JP5462591B2 - Specific content determination device, specific content determination method, specific content determination program, and related content insertion device - Google Patents

Description

  The present invention relates to a technical field for extracting content constituting a Web page.

  2. Description of the Related Art Conventionally, there is known a technique for acquiring content that is a material constituting a Web page published on a Web site and generating new content based on the acquired content. For example, Non-Patent Document 1 discloses a technology that, when a URL of image data is specified by a user, acquires image data corresponding to the URL from the Web, and automatically creates a banner based on the acquired image data. Has been.

“Automatic banner creation”, [online], [October 21, 2009 search], Internet <URL: http://hyperbannermaker.com/>

  Each web page constituting the web site is posted with content in accordance with the purpose of the web site. For this reason, the contents of the Web pages constituting the Web site are basically related to each other, but may have some characteristics. And the factor which determines the content of the web page is the content of the content (for example, text data, image data, etc.) which comprises the web page. Therefore, content that characterizes the Web page among content constituting the Web page, that is, content specific to the Web page may exist.

  The technique described in Non-Patent Document 1 extracts content unique to a Web page, but does not automatically extract content, and the user must manually specify the content. It is not possible to easily extract content peculiar to. Therefore, when the user cannot determine which content is specific to the Web page, or when the favorite content is biased depending on the user's preference, the content specific to the Web page is properly extracted. I can't. In addition, when the number of target Web pages is large, there is a problem that the user's work becomes enormous.

  Further, for example, it is possible to extract all of a specific type of content such as only an image or only text based on the description of a tag of an HTML (HyperText Markup Language) document. However, since the extracted content includes common content that is not unique to Web pages, the reliability of the extraction result is low, and the user must search for the specific content from the extraction result. It was.

  The present invention has been made in view of the above points, and a unique content determination apparatus and a specific content determination method capable of easily extracting content specific to a Web page from content constituting the Web page. An object is to provide a unique content determination program and the like.

In order to solve the above-mentioned problem, the invention described in claim 1 is characterized in that an extracting means for extracting content constituting a specified Web page among a plurality of Web pages included in a predetermined site; Among the web pages, the calculation means for counting the frequency with which each of the contents constituting the designated web page is used on another web page, and among the contents constituting the designated web page, And determining means for determining that content having a frequency of being used in another Web page is equal to or less than a predetermined value as content specific to the specified Web page.

According to this invention, the frequency with which each content which comprises the designated web page is used by several web pages among the several web pages contained in a predetermined site is counted, respectively. The content that is used less frequently on other Web pages is the content that does not appear much other than the specified Web page. Therefore, by determining whether the frequency used in other Web pages is equal to or less than a predetermined value, all content satisfying the conditions is specified as content specific to the specified Web page. Therefore, content unique to the Web page can be easily extracted.

Invention according to claim 2, in specific content determination device according to claim 1, wherein the extraction means, in units of content group consists of one or more content and content constituting the Web page The calculation means counts the frequency with which the content group constituting the designated web page is used on another web page, and the determination means constitutes the designated web page Among the content groups, it is determined that a content group whose frequency used in other Web pages is a predetermined value or less is a content group specific to the specified Web page.

  According to the present invention, content specific to the Web page is determined in units of content groups. For example, content that is displayed in a certain unit on the Web page or is related to each other is displayed. When a content group is used, it is possible to extract content that is unique to a Web page.

According to a third aspect of the present invention, in the specific content determination apparatus according to the second aspect , the extraction means is a content based on document data that is described in a predetermined markup language and that indicates the content constituting the Web page. It is characterized by extracting groups.

  According to the present invention, since the content group is extracted based on the document data indicating the content constituting the Web page, the content group can be extracted appropriately.

According to a fourth aspect of the present invention, in the specific content determination apparatus according to the third aspect , the extracting unit determines a content group based on a predetermined tag in document data indicating the content. .

  According to the present invention, the content group is extracted based on a predetermined tag. Therefore, when content specific to a Web page and content that is not specific are grouped by a predetermined tag, respectively. Therefore, it is possible to increase the accuracy of determining content unique to the Web page.

According to a fifth aspect of the present invention, in the specific content determination device according to any one of the first to fourth aspects, the extracting unit posts the article from a Web page on which the posted article is posted. A classifying unit that extracts the comment that has been extracted, and classifies each extracted comment according to the content indicated by the comment, and a setting unit that sets a threshold of appearance frequency, and the number of the content into which the comment is classified Setting means for reducing the threshold as much as possible, the calculating means calculates the appearance frequency of each of the contents into which the comment is classified in the Web page, and the determining means is calculated by the calculating means The content whose appearance frequency is not more than the set threshold value is determined to be content specific to the Web page.
The invention according to claim 6 is an extraction step of extracting content constituting a designated web page from among a plurality of web pages included in a predetermined site, and the designation among the plurality of web pages. a calculation step of the contents constituting the Web page counts the frequency used in other Web pages, of the contents constituting the Web page that the specified frequency used in other Web pages And a determination step of determining that the content equal to or less than a predetermined value is content specific to the designated Web page.
The invention described in claim 7 in specific content determination method according to claim 6, wherein the extracting step extracts a comment posted on the article from a Web page articles posted is posted, the A classification step for classifying each extracted comment according to the content indicated by the comment, and a setting step for setting a threshold of appearance frequency. A setting step for decreasing the threshold as the number of the contents into which the comment is classified increases. And the calculating step calculates the appearance frequency of each content in the Web page into which the comment is classified, and the determining step includes setting the appearance frequency calculated by the calculating step to the set threshold value. The content described below is determined as content specific to the Web page.

The invention according to claim 8 is an extraction means for extracting a content constituting a specified Web page from among a plurality of Web pages included in a predetermined site, and among the plurality of Web pages, calculation means for each content constituting the specified Web page is to count the frequency to be used in other Web pages, and, among the contents constituting the specified Web pages, other Web pages It is characterized by functioning as a determination means for determining that a content having a frequency of use or less is a content specific to the designated Web page.
The invention according to claim 9 is the specific content determination program according to claim 8 , wherein the extraction unit extracts a comment posted to the article from a Web page on which the posted article is posted, The computer is a classification unit that classifies each extracted comment according to the content indicated by the comment, and a setting unit that sets a threshold of appearance frequency. The larger the number of the content into which the comment is classified, Further functioning as a setting means for reducing the threshold value, the calculating means calculates the appearance frequency of each of the contents in which the comment is classified, and the determining means calculates the appearance frequency calculated by the calculating means. It is determined that the content that is equal to or less than the set threshold is content specific to the Web page.

