JP6800478B2 - Evaluation program for component keywords that make up a Web page - Google Patents

Description

The present invention relates to an evaluation program for component keywords constituting a Web page, and specifically, a computer program for constructing a system for knowing user preference and the relationship between words determined by a search engine evolving by artificial intelligence. It is about.

When searching a Web page (Internet homepage, etc.), a search query (words and phrases to be input at the time of search) is input in a predetermined input area. When you search, the searched Web pages are displayed in order from the top. Sometimes search results are displayed over dozens of pages. Here, what is important for the side that creates the Web page is to create the Web page so that the Web page created by oneself is displayed as high as possible. This is because it is considered that the higher the page is, the more chances it will be accessed.

Previously, Web pages containing many words included in search queries were displayed in the top search results. Therefore, the Web page creator also creates a Web page by setting a target keyword and creating a sentence containing a large number of the target keywords.

However, recent search engines (for example, Google) have adopted artificial intelligence as shown in Non-Patent Document 1 below, and the simple method as described above cannot cope with it. This point can be roughly summarized in the following two points.

(1) First, as a result of improving the accuracy of learning the features of a good Web page from the user's operation information (browsing history), the algorithm has become different for each search query (keyword).

For example, if you do a Google search with the keyword "patent attorney Tokyo", the screen shown in FIG. 1 will be displayed. In the search result, geographical information (one of the Google services) is included like reference numeral A. This is the result of learning that geographical information is important in the case of the search query "Patent Attorney Tokyo" due to the artificial intelligence of the search engine.

In addition, news, maps, images, and shopping links are displayed in the search option menu indicated by reference numeral B in FIG. This display is also made as a result of judging that there are these needs due to the artificial intelligence of the search engine.

Therefore, if you want to display the Web page at the top of the search query "Attorney Tokyo", be sure to include the map (geographical information), and also the news (= fresh information = updateability) and images ( Creating content that describes information related to (= face photo, etc.) and shopping (= books, etc.) can be said to be the creation of a Web page that corresponds to the artificial intelligence of a search engine.

Next, when a Google search is performed using the keyword "patent attorney Osaka", a screen as shown in FIG. 2 is displayed. Here, the knowledge graph card is indicated by the reference numeral C. The Knowledge Graph is a search engine's effort to capture information not just as a character string but as a relationship in the real world, and to give meaning to keywords. In the above case, it can be seen that the artificial intelligence of the search engine was recognized by meaning that "Patent Attorneys Osaka" = "Japan Patent Attorneys Association Kinki Branch".

Therefore, if you want to display a Web page at the top of the search query "Patent Attorneys Osaka", you can refer to the Japan Patent Attorneys Association Kinki Branch and create content that describes news, shopping, and images. This means creating a Web page to support the artificial intelligence of.

Next, if you do a Google search with the keyword "Tsutenkaku", the screen shown in FIG. 3 will be displayed. Here, what is indicated by the reference numeral D includes image information (one of the Google services). Links to maps, images, news, and videos are displayed in the search option menu indicated by reference numeral B. The Knowledge Graph card indicated by the symbol C means that it is a tourist attraction, its location, business hours, and the like.

Therefore, if you want to display a web page at the top of the search query "Tsutenkaku", include a wealth of image information, mention maps and news, post videos, and include content that also mentions tourist information that is recognized as meaning. Producing is the creation of a Web page to support the artificial intelligence of search engines.

From the above results, the algorithm is different for each search query (keyword), and it is necessary to create a Web page corresponding to that point.

(2) The second point is that as a result of learning the context and meaning of sentences described on Web pages, we have come to learn about unknown words.

In order to support artificial intelligence algorithms, it is necessary to understand how search engines recognize the meaning of words. Prior to incorporating conventional artificial intelligence, search engines displayed Web pages containing many words contained in search queries entered by users as search results.

However, search engines that incorporate artificial intelligence implicitly infer the text of Web pages using a mathematical approach such as vector space and singular value decomposition. Then, when a search engine analyzes the meaning of a sentence on a Web page, it infers the meaning of a keyword (word) that one would like to express from the components of the word contained in the sentence.

