link mismatch.

  Candidate calculation means 12 calculates and outputs correction candidates for the link mismatched link extracted by condition determination means 13. Here, the correction candidate specifies which item and how to correct the link information of the link causing the inconsistency.

  The correction reflecting means 14 reflects the result confirmed by the administrator on the output link mismatch and the correction candidate in the hypertext database 21.

  The storage device 2 includes a hypertext database 21 and an information storage unit 22.

  The hypertext database 21 stores a set of hypertexts existing in the site to be inspected. Note that the hypertext database 21 does not necessarily have to exist in the local storage device 2, and may be distributed via a network like a hypertext group on the Internet.

  The information storage unit 22 stores link information included in each document in the hypertext database 21. For example, the link information included in the document 101 in FIG. 2 is as shown in FIG. As can be seen from FIG. 9, the link 201 is linked from the document 101 to the document 102 with the link source notation “GX0011”, the target attribute is designated as _blank, and the style attribute is designated as st01. Although the case where the link source notation is text has been described here, when an image is specified in the link source notation, the address of the specified image file is recorded in the link source notation. Alternatively, the image file may be subjected to a character recognition module to extract text described in the image and register it in the same manner as the character string.

  Next, the operation of the present embodiment will be described in detail with reference to FIG. 1 and FIGS.

  First, based on the setting of the collection condition input from the input unit 3, the information collection unit 11 reads a document stored in the hypertext database 21 (step S1 in FIG. 10). Here, when the hypertext database 21 is WWW, the document can be accessed via HTTP (Hyper Text Transfer Protocol). Such a function is conventionally realized in a web browser such as IE (Internet Explorer) or a web crawler (spider / robot). FIG. 11 shows a collection setting screen when the hypertext database 21 is WWW. In Figure 11, the domain name of the site to be analyzed, the target number of pages to be collected, the extension of the document to be collected, the time interval for accessing the server, the number of retries if collection fails, the timeout during collection You can specify the time, the depth of the recursive hierarchy when collecting recursively by following links. When the execution button in FIG. 11 is pressed, collection of hypertext is started.

  Next, the information collecting means 11 analyzes the HTML description of the collected document, extracts link information as shown in FIG. 9, and stores it in the information storage unit 22 (step S2 in FIG. 10).

  Next, based on the extraction condition input from the input unit 3, the condition determination unit 13 extracts the corresponding link from the information storage unit 22 as a link mismatch (step S3 in FIG. 10). An extraction condition setting screen is shown in FIG. In FIG. 12, it is possible to specify which of various types of link inconsistencies such as dead links (physical inconsistencies), wrong links, links to expired information, etc. are extracted for the site to be analyzed. If it is known in advance that a link to a specific address is inconsistent, a link having that address as a link destination can also be extracted (“specific URL” in FIG. 12). Furthermore, when there are many extracted link inconsistencies, it is possible to specify how many cases are displayed on the screen. When the execution button in FIG. 12 is pressed, extraction of link mismatch is started. Of each link mismatch, the dead link can be extracted by the above-described prior art, and since it is not directly related to the present invention, description thereof is omitted. Moreover, since the method of extracting a link having a specific URL as a link destination is obvious to those skilled in the art, the description thereof is omitted. Details of the method for extracting the remaining various logical inconsistent links will be described later.

  Next, for the link extracted by the condition determining means 13 as a link mismatch, the candidate calculating means 12 obtains correction candidates for eliminating the mismatch, and outputs a result list screen (steps S4 and S5 in FIG. 10).

  An example of an output link mismatch result list screen is shown in FIG. In FIG. 13, links having the same link destination and link source notation are grouped and displayed with a mismatch type and a correction candidate, respectively. The corresponding document can be accessed by clicking on each link source address and link destination address. In the correction candidate column, correction candidates output by the system are entered. The correction candidates describe the items of link information to be corrected and how to correct them by separating them with “:”. For example, “link: delete” in FIG. 13 means that the link itself is deleted. Also, “link source notation: new arrival information” means that the link source notation is changed to “new arrival information”. This correction candidate can be rewritten after confirmation by the administrator.

  Next, the administrator confirms the output link mismatch and the correction candidate (step S6 in FIG. 10). At this time, in FIG. 13, links having the same link destination and link source notation are grouped. Therefore, the administrator does not need to check all the links, but only checks the representative examples of each inconsistency. For example, the links 271 to 274 are all expired because the link destination is the document 175, the link source notation is “XX campaign in progress”, and it is necessary to delete the link to correct it I know that there is. Therefore, the administrator need not check the remaining links 272 to 274 if he / she can access the document 171 and confirm that the inconsistency and correction candidates of the link 271 are correct without checking all the links 271 to 274. In addition, the cost required for confirmation is reduced.

  When there are a plurality of correction candidates, they are presented to the administrator separated by OR, such as “link destination: document 177 OR link source notation: product B”. In this case, the administrator need only leave the correction candidates necessary as a result of the confirmation. If it is determined that the correction candidate is wrong as a result of the confirmation, it can be corrected. For example, the correction candidates for links 278 and 279 are designed to correct the link source notation to “new arrival information”. However, if it is appropriate to change the link destination address to document 180, it is applicable. The correction candidate may be changed to “link destination: document 180”. If the administrator determines that correction is not desired as a result of confirmation, if the correction candidate is left blank, correction is not performed in a later step.

  Next, when the administrator presses the correction reflection button in FIG. 13, the correction reflection means 15 corrects each document in the hypertext database 21 based on the correction candidates confirmed by the administrator (step S7 in FIG. 10). . At this stage, if correction candidates are connected by multiple ORs, only the first correction candidate is reflected.

  In FIG. 13, there is a link “sort” in the items of link source, link destination, and link source notation. This is for sorting the extraction results using each item as a key. For example, when “sort” in the link source item is clicked, the extraction results are sorted and output using the link source document as a key. As a result, it is possible to grasp what kind of link inconsistency has occurred in each document, which can be used when correcting the inconsistency manually. When “sort” in the link destination item is clicked, the extraction results are sorted and output using the link destination document as a key. As a result, it is possible to grasp the occurrence status of inconsistency with respect to a link given to a specific document, so it is possible to focus on inconsistencies with respect to important documents such as documents with concentrated access. Furthermore, when “sort” in the item of link source notation is clicked, the extraction results are sorted and output using the link source notation as a key. As a result, it is possible to grasp what type of link source notation is likely to cause inconsistency, so that the validity of the expression used as the link source notation can be checked.