According to an eleventh aspect of the present invention, there is provided the specific content determination device according to any one of the first to sixth aspects and related content related to the content determined to be specific content by the specific content determination device. And insertion means for inserting into the designated Web page.

  According to the present invention, the content related to the content determined to be the specific content is inserted into the specified Web page, so that information related to the characteristics of the Web page can be added to the Web page. it can.

According to a twelfth aspect of the present invention, in the related content insertion device according to the eleventh aspect , the unique content determination device has the text data of the posted article as the content constituting the designated Web page. If it is included, the text data is determined to be content specific to the Web page, and the specified Web page is determined from text data of an article determined to be specific content by the specific content determination device. A feature word extracting means for extracting the feature word of each of the above, and a selecting means for selecting content related to the extracted feature word as the related content from among a plurality of contents stored in the storage means in association with each word , And the inserting means inserts the selected related content into the designated Characterized in that it inserted into the eb page.

According to this invention, if the text data of each article includes contents specific to the article, the text data of each article can be extracted by the unique content determination device. Thereby, the information relevant to the content of the article published on the web page can be added to the web page.

According to the present invention, the content that is used less frequently on other Web pages is the content that does not appear much on the plurality of Web pages other than the specified Web page, and is therefore used on other Web pages. By determining whether the frequency is equal to or less than a predetermined value, all the content satisfying the condition is specified as content specific to the designated Web page. Therefore, content unique to the Web page can be easily extracted.

It is a figure showing an example of outline composition of blog system S concerning one embodiment. It is a block diagram which shows an example of schematic structure of the blog server 1 which concerns on one Embodiment. It is a figure which shows the outline | summary of a process after advertisement content is inserted in a blog page after a blogger is designated. It is a figure which shows the structural example of a web page. It is a figure which shows an example of the DOM tree produced | generated from the HTML document. 4 is a diagram illustrating an example of content of content block correspondence information stored in a storage unit 15. FIG. It is a flowchart which shows the process example in the advertisement content insertion process of the system control part 20 of the blog server 1 which concerns on one Embodiment. It is a flowchart which shows the process example in the 1-page corresponding | compatible extraction process of the system control part 20 of the blog server 1 which concerns on one Embodiment. It is a flowchart which shows the process example in the tree search process of the system control part 20 of the blog server 1 which concerns on one Embodiment. It is a flowchart which shows the process example in the specific content block determination process of the system control part 20 of the blog server 1 which concerns on one Embodiment. It is a flowchart which shows the process example in the process at the time of the blog update of the system control part 20 of the blog server 1 which concerns on the modification of one Embodiment.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The embodiment described below is an embodiment when the present invention is applied to a server device that transmits a blog page in a blog system that provides a blog service.

[1. Overview of Blog System Configuration and Functions]
First, the configuration and outline function of the blog system S according to the present embodiment will be described with reference to FIG.

  FIG. 1 is a diagram illustrating an example of a schematic configuration of a blog system S according to the present embodiment.

  As illustrated in FIG. 1, the blog system S includes a blog server 1 as an example of a specific content determination device and a related content insertion device, a management terminal 2, and a plurality of user terminals 3. The blog server 1 and each user terminal 3 can transmit / receive data to / from each other using, for example, TCP / IP as a communication protocol via the network NW. The network NW is constructed by, for example, the Internet, a dedicated communication line (for example, a CATV (Community Antenna Television) line), a mobile communication network (including a base station), a gateway, and the like. The blog server 1 and the management terminal 2 are connected via a network such as a LAN (Local Area Network).

  In the blog system S having such a configuration, the blog server 1 is a Web server that transmits a Web page constituting the blog service site in response to a request from the user terminal 3. When a user who uses the user terminal 3 registers as a user of a blog service site, the user can operate his / her blog on the blog service site. The registered user (blogger) can access the blog service site and update his / her blog (add blog articles (records for each blog)). Therefore, the blog server 1 generates or updates a blog page on which one or a plurality of blog articles are posted as a blog web page in accordance with the update of the blog. The blog server 1 includes a blog page DB 101 and registers the blog page in the blog page DB 101.

  Further, the blog server 1 inserts advertising content (an example of related content) into the blogger blog page designated by the system administrator. The advertisement content includes, for example, text data of advertisement text, banner image data, moving image data, a rich internet application (RIA) generated by Adobe Flash (trademark), Silverlight (trademark), and the like. The advertising content inserted into each blog page is content indicating an advertisement related to a product or service related to the blog article posted on the target blog page. Therefore, the blog server 1 includes an advertisement DB 102 in which a plurality of advertisement contents are registered. Then, the blog server 1 extracts a blog article from the blog page, further extracts a feature word from the blog article, and selects advertisement content related to the extracted feature word.

  The user terminal 3 is a terminal device used by a user as a blogger or a user browsing a blog. As the user terminal 3, for example, a personal computer, a PDA, a mobile phone or the like is used.

  The management terminal 2 is a terminal device used by a system administrator of the blog system S. For example, a personal computer or the like is used as the management terminal 2.

[2. Blog server configuration and functions]
Next, the configuration and function of the blog server 1 will be described with reference to FIG.

  FIG. 2 is a block diagram illustrating an example of a schematic configuration of the blog server 1 according to the present embodiment. FIG. 3 is a diagram showing an outline of processing from when a blogger is designated until advertisement content is inserted into the blog page. FIG. 4 is a diagram illustrating a configuration example of a Web page. FIG. 5 is a diagram illustrating an example of a DOM tree generated from an HTML document. FIG. 6 is a diagram illustrating an example of the content of the content block correspondence information stored in the storage unit 15.