On the other hand, the fact that the human side that creates a Web page is compatible with an artificial intelligence search engine means that the theme (= keyword, word) that is being expressed in the text of the Web page is factorized and the word that is the component (hereinafter, "" It is to write a sentence containing "component keyword". Therefore, the selection of words (component keywords) is the most important factor. The case will be described below.

FIG. 4 is a screen for displaying the search results of the search query “flour egg pork sauce iron plate”. That is, a Web page containing "Okonomiyaki" in the title is being searched. Further, FIG. 5 is a screen for displaying the search results of the search query “flour, egg, pork, leek”. That is, a Web page containing "Chijimi" in the title is being searched.

From the above, the artificial intelligence search engine determines that the word included in the search query has a high correlation with the target keyword (in the above case, "okonomiyaki" and "chijimi") pointed to by the search result. In other words, as a result of inferring the meaning of the keyword (word) that you would like to express from the components of the word contained in the text of the Web page with the target keyword as the theme, "flour egg pork sauce iron plate" indicates "okonomiyaki". It can be seen that "flour, egg, pork, and leek" are meant to indicate "chijimi" and are being recognized.

Therefore, if you want to display the Web page at the top with the target keyword "Okonomiyaki", the component keyword group "flour egg pork sauce iron plate" with high correlation is displayed at the top with the target keyword "Chijimi". It is important to write the text using the group "flour egg pork leek". Here, "flour egg pork" is common to okonomiyaki and chijimi, but "source" and "iron plate" are characteristic words that determine the target keyword "okonomiyaki", and "leek" is a characteristic word that determines the target keyword "chijimi". Words. Of the component keywords, these words are defined as "recollection words".

For example, even when a Google search is performed using only two keywords, "source" and "iron plate", a website related to "okonomiyaki" is displayed in the search results as shown in FIG. 4A, and the search is performed using only "leek". In this case as well, as shown in FIG. 5A, the website related to "Chijimi" is displayed in the search results.

Nihon Keizai Shimbun, "Google introduces artificial intelligence into search engines for several months" [electronic version dated October 28, 2015] [searched November 29, 2016], Internet (URL: http://www.nikkei) .com / article / DGXMZO93305300X21C15A0000000 /)

If you are an expert in Web page production, you can intuitively select such component keywords, but if you are not, it is difficult to select keywords, and you should do it by trial and error. It also takes a considerable amount of time. Moreover, even if the person is an expert, it is difficult to determine whether or not the selected component keyword is optimal.

The present invention has been made in view of the above circumstances, and the problem is to efficiently select and evaluate component keywords when creating a Web page corresponding to a search engine that evolves by artificial intelligence, and to create a Web page. It is to provide an evaluation program of ingredient keywords that can contribute to the quality improvement of.

In order to solve the above problems, the component keyword evaluation program according to the present invention is
Steps to accept input of target keywords for searching web pages,
A step to automatically extract words that should compose a Web page as component keywords based on the entered target keyword,
The step of calculating the degree of correlation between the extracted individual component keywords and the target keyword as the first evaluation value, and
A step of displaying a list of the extracted component keywords and the first evaluation value, and
Steps to select a specific ingredient keyword based on the list display, and
The step of calculating the degree of correlation between all the selected component keywords and the target keyword as the second evaluation value, and
The step of displaying the calculated second evaluation value on the screen,
Is characterized by having a computer execute the above.

The operation and effect of the evaluation program with such a configuration will be described below. First, it has a step of automatically extracting words that should constitute a Web page as component keywords based on the input target keyword. The target keyword is a keyword that a search engine wants to recognize, and is a keyword that is entered in a search query when a user searches a home page. Next, it has a step of automatically extracting words that should constitute a Web page as component keywords based on the input target keyword. The component keyword is a word used in a title or content (sentence such as a description) when creating a Web page, and a plurality (many) component keywords are automatically extracted. It should be noted that the concept of "word" includes not only "word" but also sentences and the like based on words and word combinations.