  In the present embodiment, the administrator corrects the link source notation, link destination, etc. displayed in the column “correction candidates” on the result list screen of FIG. 13, but the “link source”, “ The link source, the link destination, and the link source notation may be corrected by directly overwriting the fields of “link destination” and “link source notation”. In the present embodiment, the hypertext collection setting and the link mismatch extraction condition setting are performed on separate screens. However, when the analysis is started, the conditions are set all at once on the same screen. There is also a method in which all of S1 to S5 are automated and executed, and is not limited to the method described in the present embodiment.

  Also, in this embodiment, the link inconsistency and correction candidates that were output by the administrator in step S6 were confirmed, but there is also a method in which step S6 is omitted and all steps S1 to S7 are automated and executed. There is no limitation to the method described in this embodiment.

  In the present embodiment, the case where the administrator decides and executes the inspection timing has been described. However, collection conditions and extraction conditions are set in advance, and steps S1 to S5 are automatically executed periodically. There is also a method of notifying the obtained result by e-mail or the like, and is not limited to the method described in this embodiment.

[Effect link detection embodiment]
Next, with reference to FIG. 3, FIG. 14, and FIG. 15, the operations of the condition determination means 13 and the candidate calculation means 12 when an erroneous link is detected will be described in detail. Assume that the information storage unit 22 stores link information of the document group of FIG.

  First, the condition determination means 13 groups from the information storage unit 22 links having the same link source notation, sublinks links having the same link destination in the same group, and extracts links belonging to subgroups having different link destinations. Further, an inappropriate score is assigned to each link according to the number of links included in the subgroup (step T11). FIG. 15 (a) shows an example of the link extracted at step T11 and the inappropriate score given. As shown in FIG. 15 (a), links 211, 212, 213, and 214 are grouped under link source notation “GX0011”, and links 215 and 216 are grouped under link source notation “GX0012”. Recognize. Further, in the group of the link source notation “GX0011”, three links 211, 212, and 213 have a link destination as a subgroup of the document 116, and a link 214 has a link destination as a subgroup of the document 117. In the group of the link source notation “GX0012”, the link 215 is a subgroup of the document 116 as the link destination, and the link 216 is a subgroup of the document 117 as the link destination.

  Inappropriate score assignment, first, the inappropriate score of one group is set to 1, and the distribution of the inappropriate score in inverse proportion to the number of links in the sub group is used as the inappropriate score of each sub group. Further, an inappropriate score of each link is obtained by equally dividing the inappropriate score of each subgroup by the number of links in the subgroup. For example, in FIG. 15 (a), if the inappropriate score of the group with the link source notation “GX0011” is 1, and the distribution is inversely proportional to the number of links in the subgroup, the link destination address is inappropriate for the subgroup of the document 116. The score is 1/4, and the inappropriate score of the subgroup whose link destination address is the document 117 is 3/4. Further, since the inappropriate score 1/4 of the subgroup is divided into three equally at the links 211, 212, and 213, the inappropriate score of each link is 1/12. The inappropriate score of the link 214 is 3/4. Similarly, the inappropriate scores of the links 215 and 216 are both 1/2.

  Next, the condition judging means 13 groups the links with the same link destination from the information storage unit 22, subgroups the links with the same link source notation in the same group, and extracts the links belonging to the subgroups with different link source notations. To do. Further, an inappropriate score is given to each link according to the number of links included in the subgroup (step T12). FIG. 15 (b) shows an example of the link extracted in step T12 and the inappropriate score given. Referring to FIG. 15B, it can be seen that the links 211, 212, 213, and 215 with the link destination of the document 116 are grouped, and the links 214 and 216 with the link destination of the document 117 are grouped. Further, among the groups of the document 116 linked to, three links 211, 212, and 213 are subgroups with the link source notation “GX0011”, and the link 215 is a subgroup with the link source notation “GX0012”. . Among the groups of the document 117 as the link destination, the link 214 is a subgroup with the link source notation “GX0011”, and the link 216 is a subgroup with the link source notation “GX0012”. Assignment of an inappropriate score is the same as in step T11. Therefore, the inappropriate score of the links 211, 212, and 213 at step T12 is 1/12, the inappropriate score of the link 215 is 3/4, and the inappropriate score of the links 214 and 216 is 1/2.

  Next, the condition determining means 13 groups from the information storage unit 22 links having the same link source and the same link source notation, sub-groups links having the same link destination in the same group, and sub-groups having different link destinations. Extract links belonging to. Further, an inappropriate score is assigned to each link according to the number of links included in the subgroup (step T13). FIG. 15 (c) shows an example of the link extracted in step T13 and the inappropriate score given. FIG. 15C shows that links 215 and 216 whose link source is the document 115 and whose link source notation is “GX0012” are grouped. Further, the link 215 has a link destination as a subgroup of the document 116, and the link 216 has a link destination as a subgroup of the document 117. Assignment of an inappropriate score is the same as in step T11. Accordingly, the inappropriate scores of the links 215 and 216 at step T13 are each ½.

  Next, the condition determination means 13 extracts from the information storage unit 22 a link in which the word included in the link source notation is not included in the title, heading, or highlighted character string of the link destination document, and gives an inappropriate score 1 ( Step T14). FIG. 15 (d) shows an example of the link extracted at step T14 and the inappropriate score given. In the links 214 and 215 extracted in FIG. 15D, the word included in the link source notation does not appear in the linked document in FIG.

  Next, the condition determination means 13 sums up the inappropriate scores for the links (step T15). Therefore, the inappropriate scores of the links 211, 212, and 213 are 1/12 + 1/12 = 1/6. Further, the inappropriate score of the link 214 is 3/4 + 1/2 + 1 = 9/4. The inappropriate score of the link 215 is 1/2 + 3/4 + 1/2 + 1 = 11/4. Further, the inappropriate score of the link 216 is 1/2 + 1/2 + 1/2 = 3/2.