  As shown in FIG. 2, the blog server 1 includes an operation unit 11, a display unit 12, a communication unit 13, a drive unit 14, a storage unit 15 as an example of a storage unit, an input / output interface unit 16, And a system control unit 20. The system control unit 20 and the input / output interface unit 16 are connected via a system bus 21.

  The operation unit 11 includes, for example, a keyboard and a mouse, and receives an operation instruction from a system administrator or the like, and outputs the instruction content to the system control unit 20 as an instruction signal. The display unit 12 includes, for example, a CRT (Cathode Ray Tube) display, a liquid crystal display, and the like, and displays information such as characters and images. The communication unit 13 is connected to a network NW or the like and controls a communication state with the management terminal 2, the user terminal 3, and the like. For example, the drive unit 14 reads data from a disk DK such as a flexible disk, a CD (Compact Disc), a DVD (Digital Versatile Disc), and the like, and records data on the disk DK.

  The storage unit 15 is configured by, for example, a hard disk drive or the like, and stores various programs, data, and the like. In the storage unit 15, a blog page DB 101 and an advertisement DB 102 are constructed. In the blog page DB 101, each blog page (HTML document (an example of document data) of the blog page, image data that is the material of the blog page, etc.) constituting the blog service site includes, for example, the URL and blogger identification of the page It is registered in association with the user ID which is information. In the advertisement DB 102, a plurality of advertisement contents are registered in association with keywords related to products and services to be advertised by the advertisement contents. Here, when content other than text data is included in the advertising content, the URL of the content is also registered in association with it. In addition, when the user selects the advertisement content displayed on the blog page, the URL of the Web page is also registered in association with the Web page related to the advertisement target product or service.

  The input / output interface unit 16 performs interface processing between the operation unit 11 to the storage unit 15 and the system control unit 20. The system control unit 20 includes a CPU (Central Processing Unit) 17, a ROM (Read Only Memory) 18, a RAM (Random Access Memory) 19, and the like.

  The system control unit 20 controls each unit of the blog server 1 by the CPU 17 reading and executing various programs stored in the ROM 18 and the storage unit 15. The system control unit 20 functions as an extraction unit, a calculation unit, a determination unit, and an insertion unit by executing advertisement content insertion software (an example of a specific content determination program). Note that the advertisement content insertion software or the like may be acquired from another server device or the like via the network NW, or may be recorded on a disk DK such as a CD-ROM and read via the drive unit 14. You may make it.

  Advertising content insertion software is a program for inserting advertising content into a blog page. As shown in FIG. 3, the advertisement content insertion software includes a manager unit, a material extraction engine, a sentence analysis engine, an advertisement selection unit, and the like. The manager unit controls execution of the material extraction engine, the sentence analysis engine, and the advertisement selection unit. The material extraction engine is software for extracting content as a Web material from an HTML document of a blog page and determining content specific to the blog page. Content extraction is performed in units of content blocks (an example of a content group) described later. In the present embodiment, for example, a blog article including content specific to the article corresponds to a content block specific to the blog page.

  The sentence analysis engine is software for extracting feature words of the blog page from the blog article extracted as content unique to the blog page. The advertisement selection unit is software for selecting advertisement content related to the blog page using the extracted feature words as keywords.

  Below, the outline | summary of insertion of advertisement content is demonstrated. As shown in FIG. 3, the system administrator designates the user ID of the target blogger (1). Then, the system control unit 20 acquires and analyzes the HTML documents of all the blog pages corresponding to the designated user ID from the blog page DB 101, and extracts the content as the Web material in units of content blocks. Then, as the extraction result, content block correspondence information (an example of content information) is generated for each extracted content block (2). Next, the system control unit 20 calculates the appearance frequency of each extracted content block on all blog pages corresponding to the specified user ID. The appearance frequency calculated in the present embodiment is, for example, the number of appearances (frequency). Then, the system control unit 20 determines in each blog page that a content block whose appearance frequency is equal to or less than a predetermined threshold is a content block unique to the blog page (3).

  The system control unit 20 performs analysis such as morphological analysis on the content block determined to be a specific content block, that is, a blog article, and extracts a feature word for each block page (4). There are various methods for extracting feature words, and since they are well-known, detailed description thereof is omitted. As an example, a word having the highest appearance frequency is used as a feature word.

  Next, the system control unit 20 refers to the advertisement DB 102 and selects an advertisement content related to the extracted feature word (5). Then, the system control unit 20 inserts a rule (such as a description of a tag or data itself) for inserting and displaying the selected advertisement content on the blog page and displaying it in the HTML document of the blog page (6).

  Next, a content block extraction method will be described. In the present embodiment, it is assumed that text data and image data are extracted as Web materials.

  For example, assume that the configuration (layout) of the blog page is as shown in FIG. Each content as a Web material is displayed for each certain group (group) on the blog page. Each group corresponds to a content block. Each content is divided into content blocks by a DIV tag and a TABLE tag (an example of a predetermined tag) described in the HTML document. That is, each content is blocked (grouped) by the DIV tag and the TABLE tag (hereinafter referred to as “blocked tag”).

  In FIG. 4, content blocks 301 to 306 are displayed. The content block 301 is, for example, a content block in the header portion of the page, and is composed of a text A and an image a. The content block 302 is, for example, a content block of a navigation part for moving to Web pages related to products of various categories, and is composed of, for example, text B, text C, and text D indicating links to other Web pages. Has been. The content block 303 is, for example, a content block corresponding to a blog display area, and includes a text E indicating a headline such as a blog, a content block 304, and a content block 305. In this way, the content blocks may be nested, that is, have a hierarchical structure. In this case, the content included in the content block 303 is only the text E, and the content block 304 and the content block 305 are independent of the content block 303. Each of the content blocks 304 and 305 is one blog article. The content block 304 includes text F and G indicating the title and body of the blog article. The content block 305 includes texts H, I, and J indicating the title and body of a blog article, and images b and c registered by a blogger in relation to the blog article. The content block 306 is, for example, a content block indicating copyright display, and is composed of text I.

  Of these content blocks, content blocks 301, 302, 303, and 306 appear relatively frequently on blog pages other than the blog page shown in FIG. On the other hand, the content block 304 and the content block 305 are basically used only for the blog page. Therefore, it is determined that the content block 304 or the content block 305 is a content block unique to the blog page.