Next, the degree of correlation between the extracted individual component keywords and the target keyword is calculated as the first evaluation value, and the first evaluation value for each component keyword is displayed in a list. The higher the degree of correlation, the higher the possibility that the Web page will be displayed at the top when the user searches if the Web page is created so that the component keyword is included. The degree of correlation is calculated as the first evaluation value, and for example, an arithmetic expression can be set such that the higher the numerical value, the higher the degree of correlation. The user can check which component keyword has a high evaluation value by looking at the list display. The user can select a specific component keyword based on the list display. For example, a component keyword having a high first evaluation value can be appropriately selected. Alternatively, the software may be provided with an analysis function instead of being selected by the user, and may be automatically selected by the software by adopting artificial intelligence or the like. Next, the degree of correlation between all the selected component keywords and the target keyword is calculated as the second evaluation value and displayed on the screen. By selecting a component keyword so that the second evaluation value is as high as possible, a Web page is created using the selected component keyword. As a result, it becomes possible to efficiently select and evaluate component keywords when creating a Web page corresponding to a search engine that evolves by artificial intelligence, and to contribute to improving the quality of Web page production.

In the step of calculating the first evaluation value according to the present invention.
A step of searching a Web page using the component keywords and
It is preferable to execute the step of checking whether or not the target keyword exists in the title and description of the searched Web page.

When calculating the first evaluation value, the Web page is searched using the component keywords. When searching a Web page, a search engine compatible with artificial intelligence is used. The title of the searched Web page and its description are displayed in a list, and it is determined by whether or not there is a target keyword in the title and description. For example, the evaluation value can be determined by whether or not it exists in the title, whether or not it exists in the description, and whether or not it exists in both.

In the step of calculating the first evaluation value according to the present invention
An evaluation table is stored in advance so that the higher the searched Web page is, the higher the evaluation score is, and it is preferable to have a step of calculating using this evaluation table.

By using such an evaluation table, arithmetic processing can be easily performed.

In the step of calculating the second evaluation value according to the present invention
A step of searching a Web page using the selected component keyword group, and
It is preferable to execute the step of checking whether or not the target keyword exists in the title and description of the searched Web page.

A Web page is searched using the selected component keyword group. When searching a Web page, a search engine compatible with artificial intelligence is used. The title of the searched Web page and its description are displayed in a list, and it is determined by whether or not there is a target keyword in the title and description. For example, the second evaluation value can be determined depending on whether or not it exists in the title, whether or not it exists in the description, and whether or not it exists in both. Based on this second evaluation value, it can be determined whether or not the selected component keyword is eligible.

The step of automatically extracting the component keyword according to the present invention is
It is preferable to have a step of searching the target keyword from the dictionary database and extracting words from the sentences explaining the target keyword.

As the dictionary database, for example, Wikipedia can be used. This dictionary database is used to search for component keywords and extract words from the description of the component keywords. This makes it possible to automatically extract component keywords. Also, the user does not have to think about what the ingredient keyword should be. The dictionary database is not limited to the above, and other databases may be used. When searching, a search engine compatible with artificial intelligence is used.

The step of automatically extracting the component keyword according to the present invention is
It is preferable to have a step of extracting the suggested words when the target keyword is entered in the input area of the search query.

For example, if you enter the word "Tsutenkaku" in the input area, words such as "access," "takoyaki," "kushikatsu," etc. will be suggested. These words are considered to be related to the target keyword and can be automatically extracted as component keywords. This eliminates the need for the user to think about what the ingredient keywords should be.

The step of automatically extracting the component keyword according to the present invention is
Steps to search a web page using the target keyword,
It is preferable to have a step of crawling the searched upper Web page and extracting words from the sentences constituting the Web page.

When a Web page is searched using the target keyword, the searched Web pages are displayed in order. When searching, a search engine that supports artificial intelligence is used. Next, the searched high-ranking Web pages, for example, the 1st to 10th Web pages are crawled, and words are extracted from the sentences constituting the Web page. Such words are considered to be related to the target keyword and can be automatically extracted as component keywords. This eliminates the need for the user to think about what the ingredient keywords should be.

Although three steps are given as the steps for automatically extracting the component keywords, all of them may be used, one or two as appropriate may be used, and another step is added for automatic extraction. You may do so.