  Next, the condition determination means 13 compares the total inappropriate scores between the subgroups, and extracts a link with a high inappropriate score as a link mismatch. Further, the candidate calculation means 12 obtains a correction candidate for matching the link information of the link having a high score with the link information of the link having a low score in the same group for the link extracted under each condition (step T16). ). In the group with link source notation “GX0011” in FIG. 15 (a), the sum of the inappropriate scores of the subgroups consisting of links 211, 212, and 213 is 1/6 + 1/6 + 1/6 = 1 / 2. Since the sum of the inappropriate scores of the subgroups composed of the links 214 is 9/4, the link 214 having a high inappropriate score is determined as a link mismatch. Further, it is understood that the correction candidate “link destination: document 116” is appropriate for matching the link information of the link 214 with the subgroups of the links 211, 212, and 213. Further, in the group of FIG. 15 (a) where the link source notation is “GX0012”, the total inappropriate score of link 215 is 11/4, and the total inappropriate score of link 216 is 3/2. The link 215 is determined as a link mismatch. Further, it is understood that the correction candidate “link destination: document 117” is appropriate for matching the link information of link 215 with the subgroup of link 216. Similarly, in FIG. 15 (b), link 215 is determined to be link inconsistent and “link source notation: GX0011” becomes a correction candidate, and link 214 is determined to be link inconsistent and “link source notation: GX0012” is It is obtained as a correction candidate. Similarly, in FIG. 15C, the link 215 is determined to be link inconsistent, and “link destination: document 117” becomes a correction candidate. When the above results are integrated, the link mismatch is links 214 and 215, and the correction candidates are “link destination: document 116 OR link source notation: GX0012” and “link destination: document 117 OR link source notation: GX0011”, respectively. Become.

  In this embodiment, a link having a high improper score is regarded as a link mismatch. However, a threshold of an inappropriate score is provided, and even if the inappropriate score is high, a link mismatch does not occur as long as it is below the threshold. There is also a method, and the present invention is not limited to the method described in this embodiment.

  Further, in the present embodiment, the calculation is based on the number of links in the subgroup as an example of the inappropriate score, but the number of extracted times may be simply used as the inappropriate score, and the method described in the present embodiment It is not limited to. In addition, there is a method in which the number of links in a subgroup is used as the feature vector of the link, and the average value of the distance from the feature vector of the inconsistent link given in advance as teacher data is used as an inappropriate score. It is not limited to the method described in the embodiment.

  Further, in the present embodiment, as an extraction condition for erroneous links, (1) a first inappropriate score calculated by comparing link source notation for a plurality of links with the same link destination page, (2) link source Second inappropriate score calculated by comparing linked pages for multiple links with the same notation, (3) Calculated by comparing linked pages for multiple links with the same link source notation The third inappropriate score to be calculated, and (4) the 4th inappropriate score calculated by comparing the contents between the link source notation and the linked page, was summed up. Inappropriate scores may be calculated using a plurality of types or weighting according to each condition, and is not limited to the method described in this embodiment.

[Expired link detection embodiment]
Next, with reference to FIG. 4 and FIG. 16, the operation of the condition determination means 13 and the candidate calculation means 12 when detecting an expired link will be described in detail.

  First, the condition judgment means 13 extracts a link whose date expression includes a date expression or a link pointing to a document including a date expression, calculates the expiration date from the date expression, and the current date is within the expiration date It is determined whether or not there is (step T21 in FIG. 16).

  Next, the condition determination means 13 extracts an expired expression included in the link destination document of the extracted link (step T22 in FIG. 16). Here, the term "expired" means "closed", "moved", "finished", "jumps automatically in seconds", "Has a month x day", "Thank you for your patronage", "Thank you for participating" This is an expression often used in announcements when the service is terminated, closed, or moved. In addition to the above expression, if the setting is such that the document is automatically switched after a few seconds by HTML, this is also extracted as an expired expression.

  Next, the condition determining means 13 integrates the determination result of whether or not it is within the expiration date of step T21 and the number of expired expressions extracted at step T22, and calculates an inappropriate score for the link. If this inappropriate score is less than or equal to a predetermined threshold value, a link mismatch is output (step T23 in FIG. 16). As an example of the calculation method of the inappropriate score of the link, there is a method of obtaining by multiplying the number of days from the expiration date by the number of appearances of the extracted expiration expression. In addition, the average of the distance between the result of whether it is within the expiration date and the number of occurrences of the extracted expired expression as the feature vector of the link and the feature vector of the inconsistent link previously given as teacher data There is a method of setting a value as an inappropriate score, and the method is not limited to the method described in this embodiment.

  Next, the candidate calculation means 12 extracts a transfer destination address from the link destination document for the link output as the link mismatch and sets it as a correction candidate. Here, the transfer destination address is a jump destination address when the document is automatically switched by HTML. Even if the document does not change automatically, expressions such as “Click here” and “Moved to the following URL” are extracted, and the link destination address of the link described in or near the expression is transferred. A correction candidate may be used as the destination address. On the other hand, if the destination address cannot be extracted, the correction candidate is output as “link: deleted”.

  A specific example of the operation of the condition determination means 13 and the candidate calculation means 12 will be described in the case of FIG. Note that, as described above, as a method of calculating the inappropriate score of the link, a method of multiplying the number of days from the expiration date by the number of appearances of the extracted expiration expression is used.

  First, in step T21, since there is a date expression “July 20, 2002 to August 31, 2002” in the document 125, the condition determination means 13 extracts the links 221, 222, 223, and 224. At this time, since the current date and time is August 15, 2002, it is determined that the links 221, 222, 223, and 224 are within the expiration date.

  Next, in step T22, since no expired expression appears in the document 125, nothing is extracted.

  Next, in Step T23, the inappropriate expression score of the links 221, 222, 223, and 224 is determined based on the result that the date expression was within the expiration date in Step T21 and that no expired expression was extracted in Step T22. Since the number of days from the expiration date and the number of appearances of the extracted expiration expression are both 0, 0 × 0 = 0, and it is determined that any link is appropriate.

  On the other hand, in the case of FIG. 4B, a specific example of the operation of the condition determination means 13 and the candidate calculation means 12 will be described.

  First, in step T21, since there is a date expression “July 20, 2002 to August 31, 2002” in the document 125, the condition determination means 13 extracts the link 224. At this time, since the current date and time is September 15, 2002, it is determined that the link 224 has exceeded the expiration date in the document T125.

  Next, in step T22, the condition determination means 13 extracts an expired expression “finished” from the document 125.

  Next, in Step T23, the result of the date expression being out of the valid period in Step T21 and the expired expression of “finished” in Step T22 were extracted, so the inappropriate score of link 224 is Since the number of days from 15 is 1, and the number of occurrences of the extracted expired expression is 1, 15 × 1 = 15. Here, for example, if the threshold is 10, the link 224 is determined to be a link mismatch.

