JP2003006221A - Predictive analysis type retrieval system, predictive analysis type retrieval method, and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technology by which the information a user desires to have can be accurately collected from flood of information on the Internet. SOLUTION: A predictive analysis type retrieval system is provided with a keyword creation means which creates keywords for contents retrieval on the Internet and scores the importance of keywords, a contents evaluation means which scores the text data in the contents in accordance with the keyword or the like, and a URL collection means which collects URL where user's desiring contents are predicted to appear by using the importance of keywords scored by the above keyword creation means and the contents scored by the contents evaluation means.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for efficiently collecting desired information from the flood of information on the Internet and URLs carrying the information.

[0002]

2. Description of the Related Art In recent years, the number of users of the Internet has increased rapidly, and the number of pages that can be browsed has also increased daily. In order to access the information desired by Internet users, information retrieval must first be performed. Generally, you go to a search site and enter keywords relevant to the information you want. In the search site where the keyword is input, a direct and simple method of extracting the site including the keyword as text data is adopted in the main search sites.

Methods for extracting a site containing a keyword are classified into "registration system" and "robot type". The registration system is a method in which a home page creator is registered in a search site or staff of the search site is registered and a site including a keyword is extracted from the registered sites. The robot type is a method in which a robot automatically circulates websites around the world, registers it in a database of search sites, and extracts sites containing keywords based on the registration. In the robot type, when the visited website is a website with a high update frequency, when a keyword is input, a site with a high update frequency is preferentially extracted.

[0004]

The information available on the Internet is cobblestone mixed. For this reason,
There are various innovations in the method of selecting information on "balls". For example, the robot type in which the above-mentioned "Websites with high update frequency" is preferentially extracted is
It can be said to be a search site in which "tama" = "fresh" is replaced. However, fresh information is not always what the search site user wants. Then, the subject which this invention tends to solve is providing the technique which can collect the information which a user wants exactly from the information which rebels on the internet.

[0005]

The present invention achieves the above-mentioned object by setting it in the server of the operator of the search site and utilizing it by the user of the search site.

(Claim 1) The invention according to claim 1 creates a keyword for content search on the Internet, and a keyword creating means having a function of scoring the importance of the keyword, Content evaluation means for scoring the text data according to a keyword or the like, and evaluating the content in the search result obtained based on the keyword created by the keyword creating means by the content evaluating means. Along with scoring, the scored contents are compared with the keywords to score the importance of the keywords, and the keyword creating means is used to score the keywords to be used in the next search based on the importance of the keywords. Created by updating the keywords used for searching. According to predictive analysis-type search system according to claim.

(Explanation of Terms) “Keyword” may be designated by the user, or may be reset by the operator of the search site with reference to the designation by the user. As a method of "scoring" referred to as the keyword creating means, for example, weighting according to the context of the keyword itself or the sentence containing the keyword is performed. Further, since the error is large if the number of appearances is added as it is, a method of correcting and adding the number of appearances is also adopted. As the “scoring” method referred to as the content evaluation means, a plurality of scoring functions are prepared in advance, an appropriate scoring function is selected according to the user's search purpose, and the scoring function is used for the calculation. As the scoring function, for example, when a user is a corporation and wants to collect a rumor that might affect the stock price of the corporation, a high score is given to the content judged to correspond to the "rumor". The following function is adopted.

(Operation) The keyword creating means creates a keyword for content search on the Internet. On the other hand, the content evaluation means scores the text data in the content on the Internet according to the keyword or the like. The content scored by the content evaluation means is scored by the keyword creation means. The keyword used for the next search is updated by reflecting the scored importance. By repeating this, the keywords are optimized, and useful contents according to the user's preference can be collected.

(Claim 2) According to a second aspect of the invention, a keyword creating means for creating a keyword for content search on the Internet and scoring the importance of the keyword, and text data in the content as a keyword Content evaluation means for scoring according to the
A URL that collects a URL at which it can be predicted that the content desired by the user will appear using the importance of the keyword scored by the keyword creation means and the content scored by the content evaluation means.
The present invention relates to a predictive analysis type search system including a collecting means.

(Operation) The keyword creating means creates a keyword for content search on the Internet. On the other hand, the content evaluation means scores the text data in the content on the Internet according to the keyword or the like. The content scored by the content evaluation means is scored by the keyword creation means. The keyword used for the next search is updated by reflecting the scored importance. Since the degree of importance scored by the keyword creating means is used, the keyword sufficiently reflects the user's preference, and the URL collecting means collects URLs at which it can be predicted that the content desired by the user will appear. it can. The collected URLs are predicted to have a high degree of importance, and the user can obtain a URL where useful contents can be obtained without performing a search by himself.

(Claim 3) The invention according to claim 3 limits the predictive analysis type search system according to claim 1 or 2, and the content evaluation means includes a tag for analyzing tag information in the content. The present invention relates to a predictive analysis type search system equipped with information analysis means.

(Explanation of terms) "Tag information" means, for example,
When the home page is created using the HTML language, it is the information obtained from the tag. The useful tag information is, for example, a link destination.

(Operation) Since the content evaluation means is provided with the tag information analysis means, the tag information can be used as one of the evaluation criteria for evaluating the content.