A diagram showing a screen showing search results for the search query "Patent Attorney Tokyo" A diagram showing a screen showing search results for the search query "Patent Attorney Osaka" A diagram showing a screen showing the search results of the search query "Tsutenkaku" The figure which shows the screen which displays the search result in the search query "flour egg pork sauce iron plate" The figure which shows the screen which displays the search result in the search query "source iron plate" Diagram showing a screen showing search results for the search query "flour, egg, pork, leek" Diagram showing a screen showing the search results for the search query "Chive" Block diagram showing the overall system configuration using the evaluation program Flowchart showing the procedure when using the evaluation program Flowchart showing the procedure when using the evaluation program Flowchart showing the procedure when using the evaluation program Diagram showing the input screen of the target keyword Diagram showing an example of a suggestion The figure which shows an example of the score master table (evaluation table) Diagram showing an example of calculation using the score master table Diagram showing an example of calculation using the score master table The figure which shows the screen display example of the judgment result of a component keyword candidate Diagram showing the input screen of the selected component keyword The figure which shows the example which automatically judged the recall word candidate from the component keywords and identified and displayed it. A diagram showing a screen display example of each element of the selected component keyword candidate and the target keyword.

A preferred embodiment of the component keyword evaluation program according to the present invention will be described with reference to the drawings. FIG. 6 is a block diagram showing a configuration of the entire system using the evaluation program according to the present invention.

<Overall system configuration>
In FIG. 6, various programs (computer software) are stored in the Web server 100. The keyword explorer 101 (corresponding to the evaluation program according to the present invention) includes at least a Google information acquisition system 101a, a Wikipedia acquisition system 101b, a site analysis system 101c, and a component keyword determination system 101d. The functions of each system will be described later.

Further, the Web server 100 stores a member management system 102, and at least includes an authentication system 102a and an external connection management system 102b. Since the evaluation program according to the present invention can be used only by a person who has registered as a member, an authentication system 102a for confirming whether or not the member is a member is provided. The external connection management system 102b is a system for storing and managing member data and the like in an external server.

The DB server 200 is a server provided with a database for storing various data, and operates in cooperation with the Web server 100. The crawl server 210 stores the crawl system. There are various websites on the Internet. The user registered as a member accesses the Web server 100 via the Internet using, for example, a client browser (for example, Internet Explorer) stored in a PC (personal computer), and the evaluation program according to the present invention. Can be used.

The Google API search service is stored in the server 240, and provides a search service using a search engine provided by Google. In the present embodiment, Google will be used as an example as a search engine, but the present invention is not limited to Google, and when another search engine compatible with artificial intelligence is used, the search engine is not limited to Google. Also applies.

Wikipedia is a dictionary database and is stored in the server 250. The dictionary database is not limited to Wikipedia, and other databases may be used. Moreover, you may use a plurality of dictionary databases.

<Procedure for using the evaluation program>
Next, the procedure when using the evaluation program according to the present invention will be described with reference to the flowcharts of FIGS. 7A, 7B, and 7C.

A user who intends to use the production of a Web page using this evaluation program accesses the Web server 100 via the Internet using a PC or the like and opens the keyword explorer 101. When using it, member authentication is performed by the member management system 102, but since it is a known system, the description thereof will be omitted.

A screen as shown in FIG. 8 is displayed on the user's computer screen. For convenience of explanation, only a part of the screen configuration is outlined. Hereinafter, the same applies to other screen displays.

Enter the target keyword of the website to be produced in the input area 10 (S1). The target keyword is the theme of the website and is the keyword that the search engine wants to recognize. The input target keyword is stored and registered in the storage device as a database (FIG. 7B).

If there is a separate keyword list, read it (S2). The keyword list accesses the DB server 200 by pressing the click button 11 and reads the keyword list separately (S2a). Note that this step S2 is not necessary.

Next, the Google service (map, news, etc.) included in the search result by Google is automatically acquired (S3). When retrieved, it is defined as an "important content element" and stored in the database. This is illustrated by A in FIG. In this Google service, Google's artificial intelligence reflects the needs of users from the input search query, and originally, the display area of the website ranked first is cut off and displayed. Therefore, it can be determined that the content element is regarded as important by the search engine. This step S3 is executed by the Google information acquisition system 101a.