  In the present embodiment, it is necessary to determine a content block corresponding to a blog article including the content specific to the article as a specific content block. In some cases, a plurality of blog articles including such specific contents are included in one page. For this reason, all content blocks whose appearance frequency is equal to or lower than a predetermined threshold are set as unique content blocks. For example, the threshold value is set to once. Then, a blog article including specific content is determined as a specific content block, and a blog article including only content similar to other blog articles is not determined as a specific content block. In addition, since the appearance frequency of content blocks that are common to each blog page such as the header portion, the navigation portion, and the copyright display portion is two times or more, these are not determined to be specific content blocks. The threshold value is stored in advance in the storage unit 15.

  FIG. 5 shows the HTML document of the blog page shown in FIG. 4 in a DOM (Document Object Model) tree, that is, a tree structure. In the DOM tree shown in FIG. 5, illustration of tag nodes that are not necessary for the description of the present embodiment is omitted.

  In the DOM tree, a DIV node indicating a DIV tag and a TABLE node indicating a TABLE tag are nodes that block each content into content blocks (hereinafter referred to as “blocked nodes”). For example, the system control unit 20 searches the DOM tree by depth-first search and determines the content block. Specifically, when the system control unit 20 finds a blocked node, the contents defined in each node of the partial tree having the node as a vertex are grouped into a content block. However, as a result of hierarchical definition of the content blocks, after finding a certain blocked node (hereinafter referred to as “higher level blocked node”), further blocked nodes from the descendant nodes of the blocked node When it is found (hereinafter referred to as “lower block node”), the content block is divided. For example, assuming that a node closer to the root node is a higher-level node, a content block corresponding to a subtree having a higher-level blocked node as a vertex (hereinafter referred to as “higher-level subtree”) It is divided into a content block corresponding to a subtree having a lower blocked node as a vertex (hereinafter referred to as “lower subtree”) and a content block corresponding to a portion of the upper subtree excluding the lower subtree (for example, Content block 304 and content block 303). In this case, the former content block is hierarchically lower than the latter content block. For example, the hierarchy of the content blocks 301, 302, 303, and 306 is 1, and the hierarchy of the content blocks 304 and 305 is 2. In other words, the lower the hierarchy value, the higher the hierarchy.

  This will be explained in the description of the tag of the HTML document. Basically, if a blocked tag is described, the contents whose specifications are described within the range sandwiched between the blocked tags are collectively displayed in the block. The content block corresponds to the categorized tag. However, when the blocking tag is described hierarchically, the block described below the blocking tag among the contents in which the specification is described within a range sandwiched between certain blocking tags A content block corresponding to the block tag is collectively obtained by excluding the content whose definition is described in the range between the block tags.

  When the content block is extracted in this way, the system control unit 20 temporarily stores content block correspondence information indicating the extraction result in the storage unit 15. As shown in FIG. 6, the content block correspondence information (reference numeral 401) is stored for each content block. The content block correspondence information includes a URL setting part (reference numeral 402) and block configuration information (reference numeral 403) of the extraction source HTML document. Each extracted content is set in the block configuration information. Here, for text data, the contents of the text node in the DOM tree are set. On the other hand, for the image data, instead of the image data itself, the URL of the image data is set as the src attribute of the IMG node indicating the IMG tag in the DOM tree. In the present embodiment, since feature words are extracted from a content block determined to be a blog page-specific content block, that is, a blog article, text data may be extracted, and image data is extracted. It is not necessary.

[3. Operation of the blog system]
Next, the operation of the blog system S will be described with reference to FIGS.

  FIG. 7 is a flowchart illustrating a processing example in the advertisement content insertion processing of the system control unit 20 of the blog server 1 according to the present embodiment.

  The advertisement content insertion process is started, for example, when a request for execution of the advertisement content insertion process is transmitted from the management terminal 2 based on the operation of the system administrator.

  Then, when the system administrator specifies the user ID of the blogger who operates the blog to which the advertising content is to be inserted, the system control unit 20 receives the specified user ID from the management terminal 2 as shown in FIG. (Step S1).

  Next, the system control unit 20 sets 0 to the block number NUM (step S2). The block number NUM is the number of content blocks that have been discovered at the present time. NUM is a global variable and can be accessed from a one-page extraction process and a tree search process described later.

  Next, the system control unit 20 acquires the HTML document of the first blog page corresponding to the received user ID from the blog page DB 101 (step S3). Next, the system control unit 20 designates the acquired HTML document and executes a one-page extraction process described later (step S4). In this one-page extraction process, a content block is extracted from the acquired HTML document, and the content block correspondence information is stored.

  Next, the system control unit 20 determines whether or not the content blocks of all the blog pages corresponding to the received user ID have been extracted (step S5). At this time, if there is a blog page from which no content block has been extracted (step S5: NO), the system control unit 20 acquires the HTML document of the next blog page from the blog page DB 101 (step S6). The process proceeds to step S4. And if the system control part 20 repeats the process of step S4-S6 and extracts the content block of all the blog pages (step S5: YES), it will transfer to step S7.

  In step S7, the system control unit 20 specifies the HTML document of the first blog page corresponding to the received user ID.

  Next, the system control unit 20 designates the acquired HTML document and executes a specific content block determination process described later (step S8). In the specific content block determination process, a content block is extracted from the specified HTML document, and a content block specific to the blog page is determined.

  Next, the system control unit 20 extracts feature words of the blog page from each text data constituting the content block determined to be unique (step S9). Next, the system control unit 20 inserts an advertisement page related to the blog page into the blog page based on the extracted feature words (step S10). Specifically, the system control unit 20 uses the extracted feature word as a keyword, refers to the advertisement DB 102, and selects advertisement content corresponding to the keyword. Next, the system control unit 20 inserts the rule of the selected advertisement content at a predetermined position on the identified HTML document. For example, if the advertising content includes text data, the system control unit 20 adds the content of the text data to the HTML document. For example, when image data is included in the advertisement content, the system control unit 20 adds an IMG tag for displaying the image data to the HTML document. Further, for example, the system control unit 20 adds link information to the Web page related to the advertisement target product or service to the HTML document.