  Next, in step T24, the candidate calculation means 12 attempts to extract the transfer destination address, but since the corresponding address is not described in the document 125, “link: delete” is output as a correction candidate for the link 224. .

  In this embodiment, the detection of the expired link when focusing on the date expression and the expired expression has been described. However, as with the erroneous link detection, the links with the same link destination page are grouped, and the links within the same group are linked. There are a method for detecting subgroups with different source notations, a method for grouping links with the same link source notation, and detecting subgroups with different link destinations within the same group. It is not limited.

[Embodiment for detecting inconsistency of link source notation]
Next, with reference to FIG. 5, FIG. 17, and FIG. 18, the operations of the condition determination means 13 and the candidate calculation means 12 when detecting inconsistencies in the link source notation will be described in detail.

  First, the condition determination unit 13 groups links with the same link destination from the information storage unit 22, subgroups links with the same link source notation in the same group, and extracts links belonging to sub groups with different link source notations. . Further, an inappropriate score is given to each link according to the number of links included in the subgroup (step T31 in FIG. 17). When the document group is shown in FIG. 5, the link extracted in step T31 and the inappropriate score to be given are as shown in FIG. As can be seen from FIG. 18, the links 231, 232, 233, and 234 of which the link destination is the document 135 are grouped. Further, three links 231, 232, and 233 are subgroups whose link source notation is “GX Series”, and links 234 are subgroups whose link source notation is “gX Series”.

  Inappropriate score assignment, first, the inappropriate score of one group is set to 1, and the distribution of the inappropriate score in inverse proportion to the number of links in the sub group is used as the inappropriate score of each sub group. Further, an inappropriate score of each link is obtained by equally dividing the inappropriate score of each subgroup by the number of links in the subgroup. Accordingly, the inappropriate score of the links 231, 232, and 233 at step T 31 is 1/12, and the inappropriate score of the link 234 is 3/4. Here, the condition determination means 13 compares the sum of the inappropriate scores between the subgroups, and extracts a link having a high inappropriate score as a link mismatch. In FIG. 18, since the inappropriate score 3/4 of the link 134 is higher than the total 1/4 of the inappropriate scores of the links 231, 232, and 233, the link 134 is extracted as a link mismatch.

  Next, the candidate calculation means 12 checks whether or not the link source notation of the extracted link is registered in the term dictionary (step T32 in FIG. 17). Here, the term dictionary is a table having expressions that should be unified as values by using notation fluctuation as a key. For example, “free software” means software that can be used free of charge, and there is a fluctuation of notation such as “freeware” and “free software”. It is only necessary to register “free software” and “free software” as values in the term dictionary. If the link source notation of the extracted link is registered in the term dictionary, the candidate calculation means 12 outputs the expression to be unified corresponding to the key as a correction candidate (step T33 in FIG. 17). It should be noted that a fuzzy search may be used when searching for keys in order to sufficiently absorb notation fluctuations. In addition, an expression to be unified is searched fuzzyly without using a word of fluctuation of the notation, and if the similarity of a character string is equal to or greater than a threshold value, the retrieved expression to be unified may be set as a correction candidate.

  In the case of FIG. 18, it is assumed that neither “GX Series” nor “gX Series” is registered in the term dictionary.

  On the other hand, if the link source notation of the extracted link is not registered in the term dictionary, the candidate calculation means 12 displays the link source notation of the link with a large inappropriate score within the same group, and the link source of the link with a low score. Correction candidates that match the notation are obtained (step T34 in FIG. 17). In the case of FIG. 18, “link source notation: GX Series” is output as a correction candidate.

  In this embodiment, the calculation is based on the number of links in the subgroup as an example of the inappropriate score. However, the number of links in the subgroup is used as a feature vector of the link, and is given in advance as teacher data. There is a method of setting an average value of the distances to the feature vectors of inconsistent links as an inappropriate score, and the method is not limited to the method described in the present embodiment.

[Embodiment for detecting inconsistency in style of link source notation]
Next, with reference to FIG. 6, FIG. 19, and FIG. 20, the operation of the condition determination unit 13 and the candidate calculation unit 12 when detecting inconsistency in the style of the link source notation will be described in detail.

  First, the condition determination means 13 groups from the information storage unit 22 links with the same link source document, subgroups links with the same target attribute within the same group, and extracts links belonging to subgroups with different target attributes. . Further, an inappropriate score is assigned to each link according to the number of links included in the subgroup (step T41 in FIG. 19). When the document group is shown in FIG. 6, the link extracted in step T41 and the inappropriate score to be given are as shown in FIG. As can be seen from FIG. 20, the links 241, 242, 243, and 244 whose source is the document 141 are grouped. Further, three links 241, 242, and 243 are subgroups with a target attribute “_blank”, and link 244 is a subgroup with an unspecified target attribute.

  Inappropriate score assignment, first, the inappropriate score of one group is set to 1, and the distribution of the inappropriate score in inverse proportion to the number of links in the sub group is used as the inappropriate score of each sub group. Further, an inappropriate score of each link is obtained by equally dividing the inappropriate score of each subgroup by the number of links in the subgroup. Therefore, the improper score of the links 241, 242 and 243 at step T41 is 1/12, and the improper score of the link 244 is 3/4. Here, the condition determination means 13 compares the sum of the inappropriate scores between the subgroups, and extracts a link having a high inappropriate score as a link mismatch. In FIG. 20, since the inappropriate score 3/4 of the link 144 is higher than the total 1/4 of the inappropriate scores of the links 241, 242, and 243, the link 144 is extracted as a link mismatch.

  Next, the candidate calculation means 12 obtains a correction candidate that matches the target attribute of a link with a large inappropriate score with the target attribute of a link with a small score within the same group (step T42 in FIG. 19). In the case of FIG. 20, “target attribute: _blank” is output as a correction candidate.

  In the present embodiment, the links to be grouped in step T41 are the same links in the link source document. However, in the link group in which the link source document is the same, a link existing in a specific area such as a table or a list of links is used. There is also a method of grouping only, and the method is not limited to the method described in this embodiment. There is also a method for grouping links between multiple documents, such as documents stored in the same directory as a specific document, based on the style, and detecting inconsistencies in the link style of the peripheral pages of the specific document. The method is not limited to the method described in the embodiment.

  Further, in the present embodiment, the method of detecting the target attribute inconsistency and obtaining the correction candidate has been described. However, the style attribute inconsistency detection and the correction candidate can be obtained by the same method.