Next, the search option menu item of the search result by Google is automatically acquired (S4). The acquired item is defined as a "content element" and is stored in the database as an index for creating content according to this. The search option menu item is an item displayed in the area indicated by reference numeral B in FIG. The artificial intelligence of the search engine works here as well, and the results of determining that there are these needs are displayed. This step S4 is executed by the Google information acquisition system 101a.

Next, the knowledge graph card of the search result by Google is automatically acquired (S5). This is illustrated by reference numeral C in FIG. When the Knowledge Graph card is displayed in the Google search results, it indicates that it is functioning as semantic information pointing to the entered search query. This information is classified as "semantic elements" and stored in the database as an index for creating content corresponding to this. This step S5 is finally executed by the Google information acquisition system 101a.

Next, words are automatically extracted from Wikipedia, which is the target keyword (S6). Google has announced that it uses Wikipedia's database, and words extracted from Wikipedia can be candidates for component keywords with high probability. These words are stored in the database as component keyword candidates (S6a). This step S6 is executed by the Wikipedia acquisition system 101b. The Wikipedia acquisition system 101b accesses database 250 to acquire words.

When extracting words, it is easy if it is an English sentence because it is divided by word, but if it is a Japanese sentence, it is not divided by word, so it is easy to extract from the sentence using morphological analysis technology. Break down words. The morphological analysis is a known analysis method.

Next, a word is automatically extracted from a search query (Google suggestion) including a target keyword (S7). FIG. 9 shows an example in which "Shin-Osaka" is entered in the search query, and the words "station", "lunch", "hotel", etc. are suggested below it in a list display. Google suggestion is a function that displays candidates when a word is entered on the Google search screen. Since these words are the user's needs themselves and can be candidates for component keywords with high probability, these words are stored in the database as component keyword candidates (S6a). This step S7 is executed by the Google information acquisition system 101a.

Next, the websites at the top of the Google search results of the target keyword are crawled, and words and feature words are automatically extracted (S8). It is considered that the words used on the higher-ranking websites are words that Google has evaluated to a certain degree as component keywords, and can be candidates for component keywords with high probability. Therefore, this information is acquired as a component keyword candidate from the website on the Internet and stored in the database (S6a). The method of decomposing words from sentences uses the above-mentioned morphological analysis technique. This step S8 is performed by the crawl system on the crawl server 210.

On the other hand, Google demands originality in the production of websites, and it is not preferable to use only the words of websites that are generally displayed at the top. Therefore, a feature word is acquired for each page in each of the top-displayed websites, and what kind of characterization of the word in the website is automatically calculated. Instead of constructing similar feature words, the website to be created uses it as an index for composing words as unique feature words. Feature words are automatically acquired using the TF-IDF method. The acquired feature words are stored in the database (S8a). The TF-IDF method is a known method and is an index of features widely used in information retrieval and document recommendation. The acquisition of this feature word is executed by the function of the site analysis system 101c.

Next, the component keyword candidates saved in the database in step S6a are automatically determined by themselves (S9). This step 9 is executed by the component keyword determination system 101d of the keyword explorer 101. To specifically explain the logic used at the time of this determination, a Google search is performed using the component keyword as a search query. It is determined whether or not the target keyword is included in the title and description of the website given as the search result.

The title and description of the website obtained by the search are direct information that can be visually recognized by the user. Whether or not the target keyword is included here is a very important factor in determining whether or not the information is the information that the user wants to obtain. It is thought that Google's artificial intelligence also attaches great importance to the information displayed in the title and description.

FIG. 10 is a diagram showing a score master table (an example of an evaluation table) used in the determination logic. The ranking of the search results is evaluated from 1st to 10th. However, the number of evaluation targets can be set as appropriate. The score is shown when the target keyword is present in the title, in the description, and in both. If it exists in both, the score set separately is applied instead of adding the score if it exists in the title and the score if it exists in the description.

As a matter of course, the higher the website, the higher the score. In addition, the points are not simply added, but the highest point is 100 points, so the maximum score (upper limit) of the total points from 1st to 3rd place is 40 points, and the total points from 4th to 7th place. The maximum score is 35 points, and the total score of 8th to 10th is set to 25 points. The specific score setting is not limited to this, and can be set as appropriate.