  When the advertisement content rule is inserted into the specified HTML document, the system control unit 20 updates the HTML document registered in the blog page DB 101 with the HTML document (step S11).

  Next, the system control unit 20 determines whether or not advertisement content has been inserted into all the blog pages corresponding to the received user ID (step S12). At this time, if there is a blog page in which no advertising content is inserted (step S12: NO), the system control unit 20 specifies the HTML document of the next blog page (step S13), and proceeds to step S8. Transition. And if the system control part 20 repeats the process of step S8-S13 and inserts an advertising content in all the blog pages (step S12: YES), all the content block corresponding | compatible information preserve | saved at the memory | storage part 15 will be shown. The data is deleted from the storage unit 15 (step S14). After completing this process, the system control unit 20 ends the advertisement content insertion process.

  FIG. 8 is a flowchart showing a processing example in the one-page extraction process of the system control unit 20 of the blog server 1 according to this embodiment.

  As shown in FIG. 8, the system control unit 20 first generates a DOM tree of the acquired HTML document on the RAM 19 (step S21).

  Next, the system control unit 20 sets 0 in the hierarchy LV (step S22). The hierarchy LV is a hierarchy of content blocks to which the currently searched node belongs in the DOM tree. LV is a global variable and can be accessed from the one-page correspondence extraction process and the tree search process described later.

  Next, the system control unit 20 designates the root node of the DOM tree (step S23) and executes tree search processing (step S24). The tree search process can be recursively called. By this tree search process, all content blocks are extracted from the Web page, and content block correspondence information is generated.

  Next, the system control unit 20 saves each content block correspondence information generated by the tree search process in the storage unit 15 (step S25). When completing this process, the system control unit 20 ends the one-page correspondence extraction process.

  FIG. 9 is a flowchart illustrating a processing example in the tree search process of the system control unit 20 of the blog server 1 according to the present embodiment.

  As shown in FIG. 9, the system control unit 20 first determines the type of the designated node (step S31). At this time, when the designated node type is a DIV node or a TABLE node (blocked node), that is, when a content block is found (step S31: DIV or TABLE), the process proceeds to step S32.

  In step S32, the system control unit 20 adds 1 to the block number NUM and adds 1 to the hierarchy LV. Next, the system control unit 20 sets NUM in the block number BN [LV] (step S33). The block number BN [LV] is the block number of the content block indicated by the hierarchy LV to which the currently searched node belongs. This block number is assigned in the order of discovery of content blocks. BN [LV] is a global variable.

  Next, the system control unit 20 initializes content block correspondence information corresponding to the content block of the block number BN [LV] (step S34). Specifically, the system control unit 20 sets an area for storing the content block correspondence information on the RAM 19 and sets the URL of the acquired HTML document in the area.

  Next, the system control unit 20 determines whether there is a child node that has not been searched among the child nodes of the designated node (step S35). At this time, if there is a child node that has not been searched yet (step S35: YES), the system control unit 20 proceeds to step S36.

  In step S36, the system control unit 20 designates one of the unsearched child nodes and executes tree search processing (step S37). After completing the tree search process, the system control unit 20 proceeds to step S35.

  And the system control part 20 will repeat the process of step S35-S37, will complete | finish the tree search process of all the child nodes (step S35: NO), and will transfer to step S38. Note that the system control unit 20 also proceeds to step S38 when there is no child node of the designated node. In step S38, the system control unit 20 subtracts 1 from the hierarchy LV and ends the tree search process.

  In step S31, when the type of the designated node is a text node (step S31: text), the system control unit 20 displays the content (text data) of the designated node with the block number BN [LV]. It is additionally set in the block configuration information in the content block correspondence information corresponding to the content block (step S39). After completing this process, the system control unit 20 ends the tree search process.

  In step S31, if the type of the designated node is an IMG node (step S31: IMG), the system control unit 20 acquires the URL of the image data set as the src attribute of the designated node. The acquired URL is additionally set in the block configuration information in the content block correspondence information corresponding to the content block of the block number BN [LV] (step S40). After completing this process, the system control unit 20 ends the tree search process.

  In step S31, when the type of the designated node is not any of the DIV node, the TABLE node, the text node, and the IMG node (step S31: Other), the system control unit 20 is a child of the designated node. It is determined whether there is a child node that has not been searched for among the nodes (step S41). At this time, if there is a child node that has not been searched yet (step S41: YES), the system control unit 20 designates one of the child nodes that have not been searched (step S42). Then, tree search processing is executed (step S43). After completing the tree search process, the system control unit 20 proceeds to step S41.

  On the other hand, when the system control unit 20 finishes the tree search process for all the child nodes of the designated node, or when there is no child node of the designated node (step S41: NO), The tree search process is terminated.

  FIG. 10 is a flowchart illustrating a processing example in the specific content block determination process of the system control unit 20 of the content generation server 1 according to the present embodiment.

  As shown in FIG. 10, the system control unit 20 first generates a DOM tree of the specified HTML document (step S61), sets 0 for the block number NUM and the hierarchy LV, as in the one-page extraction process. (Step S62), the root node of the DOM tree is designated (step S63), and the tree search process is executed (step S64).

  Next, the system control unit 20 sets 1 to the block number i (step S65). Next, the system control unit 20 calculates the appearance frequency of the content block with the block number i (step S66).

  Specifically, the system control unit 20 stores the block configuration information of the content block correspondence information i (the content block correspondence information corresponding to the content block with the block number i) generated in the tree search process in step S64 and the storage unit 15. The block configuration information of each stored content block correspondence information is compared. At this time, if the contents of the block configuration information match, the system control unit 20 counts the number of appearances as one. At this time, the system control unit 20 may ignore the content order in the block configuration information. In addition, the system control unit 20 includes a part of the contents specified in the block configuration information of the content block correspondence information i in which all contents specified in the block configuration information of the content block correspondence information stored in the storage unit 15 Even if it matches the content, it may be counted as one appearance. Furthermore, when comparing the text data defined in the block configuration information of the content block correspondence information, the system control unit 20 does not determine whether or not the sentences etc. indicated by the text data match. Instead, the substantial contents expressed by the sentences may be compared. For example, the system control unit 20 may extract words from the text data by performing morphological analysis of each text data and compare the extracted words. Then, the system control unit 20 may determine that the text data match when all the words match, or determine that the text data match when the words match at a predetermined ratio or more. May be. In this way, the system control unit 20 compares the block configuration information of the content block correspondence information i with the block configuration information of all the content block correspondence information stored in the storage unit 15, and calculates the appearance frequency. .