  In the present embodiment, the calculation is based on the number of links in the subgroup as an example of the inappropriate score. However, the number of links in the subgroup is used as a feature vector of the link, and is given in advance as teacher data. There is a method of setting an average value of the distances to the feature vectors of inconsistent links as an inappropriate score, and the method is not limited to the method described in the present embodiment.

[Embodiment of ghost link detection]
Next, with reference to FIG. 7 and FIG. 21, the operations of the condition determination means 13 and the candidate calculation means 12 when detecting a ghost link will be described in detail.

  First, the condition determination means 13 extracts a link for which an invisible link source notation is designated from the information storage unit 22 (step T51 in FIG. 21). Here, the invisible link source notation means an empty character string, a transparent image, a very small image or character, an image or character having the same color as the background, and the like. In FIG. 7 (a), links for which an empty character string is specified in the link source notation are extracted.

  Next, the candidate calculation means 12 outputs the correction candidate as “link: delete” so as to delete the link (step T52 in FIG. 21).

[Embodiment of loop link detection]
Next, with reference to FIG. 8 and FIG. 22, the operation of the condition determining means 13 when detecting a loop link will be described in detail. The candidate calculation means 12 does not operate when a loop link is detected.

  First, the condition determination means 13 divides the link source notation of the link stored in the information storage unit 22 into words (step T61 in FIG. 22). There are several ways to divide the link source notation into words, such as using morphological analysis, cutting at the change of character type, or cutting every n characters.

  Next, the condition determining means 13 extracts a link group that is a series of links forming a loop and that has the same word in the link source notation corresponding to the series of links (step T62 in FIG. 22). ). In FIG. 8, since the links 261, 262, and 263 including the word “present” form a loop, they are output as a loop link.

  In this embodiment, a case has been described in which a loop link in which all the words in the link source notation are the same is extracted. However, a characteristic word dictionary is provided for each topic, and the word in the link source notation is There is also a method of extracting loop links that all belong to the same topic, and is not limited to the method described in this embodiment.

[Link mismatch detection method focusing on time change]
In the present embodiment, the method for detecting various link inconsistencies based on the link information of each link collected at a certain point in time has been described. However, the collection of link information is periodically repeated, and the time series of link information is obtained. There is also a method of detecting various link inconsistencies by paying attention to the change. With reference to FIGS. 4, 23, and 24, the operations of the condition determination unit 13 and the candidate calculation unit 12 when detecting various types of link inconsistency by paying attention to the time series change of the link information will be described in detail.

  It is assumed that the information storage means 22 stores link information at time T and time T ′.

  First, the condition determination means 13 groups links having the same item of link information at time T and time T ′ (step T71 in FIG. 23). In the case of FIG. 4, when the link information of the link 125 as of August 15, 2002 and the link information as of September 15, 2002 are grouped with the link of the document 125, the result is as shown in FIG.

  Next, a link whose link information has changed in the same group is extracted as a link mismatch (T72 in FIG. 23). In the case of Figure 23, as of August 15, 2002, there are four links to the document 125, but as of September 15, 2002, there is only one link to the document 125. Absent. Therefore, the link 224 is extracted as a link mismatch.

  Next, the candidate calculation means 12 outputs a correction candidate corresponding to the change occurring at time T and time T ′ (step T72 in FIG. 23). In the case of FIG. 23, since link deletion occurs on August 15, 2002 and September 15, 2002, “link: deletion” is output as a correction candidate.

  In this embodiment, the case has been described in which the deletion of the link occurs when the linked documents at time T and time T ′ group the same link, but the link source notation has changed. The candidate calculation means 12 outputs correction candidates so that the link source notation is changed, assuming that the contents of the linked document at time T have changed at time T ′.

  In the present embodiment, the method of grouping links having the same link destination document at time T and time T ′ has been described. There is a method for detecting a change in the target attribute, and the method is not limited to the method described in this embodiment.

  Next, the effect of this embodiment will be described.

  In this embodiment, various logical inconsistencies can be detected. That is, in the present embodiment, the link information is extracted from the hypertext database, and the links are grouped for each item of the link information, so that a unique link out of the group is detected as a link inconsistency. It is possible to detect logical inconsistencies such as () link errors, (2) links to expired information, (3) link source notation inconsistencies, and (4) link source notation style inconsistencies. (2) Detecting logical inconsistencies, such as links to expired information, by periodically collecting link information and detecting various link inconsistencies by focusing on time-series changes in link information can do. Furthermore, by detecting a link without a link source notation, (5) a ghost link, which is a form of logical inconsistency, can be detected, and is a sequence of links forming a loop, and the link sequence (6) Loop link, which is a form of logical inconsistency, can be detected by detecting all link source notations corresponding to the same topic.

  In the present embodiment, correction candidates for logical inconsistencies can be presented to the administrator. In other words, in this embodiment, since the correction candidate is obtained by processing such as automatically calculating the correction candidate so that the link information of the unique link out of the group becomes the same link information as the group, the administrator is not required. There is no need to consider how the alignment should be corrected, and the correction can be automatically reflected.

  In this embodiment, link inconsistencies are grouped and displayed together. Therefore, if the administrator confirms a part of the links, it can determine whether or not the remaining links are inconsistent as well, and the efficiency of checking by the administrator is greatly improved.

  Further, in the present embodiment, link mismatch is displayed as a list using any one of three items: (1) link source notation, (2) link source page identification information, and (3) link destination page identification information as a sort key. To do. Therefore, the administrator can grasp correction items for each page, focus on inconsistencies with important pages, and check the validity of expressions used as link source notations.

  Although the data processing apparatus 1 of the present embodiment includes the information collecting unit 11, the information collecting unit 11 collects and stores information about pages and links from the hypertext database 21 in another data processing apparatus. In such a form, the information collecting means 11 of the data processing apparatus 1 can be omitted. Further, in the configuration in which the inconsistent portion of the hypertext database 21 is corrected manually by the administrator himself / herself by looking at the result list screen of FIG. 13, the correction reflecting means 14 can be omitted. Furthermore, although the burden on the administrator increases, the candidate calculation means 12 can be obtained by the administrator himself / herself from the remaining information even if there is no type of inconsistency in the result list screen in FIG. A configuration in which is omitted can also be adopted.

“Second Embodiment”
Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 25, in the second embodiment of the present invention, the data processor 5 further calculates the importance in addition to the configuration of the data processor 2 in the first embodiment shown in FIG. It differs from the first embodiment in that the means 15 is provided.