FIG. 11 is a diagram showing an example of calculation using the score master table of FIG. The table on the left shows an example of search results, in which the target keyword exists in the title of the website in the 5th place and the description exists in the 3rd, 5th, and 10th places. Specific examples of points are shown in the table on the right. The 5th place on the website is 25 points because it exists in both the title and the description. The total of 1st to 3rd place is 18 points, the total of 4th to 7th place is 25 points, the total of 8th to 10th place is 9 points, and the total is 52 points. Since none of these total points have reached the perfect score (upper limit value), each score is applied, and the final evaluation score (corresponding to the first evaluation value) is also 52 points.

FIG. 12 is a diagram showing another calculation example. Google search for component keyword candidates in the same way, and the target keywords are included in the titles of the 1st, 2nd, 3rd, 5th, 7th, and 9th websites, and the 1st, 3rd, 5th, and so on. This is the case when the target keyword exists in the explanations of the 6th, 9th, and 10th places. The total score of the 1st to 3rd places is 83 points, but since it exceeds the perfect score, it is 40 points. The total score of the 4th to 7th places is 58 points, but since it exceeds the perfect score, it is 58 points. The total score of 8th to 10th place is 29 points, but since it exceeds the perfect score, it is 25 points. The final first evaluation value is 100 points.

It can be said that the closer the first evaluation value is to 100 points, the higher the degree of correlation between the component keyword and the target keyword. The determination result of the component keyword alone is stored in the database (S9a).

Next, the determination result of each component keyword candidate is displayed on the screen (S10). An example of the screen display is shown in FIG. In this example, the target keyword is "Shinjuku". Under the item of related keywords, "movie", "lunch", "ramen", etc. are shown as candidates for component keywords. The first evaluation value is shown as a score under the item called Brain Score. This is the determination result using the score master table of FIG. On the far right of the list display, there is a selection button for each component keyword.

The user can select the one with the highest first evaluation value by clicking the selection button while looking at the determination result. For example, "beer garden", "movie", and "cafe" with high evaluation values can be selected. Alternatively, the software may be provided with an analysis function instead of being selected by the user, and may be automatically selected by the software by adopting artificial intelligence or the like.

Although the flowchart of FIG. 7B does not automatically determine the feature word in step S9, the automatically extracted feature word may also be automatically determined in step S9.

Next, the selected component keyword candidate is Google-searched and automatically determined (S11). FIG. 14 is an input screen for the selected component keyword. Here, two of "women's association" and "LUMINE" are selected as the component keywords (group). The target keyword is "Shinjuku". It is automatically input by clicking the selection button in FIG. Alternatively, the user may be able to input by himself / herself. The logic of automatic determination is the same as in step S10, and the same score master table is used. In step S10, the evaluation was for a single component keyword (one), but in step S11, the evaluation is for a plurality of component keywords (composite words). When the "examine" button is clicked in FIG. 14, the determination logic operates, and the relevance evaluation is calculated and displayed (S12). In FIG. 14, 100 points (perfect score) are shown as the evaluation value (corresponding to the second evaluation value). The determination result is saved in the database (S11a).

FIG. 14A shows an example in which candidates for characteristic words (remembering words) that determine the target keyword are automatically determined and identified from the component keywords. As a method for identifying and displaying, various methods such as attaching a colored marker, changing the font and size, and the like can be considered. In this case, it is automatically determined that "Chive" is a recall word for the target keyword "Chijimi". Regarding the determination of the recall word, for example, the feature word extracted in step S8 can be automatically used as the recall word. Alternatively, as a result of the automatic determination in step S9, the component keyword having a high first evaluation value can be used as the recall word.

The second evaluation value can be calculated as many times as necessary while changing the selection of the component keyword, whereby an appropriate component keyword can be selected.

Component keyword candidates with a low second evaluation value are automatically or selected and excluded. For example, a component keyword having a low evaluation value in FIG. 14 can be selectively excluded from the list display in FIG. 13 by the user, or automatically excluded (process to be hidden). .. Further, when displaying the list in FIG. 13, if the brain score (first evaluation value) is 0 points, processing may be performed so that the brain score is not displayed from the beginning.