  When calculating the appearance frequency, the system control unit 20 determines whether the calculated appearance frequency is equal to or less than the threshold stored in the storage unit 15 (step S67). At this time, if the appearance frequency is equal to or lower than the threshold (step S67: YES), the system control unit 20 determines that the content block with the block number i is one of the specific content blocks (step S68). . That is, the system control unit 20 adds the content block with the block number i to the content block specific to the blog page corresponding to the designated HTML document.

  The system control unit 20 adds 1 to the block number i (step S69) when the appearance frequency is greater than the threshold (step S67: NO) or when the process of step S68 is completed. It is determined whether i is larger than the number of blocks NUM (step S70). At this time, if the block number i is equal to or less than the value of the block number NUM (step S70: NO), the system control unit 20 proceeds to step S66. When the system control unit 20 calculates the appearance frequencies of all the content blocks extracted in the tree search process (step S70: YES), the system control unit 20 ends the specific content block determination process.

  The system control unit 20 has extracted the content block by the tree search process in step S64. However, in the one-page correspondence extraction process (step S4 in FIG. 7) executed from the advertisement content insertion process, the received blogger user Content blocks are extracted for all the blog pages corresponding to the ID, and as a result, the content block correspondence information is stored in the storage unit 15. Therefore, it is not necessary to extract the content blocks again. In that case, based on the URL of the designated HTML document, the content block correspondence information of each content block constituting the blog page to which the HTML document corresponds can be acquired from the storage unit 15.

[4. Modification 1]
Next, a modification of this embodiment will be described with reference to FIG.

  In the explanation so far, when the blogger is designated by the system administrator, the advertising content is inserted into the blog page of the designated blogger. However, even if the advertising content is inserted at the timing when the blog is updated. good.

  FIG. 11 is a flowchart illustrating a processing example in the blog update processing of the system control unit 20 of the blog server 1 according to the modification of the present embodiment. In FIG. 11, the same steps as those in FIG. 7 are given the same step numbers.

  First, prior to updating the blog, the blogger accesses the blog service site by operating the user terminal 3, and logs in to the blog service site by inputting his / her user ID and password. By this login, the blog server 1 issues a session ID to the user terminal 3, and manages the session ID and the user ID in association with each other. Since the request from the user terminal 3 to the blog server 1 includes the session ID, the blog server 1 can specify which blogger the request is from.

  Then, when the blogger performs a new blog article registration operation, the user terminal 3 transmits blog article data (text data such as titles and body text, image data, etc.) to the blog server 1, as shown in FIG. The system control unit 20 of the blog server 1 receives blog article data (step S71). Next, the system control unit 20 acquires an HTML document of the blog page to be updated from the blog page DB 101 from among the blog pages corresponding to the blogger user ID (step S72). Next, the system control unit 20 updates the acquired HTML document based on the received blog article data (step S73). For example, the system control unit 20 adds a TABLE tag or DIV tag for a blog article to the acquired HTML document, and adds the title of the received blog article, text data of the body text, etc. sandwiched between the tags. To do. Next, the system control unit 20 updates the HTML document registered in the blog page DB 101 with the HTML document to which the blog article data is added (step S74).

  Next, the system control unit 20 extracts content blocks from all blog pages corresponding to the blogger user ID (steps S3 to S6).

  Next, the system control unit 20 designates the HTML document updated in step S73, executes the specific content block determination process (step S8), and determines the blog page from each text data constituting the content block determined to be specific. Feature words are extracted (step S9).

  Next, the system control unit 20 deletes the existing advertisement content rule from the specified HTML document (step S75), and inserts the related advertisement content rule using the extracted feature word as a keyword (step S10). . That is, the system control unit 20 changes the advertising content displayed on the blog page.

  Then, the system control unit 20 updates the HTML document registered in the blog page DB 101 with the HTML document into which the advertisement content rule is inserted (step S11), and deletes all content block correspondence information from the storage unit 15. (Step S14).

  Note that the processing when a blog page has to be newly generated as the blog is updated may be basically the same as the processing described above. However, since the advertisement content has not yet been inserted in the newly generated blog page, the regulation of the advertisement content in step S75 is not performed.

[5. Modification 2]
In the description so far, one time is set as the threshold value used for determining the content specific to the blog page, but a value of two times or more may be set as the threshold value.

  For example, when the threshold is set to once, a content block (blog article) having an appearance frequency of once is extracted as content specific to the blog page, and feature words are extracted from the text data of the extracted blog article. The At this time, if the amount of text data of each extracted blog article is small, the number of words extracted from the text data decreases. If a sufficient number of words cannot be extracted, it may not be possible to determine at all which word is a feature word, or may not be able to accurately determine. Therefore, by increasing the threshold value and loosening the condition for determining content specific to the blog page, the number of blog articles that are the target of feature word extraction is increased. As a result, feature words can be extracted.

  Specifically, the system control unit 20 of the blog server 1 initially sets a threshold value to once, and determines a content block specific to the blog page, thereby extracting a blog article that appears once. To extract feature words. At this time, when it is determined that the feature word cannot be extracted, the system control unit 20 changes the threshold value to two times to extract the blog article and the feature word. If the system control unit 20 determines that the feature word cannot still be extracted, the system control unit 20 changes the threshold to three times to extract the blog article and the feature word. The system control unit 20 continues such processing until a feature word can be extracted. That is, the threshold value is raised when the processing based on the extraction result of the specific content block cannot be normally performed.

  However, if the threshold value is increased indefinitely, non-blog articles are also extracted. Therefore, the processing is interrupted when the threshold value is increased to some extent. For example, if the threshold value rises to the value of the number of pages of the blog page corresponding to the specified blogger, the content block used in common in each blog page is extracted, so the threshold value is the number of pages of the blog page. Processing may be interrupted when the value is reached.