  The importance level calculation means 15 calculates and ranks the importance level of the link inconsistency extracted by the condition determination means 13 according to the frequency of access to the document in which the link inconsistency is detected and the seriousness of the inconsistency. Output.

  Next, the operation of the present embodiment will be described in detail with reference to the drawings.

  The operations of the information collecting unit 11 and the condition determining unit 13 in the present embodiment shown in steps S1 to S3 in FIG. 26 are the same as the operations of the units 11 and 13 in the first embodiment, and thus description thereof is omitted. To do. Further, the candidate calculation means 12 is the same as the means 12 of the first embodiment in that a correction candidate for eliminating the mismatch is obtained for the link extracted by the condition determination means 13 as the link mismatch (step S4), the result list screen as shown in FIG. 13 is not output, and the control is transferred to the importance calculation means 15.

  The importance level calculation means 15 calculates the importance levels for the links extracted by the condition determination means 13 as link mismatches, ranks them, and outputs them (steps S8 and S9 in FIG. 26). Here, the importance levels are (1) type of error / inappropriate location detected, (2) error / incorrect accuracy of detected location, and (3) number of linked pages of the page including the detected location. , (4) User access to the page including the detected location, and (5) Hierarchy level in the hypertext of the page including the detected location.

  The output link mismatch ranking screen is shown in FIG. The difference from the result list screen of Fig. 13 is that the links with the same link destination and link source notation are grouped, and in addition to the type of inconsistency and correction candidates, the importance of inconsistency is given, and the importance of inconsistency The points are displayed in descending order. Therefore, the administrator can perform operations such as confirmation and rewriting of correction candidates in step S6 in descending order of importance of inconsistency. Thereafter, as in the first embodiment, each correction in the hypertext database 21 is corrected by the correction reflecting means 14 based on correction candidates confirmed or corrected by the administrator (step S7 in FIG. 26).

  In the present embodiment, the case has been described in which the importance calculation means 15 calculates and ranks the importance after the candidate calculation means 12 obtains the correction candidate, and the importance calculation means 15 first outputs the ranking. There is also a method in which the importance is calculated and ranked, and the candidate calculation means 12 later obtains correction candidates, and is not limited to the method described in the present embodiment.

  In this embodiment, the administrator checks the link inconsistency and correction candidate output in step S6, but omits step S6 and automates all of steps S1 to S4, S8, S9, and S7. There is also a method that is executed, and the method is not limited to the method described in this embodiment.

  Further, in the present embodiment, the case where the administrator decides and executes the inspection timing has been described. However, collection conditions and extraction conditions are set in advance, and steps S1 to S4, S8, and S9 are automatically performed periodically. There is also a method of executing the above and notifying the obtained result by e-mail or the like, and is not limited to the method described in this embodiment.

  Although the data processing device 5 of the present embodiment includes the information collecting unit 11, the information collecting unit 11 collects and stores information about pages and links from the hypertext database 21 in another data processing device. In such a form, the information collecting means 11 of the data processing device 5 can be omitted. In addition, the correction reflecting means 14 can be omitted in the configuration in which the inconsistent portion of the hypertext database 21 is corrected by the administrator's own manual operation while viewing the result list screen of FIG. In this case, although the burden on the administrator increases, it is possible to obtain correction candidates from the remaining information even if there is no type of inconsistency in the result list screen in FIG. A configuration in which 12 is omitted can also be adopted.

“Third embodiment”
Next, a third embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 28, in the third embodiment of the present invention, the data processing device 6 excludes the correction reflecting means 14 from the configuration of the data processing device 5 in the second embodiment shown in FIG. The second embodiment is different from the second embodiment in that a total score calculation means 16 is added.

  The total score calculation means 16 calculates the total score of the consistency of the site to be diagnosed based on the link mismatch output by the condition determination means 13 and the importance of the mismatch calculated by the importance level calculation means 15. Here, the total score is a method of using the number of link inconsistencies and the ratio of the number of link inconsistencies to the total number of links in addition to using the sum of inconsistency importance calculated by the importance calculation means 15. is there.

  Hereinafter, the operation of the present embodiment will be described in detail with reference to the drawings.

  The operations of the information collection unit 11, the candidate calculation unit 12, the condition determination unit 13, and the importance level calculation unit 15 in the present embodiment indicated by steps S1 to S4 and S8 in FIG. 29 are the same as those in the second embodiment. Since it is the same as the operation of 11, 12, 13, and 15, the description is omitted.

  In the second embodiment, after detecting the link mismatch, the correction is reflected in the hypertext database 21 according to the correction candidates. In the present embodiment, after detecting the link mismatch, the total score calculation means 16 calculates and outputs the total score of the entire diagnosis target site based on the importance obtained by the importance calculation means 15 (FIG. 29). Step S10).

  By periodically performing this total score calculation and outputting the time change as shown in FIG. 30, the progress of the site quality improvement can be known. In FIG. 30, the increase in the total score is saturated with time, and it can be seen that the work for improving the site quality is approaching convergence.

  Also, this total score calculation is performed periodically, and the site quality deteriorates by notifying the alert when a predetermined condition is met, such as the importance of the part detected as total score or link mismatch exceeds the threshold If you do, the site administrator can receive a warning.

  Further, by performing total score calculation for a plurality of different sites A to M, ranking and outputting as shown in FIG. 31, the site quality can be quantitatively compared. In FIG. 31, it can be seen that the quality of site A is about twice as good as that of site E.

  Next, the effect of this embodiment will be described.

  In the present embodiment, a total score of site quality to be diagnosed is calculated based on the number of detected link mismatches and the importance. Therefore, it is possible to grasp the progress of site quality improvement and to quantitatively compare the quality between different sites.

  Although the data processing device 6 of the present embodiment includes the information collecting unit 11, the information collecting unit 11 collects and stores information about pages and links from the hypertext database 21 in another data processing device. In such a form, the information collecting means 11 of the data processing device 6 can be omitted. Further, in the description of the present embodiment, the reflection (correction) of the detected inconsistent portion to the hypertext database 21 has not been described, but the reflection may or may not be performed. . In the case of reflecting, the inconsistent portion of the hypertext database 21 may be corrected manually by the administrator himself / herself by looking at the result list screen of FIG. 27, or the correction reflecting means similar to the second embodiment A configuration in which 14 is provided is also conceivable. Further, although the burden on the administrator increases, the candidate calculation means 12 can be obtained by the administrator himself / herself from the remaining information even if there is no inconsistency type and correction candidate in the result list screen of FIG. A configuration in which is omitted can also be adopted.