In order to raise the evaluation by the artificial intelligence installed in Google, it is important to adopt a word that is recognized as a component keyword and has a high numerical value indicated by the TF-IDF method as a characteristic word. Therefore, the information that satisfies it is displayed and selected or excluded.

Next, each element of the selected component keyword candidate and the target keyword is displayed (S13). An example of the screen display is shown in FIG. Each element of the target keyword is an "important content element", a "content element", and a "knowledge graph card" displayed on FIG. The result is saved in the database (S13a).

As the final judgment result, the elements of the content creation method reflecting the user preference of the target keyword, "important content", "content", and "semantic element (with Knowledge Graph card)" are displayed. In addition, the automatic determination result (first evaluation value) of each component keyword, "important content", "content", and "semantic element (with knowledge graph card)" are displayed, and the result list can be saved.

The evaluation results performed in the past can be displayed and compared. In addition, the result list can be downloaded and used for creating a website.

The website creator will be able to create a website to support Google's artificial intelligence by proceeding with the website creation based on the information obtained from this system.

<Another Embodiment>
Words in the present invention are not limited to nouns, but also include adjectives and verbs. Further, the "word" in the present invention includes a phrase, a sentence, etc. in which a plurality of words, particles, etc. are combined.

The component keyword in the present invention is a concept including a component keyword finally selected by the user and a candidate (component keyword candidate) until the selection.

As the hardware on which the database is actually constructed, an appropriate storage device such as a hard disk, a semiconductor memory, or an external storage medium (DVD or the like) can be used.

In the present embodiment, the first evaluation value and the second evaluation value are evaluated according to the magnitude of the numerical values, but the evaluation is not limited to this, and evaluations such as A, B, and C may be used. .. Further, when a numerical value is used, the smaller the numerical value, the higher the evaluation may be set.

In the present embodiment, the evaluation program is installed on the Web server, but it may be installed on each user's PC.

10 Input area 11 Click button 100 Web server 101 Keyword explorer 101a Google information acquisition system 101b Wikipedia acquisition system 101c Site analysis system 101d Component keyword judgment system 102 Member management system 200 DB server 210 Crawl server

Claims (7)

Steps to accept input of target keywords for searching web pages,
A step to automatically extract component keywords as words that should compose a Web page based on the entered target keywords,
The step of calculating the degree of correlation between the extracted individual component keywords and the target keyword as the first evaluation value, and
A step of displaying a list of the extracted component keywords and the first evaluation value, and
Steps to select a specific ingredient keyword based on the list display, and
The step of calculating the degree of correlation between all the selected component keywords and the target keyword as the second evaluation value, and
The step of displaying the calculated second evaluation value on the screen,
An evaluation program for component keywords that make up a Web page, which is characterized by having a computer execute the above.

In the step of calculating the first evaluation value,
A step of searching a Web page using the component keywords and
The evaluation program according to claim 1, wherein the step of checking whether or not the target keyword exists in the title and description of the searched Web page and the execution of the step. In the step of calculating the first evaluation value,
The evaluation program according to claim 2, wherein an evaluation table is stored in advance so that the higher the searched Web page is, the higher the evaluation score is, and the evaluation table has a step of calculating using the evaluation table. .. In the step of calculating the second evaluation value,
A step of searching a Web page using the selected component keyword group, and
The evaluation program according to claim 2, wherein the step of checking whether or not the target keyword exists in the title and description of the searched Web page and the execution of the step. The step of automatically extracting the component keywords is
The evaluation program according to any one of claims 1 to 4, wherein the evaluation program includes a step of searching the target keyword from a dictionary database and extracting words from a sentence explaining the target keyword. The step of automatically extracting the component keywords is
The evaluation program according to any one of claims 1 to 5, wherein the evaluation program includes a step of extracting words to be suggested when a target keyword is input in an input area of a search query. The step of automatically extracting the component keywords is
Steps to search a web page using the target keyword,
The evaluation program according to any one of claims 1 to 6, wherein the evaluation program includes a step of crawling a searched upper Web page and extracting words from sentences constituting the Web page.