  Further, for example, the system administrator may determine in advance that a content block that appears only once per a predetermined number of pages of the blog page is a content block specific to the blog page. In this case, the number of appearances as the threshold value may be changed in proportion to the number of blog pages corresponding to the designated blogger.

[6. Modification 3]
In the description so far, the number of appearances (frequency) is used as the appearance frequency used to determine the content specific to the blog page. However, the number of appearances for all content blocks of the blog page corresponding to the specified blogger. A ratio (relative frequency) may be used.

  For example, it is assumed that a comment can be registered from another user on a blog article registered by a blogger, and the comment can be viewed together with the blog article. The text data of this comment is also one of the contents constituting the blog page. When adding text data of a comment to a blog page, the system control unit 20 of the blog server 1 adds a description of a blocking tag to the HTML document of the blog page, and then adds the text data. Is a content block independent of the text data of blog articles and other comments. Then, the system control unit 20 extracts the text data of the comment as a content block, and when the extracted text data of the comment has a specific content, the advertising content related to the comment is inserted into the blog page To do.

  By the way, when a plurality of comments are registered for a certain blog article, the contents of the plurality of comments are not so much as contents that frequently appear, for example, opinions of majority and minority. It may be divided into contents that do not appear frequently. At this time, the opinions of the majority can be considered as general opinions and not very characteristic content. On the other hand, the opinions of minorities are unique opinions and can be considered as content specific to blog pages. In such a case, I want to extract comments that show minority opinions as content specific to the blog page.

  However, the number of majority opinions and the number of minority opinions are relative and vary with the total number of comments. In such a case, if the frequency is used as the appearance frequency and the threshold is set to, for example, once, contents that do not appear frequently (minority opinions) may not be appropriately extracted. Therefore, the relative frequency is used as the appearance frequency, and the threshold is set to a predetermined ratio. The threshold value at this time can be set arbitrarily. For example, when the content of the extracted content block is divided into N patterns (N is an integer of 2 or more), in order to distinguish minority opinions, a threshold value is set within a range of less than 1 / N. You may do it. As described above, the system control unit 20 may change the threshold according to the situation at that time.

  In addition, as a system capable of registering comments and the like for articles such as blogs, for example, other users can register a tweet that follows a tweet registered by a certain user. Twitter (trademark) and electronic bulletin boards.

  As described above, according to the present embodiment, the system control unit 20 of the blog server 1 extracts the contents constituting the blog pages that are sequentially designated by the designation of the HTML document, and is designated. The appearance frequency of each content constituting the blog page is calculated, and among the contents constituting the specified blog page, the content whose appearance frequency is equal to or less than a predetermined threshold is determined to be content specific to the blog page. .

  Therefore, since the content having a smaller appearance frequency is a content that does not appear much other than the designated blog page, all content satisfying the condition is determined by determining whether the appearance frequency is equal to or less than a threshold. Identified as content specific to the specified blog page. Therefore, content specific to the blog page can be easily extracted.

  Further, the system control unit 20 of the blog server 1 inserts advertisement content related to content specific to the designated blog page into the blog page.

  Therefore, information related to the characteristics of the blog page can be added to the Web page.

  In addition, when the system control unit 20 of the blog server 1 includes text data of a blog article as the content constituting the designated blog page, the text data is converted into the text data unique to the blog page. It is determined that the content is the content, the feature word of the blog page is extracted from the text data of the blog article, and the advertisement content associated in advance is inserted into the blog page using the feature word as a keyword.

  Therefore, an advertisement related to the content of the blog posted on the blog page can be added to the blog page.

  Further, the system control unit 20 of the blog server 1 calculates the appearance frequency of each content on a plurality of blog pages included in the blog service site.

  Therefore, the appearance frequency of each content constituting the specified blog page is calculated on a plurality of Web pages (for example, a plurality of blog pages corresponding to the specified blogger user ID) included in the blog service site. Therefore, it is possible to determine that the content used in common in the blog service site is not unique content, and the determination accuracy can be improved.

  Further, the system control unit 20 of the blog server 1 extracts the contents constituting the blog page in units of content blocks composed of one or more contents, and configures each designated blog page. The content block appearance frequency is calculated, and among the content blocks constituting the designated blog page, a content block whose appearance frequency is equal to or less than a threshold is determined to be a content block specific to the blog page.

  Therefore, on the blog page, for example, when one or more contents are collectively displayed as a content block, such as a header part, a navigation part, a part where a blog is displayed, a copyright display part, etc., A content block specific to a blog page can be extracted.

  Further, the system control unit 20 of the blog server 1 extracts the contents constituting the blog page based on the HTML document of the blog page, and determines the content block based on the DIV tag or the TABLE tag in the HTML document.

  Therefore, one or more contents explicitly blocked when creating an HTML document can be specified by the DIV tag, and one of the contents that is blocked and displayed in a table format by the TABLE tag. Since the above content can be specified, for example, when the content specific to the blog page and the non-specific content are blocked by these tags, the accuracy of determining the content specific to the web page Can be raised.

  In the above embodiment, text data and image data are extracted as the contents constituting the Web page, but the contents to be extracted are not limited to these. For example, it may be content displayed on a Web page or content that is played back when a Web page is displayed (for example, moving image data, audio data, electronic document, etc.). Further, only a predetermined type of content may be extracted.

  Moreover, in the said embodiment, each content block corresponding information which comprises all the blog pages corresponding to the user ID of the designated blogger for the content block corresponding information corresponding to each content block which comprises the designated blog page. Each appearance frequency was calculated by comparing with. In other words, when calculating the appearance frequency of each content block making up the specified blog page, the frequency of appearance in the range targeting all the blog pages corresponding to the specified blogger was calculated. However, the target range is not limited to this. For example, blog pages corresponding to a predetermined number of pages may be targeted, or all blog pages constituting the blog service site may be targeted.

  Moreover, in the said embodiment, although the content pinched | interposed into the DIV tag and the content pinched | interposed into the TABLE tag were extracted by grouping as a content block, as a tag which groups content, It is not limited only to these.

  In the above embodiment, content specific to a Web page is extracted in units of content blocks, but each content may be extracted one by one as it is.