"Fourth embodiment"
Next, a fourth embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 32, the fourth embodiment according to the present invention includes an input means 501, a data processing device 502, an output means 503, and a storage device 504, as in the first embodiment of the present invention. Furthermore, a hypertext inspection program 500 for realizing the keyword extraction device of the first embodiment is provided.

  The input unit 501 is a device for inputting instructions from an operator, such as a mouse and a keyboard. The output unit 503 is a device that outputs a processing result of the data processing device 502 such as a display screen or a printer.

  The hypertext inspection program 500 is read into the data processing device 502, controls the operation of the data processing device 502, generates the input memory 505 and the work memory 506 in the storage device 504, and the first on the data processing device 502. The information collecting unit 11, the candidate calculating unit 12, the condition determining unit 13, and the correction reflecting unit 14 of FIG. The data processing device 502 executes the same processing as in the first embodiment by controlling a program for realizing the hypertext inspection device.

  The data processing device 1 in FIG. 1 corresponds to the data processing device 502 in FIG. 32, and the storage device 2 in FIG. 1 corresponds to the storage device 504 in FIG. However, the hypertext database 21 to be processed is obtained by accessing the database outside by the data processing device 502 via the network (for example, the Internet) in addition to using the data stored in the storage device 504. It may be.

"Fifth embodiment"
Next, a fifth embodiment of the present invention will be described in detail with reference to the drawings.

  As in the fourth embodiment, the fifth embodiment uses the configuration shown in FIG. The hypertext inspection program 500 is read into the data processing device 502, controls the operation of the data processing device 502, generates the input memory 505 and the work memory 506 in the storage device 504, and the second on the data processing device 502. The information collection unit 11, the candidate calculation unit 12, the condition determination unit 13, the correction reflection unit 14, and the importance calculation unit 15 shown in FIG. The data processing device 502 executes the same processing as that of the second embodiment under the control of a program for realizing the hypertext inspection device.

  The data processing device 5 in FIG. 25 corresponds to the data processing device 502 in FIG. 32, and the storage device 2 in FIG. 25 corresponds to the storage device 504 in FIG. However, the hypertext database 21 to be processed is obtained by accessing the database outside by the data processing device 502 via the network (for example, the Internet) in addition to using the data stored in the storage device 504. It may be.

"Sixth embodiment"
Next, a sixth embodiment of the present invention will be described in detail with reference to the drawings.

  The sixth embodiment uses the configuration shown in FIG. 32 as in the fourth embodiment. The hypertext inspection program 500 is read into the data processing device 502, controls the operation of the data processing device 502, generates the input memory 505 and the work memory 506 in the storage device 504, and also executes the third on the data processing device 502. The information collecting unit 11, the candidate calculating unit 12, the condition determining unit 13, the importance calculating unit 15, and the total score calculating unit 16 shown in FIG. The data processing device 502 executes the same processing as that of the third embodiment under the control of a program for realizing the hypertext inspection device.

  The data processing device 6 in FIG. 28 corresponds to the data processing device 502 in FIG. 32, and the storage device 2 in FIG. 28 corresponds to the storage device 504 in FIG. However, the hypertext database 21 to be processed is obtained by accessing the database outside by the data processing device 502 via the network (for example, the Internet) in addition to using the data stored in the storage device 504. It may be.

1 is a block diagram showing a configuration of a first exemplary embodiment of the present invention. It is a figure which shows the example of the designation | designated method of the link in hypertext, and the display example on a browser. It is a figure which shows the example of the logical inconsistency by an incorrect link. It is a figure which shows the example of the logical inconsistency by an expired link. It is a figure which shows the example of the logical inconsistency by inconsistency of a link origin description. It is a figure which shows the example of the logical inconsistency by the unification of the style of a link origin notation. It is a figure which shows the example of the logical inconsistency by a ghost link. It is a figure which shows the example of the logical inconsistency by a loop link. It is a figure which shows the example of the link information stored in an information storage part. 3 is a flowchart showing the operation of the first exemplary embodiment of the present invention. FIG. 5 is a diagram showing an example of a document collection condition setting screen in the first embodiment of the present invention. FIG. 6 is a diagram showing an example of a setting screen for a link mismatch extraction condition in the first embodiment of the present invention. It is a figure which shows the example of the inconsistency extraction result in the 1st Embodiment of this invention. 6 is a flowchart showing an operation of extracting an erroneous link in the first embodiment of the present invention. It is a figure which shows the example of the link information extracted at the time of the incorrect link extraction in the 1st Embodiment of this invention. 3 is a flowchart showing an operation of extracting an expired link in the first embodiment of the present invention. 5 is a flowchart showing an operation of extracting ununiformity of link source notation in the first exemplary embodiment of the present invention. It is a figure which shows the example of the link information extracted at the time of the ununiform extraction of the link origin notation in the 1st Embodiment of this invention. 6 is a flowchart showing an operation of extracting ununiformity of styles of a link source notation in the first exemplary embodiment of the present invention. It is a figure which shows the example of the link information extracted at the time of ununiform extraction of the style of the link origin notation in the 1st Embodiment of this invention. 6 is a flowchart showing an operation of extracting a ghost link in the first embodiment of the present invention. 3 is a flowchart showing an operation of loop link extraction in the first exemplary embodiment of the present invention. 5 is a flowchart showing an operation of extracting time change of link information in the first exemplary embodiment of the present invention. It is a figure which shows the example of the link information extracted at the time of the time change extraction of the link information in the 1st Embodiment of this invention. FIG. 5 is a block diagram showing a configuration of a second exemplary embodiment of the present invention. 6 is a flowchart showing the operation of the second exemplary embodiment of the present invention. It is a figure which shows the example of the mismatch extraction result in the 2nd Embodiment of this invention. FIG. 10 is a block diagram showing a configuration of a third exemplary embodiment of the present invention. 10 is a flowchart showing the operation of the third exemplary embodiment of the present invention. FIG. 16 is a diagram showing an example of a screen for outputting a change in total score over time in the third embodiment of the present invention. FIG. 10 is a diagram showing an example of a site ranking screen based on a total score in the third embodiment of the present invention. FIG. 10 is a block diagram showing configurations of fourth, fifth, and sixth embodiments of the present invention.

Explanation of symbols

1, 5, 6, 502 Data processing device
2,504 storage device
3,501 Input method
4, 503 Output means
11 Information collection means
12 Candidate calculation means
13 Condition judging means
14 Correction reflection means
15 Importance calculation method
16 Total score calculation means
21 Hypertext database
22 Information storage
500 Hypertext inspection program
505 input memory
506 Work memory

Claims (18)

A hypertext database that stores a collection of documents structured by hyperlinks;
Information collecting means for taking out each document from the hypertext database and taking out a link source address, a link destination address and a link source notation as link information included in each document;
Information storage means for storing the text and link information of the document retrieved by the information collecting means;
The link information including the date expression in the link source notation and the link information having the document whose date expression includes the date expression as the link destination address are extracted from the information storage means, and the date expression is compared with the current date and time. In addition, an expired expression is extracted from the document indicated by the link destination address of the extracted link information, and an inappropriate score of the link information is obtained based on a comparison result between the date expression and the current date and time and an extraction result of the expired expression. A hypertext inspection apparatus comprising: a condition determining unit that calculates link information with a high inappropriate score calculated as link information to expired information and outputs the detection result from an output unit .   Extracting the transfer destination address from the document indicated by the link destination address of the link information to the expiration information detected by the condition determining means, and correcting the link destination address of the link information with the extracted transfer destination address; The hypertext inspection apparatus according to claim 1, further comprising candidate calculation means for outputting a correction candidate to be indicated from the output means.   3. The hypertext inspection apparatus according to claim 2, wherein the candidate calculation unit outputs a correction candidate indicating that the link information is deleted from the output unit when the transfer destination address cannot be extracted. .   The correction reflecting means for correcting the link information to the expired information in the document stored in the hypertext database based on the correction candidate calculated by the candidate calculating means. 3. The hypertext inspection device according to 3.   5. The importance calculating means according to claim 1, further comprising importance calculating means for calculating importance of link information to the expiration information detected by the condition determining means and outputting the importance from the output means. Hypertext inspection device.   The importance calculated by the importance calculation means, the number of link information to the expiration information detected by the condition determination means, the number of link information to the expiration information detected by the condition determination means for the total number of links 6. A total score calculating means for calculating a total score related to the document set structured by the hyperlink by a combination of one factor or a plurality of factors of the ratio and outputting the total score from the output means. The described hypertext inspection device. An information collecting unit that extracts each document from a hypertext database storing a document set structured by hyperlinks, and extracts a link source address, a link destination address, and a link source notation as link information included in each document;
The information collecting means storing the text and link information of the retrieved document in an information storage means;
Condition determining means extracts from the information storage means link information whose link source notation includes a date expression and link information whose text contains a date expression as a link destination address. Compare the date and time, extract an expired expression from the document indicated by the link destination address of the extracted link information, and based on the comparison result of the date expression and the current date and time and the extracted result of the expired expression Calculating an inappropriate score for information;
A hypertext inspection method comprising: detecting the link information with the calculated inappropriate inappropriate score as link information to expired information, and outputting the detection result from the output means; .   The candidate calculation means extracts the transfer destination address from the document indicated by the link destination address of the link information to the expiration information detected by the condition determination means, and the link destination address of the link information by the extracted transfer destination address The hypertext inspection method according to claim 7, further comprising a step of outputting a correction candidate indicating that the correction is to be performed from the output unit.   9. The hyperlink according to claim 8, further comprising the step of outputting, from the output unit, a correction candidate indicating that the link information is to be deleted when the candidate calculation unit cannot extract the transfer destination address. Text inspection method.   The correction reflecting means includes a step of correcting link information to expiration information in a document stored in the hypertext database based on the correction candidates calculated by the candidate calculating means. Item 10. The hypertext inspection method according to Item 8 or 9.   11. The importance level calculation means includes a step of calculating the importance level of link information to expired information detected by the condition determination means and outputting the importance level from the output means. The hypertext inspection method described in 1.   Total score calculation means, the importance calculated by the importance calculation means, the number of link information to the expiration information detected by the condition determination means, the expiration information detected by the condition determination means for the total number of links A step of calculating a total score relating to the document set structured by the hyperlink by a combination of one factor or a plurality of factors of the ratio of the number of link information and outputting from the output means is provided. Item 12. The hypertext inspection method according to Item 11. A computer comprising a hypertext database for storing a set of documents structured by hyperlinks and information storage means,
Information that retrieves each document from the hypertext database, retrieves a link source address, a link destination address, and a link source notation as link information included in each document, and stores the text and link information of the retrieved document in the information storage unit Collection means;
The link information including the date expression in the link source notation and the link information having the document whose date expression includes the date expression as the link destination address are extracted from the information storage means, and the date expression is compared with the current date and time. In addition, an expired expression is extracted from the document indicated by the link destination address of the extracted link information, and an inappropriate score of the link information is obtained based on a comparison result between the date expression and the current date and time and an extraction result of the expired expression. A hypertext inspection program for detecting the calculated link information having a high inappropriate score as link information to the expired information and functioning as a condition determination means for outputting the detection result from the output means.   The computer further extracts a transfer destination address from the document indicated by the link destination address of the link information to the expiration information detected by the condition determination means, and the link destination of the link information by the extracted transfer destination address 14. The hypertext inspection program according to claim 13, wherein the hypertext inspection program functions as candidate calculation means for outputting a correction candidate indicating that the address is corrected from the output means.   15. The hypertext inspection program according to claim 14, wherein the candidate calculation means outputs a correction candidate indicating that the link information is deleted from the output means when the transfer destination address cannot be extracted. .   Further, the computer is caused to function as a correction reflection unit that corrects link information to the expiration information in the document stored in the hypertext database based on the correction candidates calculated by the candidate calculation unit. The hypertext inspection program according to claim 14 or 15.   16. The computer according to claim 13, 14, 15, or 15 for causing the computer to further function as an importance calculation means for calculating the importance of link information to the expiration information detected by the condition determination means and outputting the importance from the output means. 16. The hypertext inspection program according to 16.   The computer further includes the importance calculated by the importance calculation means, the number of link information to the expiration information detected by the condition determination means, and the expiration information detected by the condition determination means for the total number of links. For calculating the total score for the document set structured by the hyperlink by one factor or a combination of a plurality of factors in the ratio of the number of link information and outputting from the output means The hypertext inspection program according to claim 17.

Images