  In the above embodiment, advertising content indicating an advertisement related to a product or service has been inserted into the Web page as content related to content unique to the Web page. It is not something that can be done. For example, image data (still image or moving image) related to content such as a blog article determined to be unique content may be inserted as a background image or an insertion image (illustration). Specifically, for example, a database for image data is constructed, and image data and keywords are associated and registered in the database. The keyword associated with the image data is a word indicating an image represented by the image data or a word related to the image. Then, feature words are extracted from the content determined to be unique content, and related image data is selected from the database using the extracted feature words as keywords. Then, the URL of the selected image data is inserted as a background attribute in the BODY tag of the target HTML document, or an IMG tag for displaying the selected image data is inserted at a predetermined position of the target HTML document. As a result, an image suitable for the content content such as a blog article determined to be unique content can be inserted into the Web page.

  Further, the use of content specific to a Web page is not limited to only inserting related content into a Web page. For example, new content may be generated based on content unique to the Web page.

  In the above embodiment, the specific content determination device of the present invention is applied to the server device. However, if the HTML document can be acquired from the storage means or the network, the special content is determined for the terminal device or the like. A determination device may be applied.

  In the above-described embodiment, the document data of the present invention is applied to an HTML document. However, data (for example, XHTML (Extensible HyperText Markup Language) that is described in a markup language and indicates content constituting a Web page. Document data may be applied to a document).

  Moreover, in the said embodiment, although the content which comprises the blog page in a blog service site was extracted, the kind of site and page made into object are not restricted only to these.

DESCRIPTION OF SYMBOLS 1 Blog server 2 Management terminal 3 User terminal 11 Operation part 12 Display part 13 Communication part 14 Drive part 15 Storage part 16 Input / output interface part 17 CPU
18 ROM
19 RAM
20 System Control Unit 21 System Bus 101 Blog Page DB
102 Advertising DB
NW Network S Blog System

Claims (11)

An extracting means for extracting content constituting a designated web page from a plurality of web pages included in a predetermined site;
Calculating means for counting the frequency with which each of the contents constituting the designated web page is used on another web page among the plurality of web pages;
A determination unit that determines that content that is used in other Web pages is less than or equal to a predetermined value among the contents that constitute the specified Web page is content specific to the specified Web page;
A unique content determination apparatus comprising: In the specific content determination apparatus according to claim 1,
The extraction means extracts the content constituting the web page in units of content groups composed of one or more contents,
The calculation means counts the frequency at which the content group constituting the designated web page is used on another web page,
The determination means is a content group specific to the designated web page, among the content groups constituting the designated web page, a content group whose frequency used in other web pages is a predetermined value or less. A unique content determination apparatus characterized by determining that In the specific content determination apparatus according to claim 2,
The specific content determination apparatus, wherein the extraction unit extracts a content group based on document data that is described in a predetermined markup language and indicates content that constitutes a Web page. In the specific content determination apparatus according to claim 3,
The specific content determination apparatus, wherein the extraction unit determines a content group based on a predetermined tag in document data indicating the content. In the specific content determination apparatus according to any one of claims 1 to 4,
The extraction means extracts a comment posted to the article from a web page on which the posted article is posted,
Classifying means for classifying each extracted comment according to the content indicated by the comment;
Setting means for setting a threshold of appearance frequency, and setting means for decreasing the threshold as the number of the contents into which the comments are classified increases.
Further comprising
The calculating means calculates the appearance frequency of each of the contents in which the comment is classified in a Web page,
The determination unit determines that the content whose appearance frequency calculated by the calculation unit is equal to or less than the set threshold is content specific to the Web page. An extracting step of extracting content constituting the designated web page among a plurality of web pages included in a predetermined site;
A calculation step of counting the frequency with which each content constituting the designated web page is used on another web page among the plurality of web pages;
A determination step of determining, among the contents constituting the specified web page, content whose frequency used in other web pages is a predetermined value or less as content specific to the designated web page;
A unique content determination method characterized by comprising: The specific content determination method according to claim 6,
The extraction step extracts a comment posted to the article from a web page where the posted article is posted,
A classification step of classifying each extracted comment according to the content indicated by the comment;
A setting step of setting a threshold of appearance frequency, a setting step of decreasing the threshold as the number of the contents into which the comment is classified is increased,
Further including
The calculation step calculates the appearance frequency of each content in the web page into which the comment is classified,
The determination step is characterized by determining that the content whose appearance frequency calculated by the calculation step is equal to or less than the set threshold is content specific to the Web page. Computer
Extracting means for extracting content constituting a designated web page from a plurality of web pages included in a predetermined site;
A calculating means for counting the frequency with which each content constituting the designated web page is used in another web page among the plurality of web pages; and
A determination unit that determines that content that is used in other Web pages and has a frequency equal to or lower than a predetermined value among the contents that constitute the specified Web page is content specific to the specified Web page;
A unique content determination program characterized in that it is made to function as: In the specific content determination program according to claim 8,
The extraction means extracts a comment posted to the article from a web page on which the posted article is posted,
The computer,
Classification means for classifying each extracted comment according to the content indicated by the comment, and
Setting means for setting a threshold of appearance frequency, and setting means for decreasing the threshold as the number of the contents into which the comment is classified increases.
Further function as
The calculating means calculates the appearance frequency of each of the contents in which the comment is classified in a Web page,
The determination unit determines that the content whose appearance frequency calculated by the calculation unit is equal to or less than the set threshold is content specific to the Web page. The specific content determination device according to any one of claims 1 to 5,
Insertion means for inserting related content related to content determined to be specific content by the specific content determination device into the specified web page;
A related content insertion device comprising: The related content insertion device according to claim 10,
When the specific content determination device includes text data of a posted article as content constituting the designated web page, the text data is content specific to the web page. Judgment,
Feature word extraction means for extracting feature words of the designated web page from text data of an article determined to be unique content by the unique content determination device;
Selecting means for selecting, as the related content, content related to the extracted feature word from a plurality of contents stored in the storage means in association with each word;
Further comprising
The related content insertion apparatus, wherein the insertion means inserts the selected related content into the designated Web page.

Images