(Claim 4) The invention according to claim 4 limits the predictive analysis type search system according to any one of claims 1 to 3 and the keyword creating means is a keyword created by a user. The present invention relates to a predictive analysis type search system that includes an input receiving unit that receives an input and that the keyword creating unit uses the keyword received by the input receiving unit to re-score.

(Operation) Since the input receiving means is provided,
It is possible to deal with addition and modification of keywords by the user.

(Claim 5) The invention according to claim 5 limits the predictive analysis type search system according to any one of claims 1 to 4, and the keyword creating means includes:
The present invention relates to a predictive analysis type search system having a function of extracting a new wording or a context from the collected information and automatically generating a new keyword based on the evaluation result of the content evaluation means. (Operation) When the evaluation result by the content evaluation means is high, a new keyword or a new context is extracted from the information to newly generate the keyword. As a result, new keywords can be generated based on the content the user has shown interest in, and the keywords can be optimized. (Claim 6) The invention according to claim 6 limits the predictive analysis type search system according to any one of claims 1 to 5, and the content scored by the content evaluation means has a high score. The present invention relates to a predictive analysis type search system that includes a sorting unit that sorts in order, and that can output the sorting result by the sorting unit to a user interface that provides a search result to a user.

(Operation) The content scored by the content evaluation means is sorted by the sorting means in the order of high score. Then, the sorting result by the sorting means can be output to the user interface for the user to obtain the search result. Therefore, the user may browse in order from the content with the highest score, which is efficient.

(Claim 7) The invention according to claim 7 creates a keyword for content search on the Internet, and a keyword creating procedure having a function of scoring the importance of the keyword, Content evaluation procedure for scoring the text data according to a keyword or the like, and evaluating the content in the search results obtained based on the keyword created by the keyword creation procedure by the content evaluation procedure. Along with scoring, the scored content and the keyword are compared to score the importance of the keyword, and the keyword creation procedure is based on the importance of the scored keyword used in the next search. Created by updating the keywords used for searching. According to predictive analysis-first search method characterized.

(Claim 8) According to the invention of claim 8, a keyword creating procedure for creating a keyword for content search on the Internet and scoring the importance of the keyword, and the text data in the content as a keyword Content evaluation procedure for scoring according to
A URL that collects a URL that can predict that the content desired by the user will appear using the importance of the keyword scored by the keyword creation procedure and the content scored by the content evaluation procedure.
And a predictive analysis type search method including a collection procedure.

(Claim 9) The invention according to claim 9 limits the predictive analysis type search method according to claim 7 or 8, and the content evaluation procedure includes a tag for analyzing tag information in the content. The present invention relates to a predictive analysis type search method including an information analysis procedure.

(Claim 10) The invention according to claim 10 is
The predictive analysis type search system according to any one of claims 7 to 9 is limited, and the keyword creation procedure includes an input acceptance procedure for accepting a keyword input by a user and is accepted by the input acceptance procedure. The present invention relates to a predictive analysis type search method including a rescoring procedure for rescoring using a keyword.

(Claim 11) The invention according to claim 11 is
The predictive analysis type search method according to any one of claims 7 to 10 is limited, and the keyword creation procedure is based on the evaluation result of the content evaluation procedure, and a new wording is selected from the collected information. The present invention relates to a predictive analysis type search method characterized by including a procedure for automatically generating a new keyword by extracting a keyword or a context. (Claim 12) The invention according to claim 12 limits the predictive analysis type search method according to any one of claims 7 to 11, and gives a high score to the content scored by the content evaluation procedure. The present invention relates to a predictive analysis type search method including a sorting procedure for sequentially sorting and a user output procedure that enables a sorting result by the sorting procedure to be output to a user.

(Claim 13) The invention according to claim 13 is
A program for realizing a predictive analysis type search method for a computer, which program has a function of creating a keyword for content search on the Internet and scoring the importance of the keyword. A keyword creation procedure and a content evaluation procedure for scoring text data in the content according to a keyword etc. are provided, and the content in the search result obtained based on the keyword created by the keyword creation procedure is The content is evaluated and scored, and the scored content is compared with the keywords to score the importance of the keywords, and the keyword creation procedure scores the keywords to be used in the next search. Created based on the importance of the keywords According to a program for executing that it will update the keywords used more during the search on the computer.

(Claim 14) The invention according to claim 14 is
A program for realizing a predictive analysis type search method for a computer, the program creating a keyword for content search on the Internet, and a keyword creating procedure for scoring the importance of the keyword, The user wants to use the content evaluation procedure for scoring the text data in the content according to the keywords, etc., the importance of the keywords scored by the keyword creation procedure, and the content scored by the content evaluation procedure. A program for causing a computer to execute a URL collection procedure for collecting a URL at which content can be expected to appear.

(Claim 15) The invention according to claim 15 is
The program according to claim 13 or 14, wherein the content evaluation procedure includes a tag information analysis procedure for analyzing tag information in the content.

(Claim 16) The invention according to claim 16 is
The keyword creation procedure comprises an input acceptance procedure for accepting a keyword input by a user, and a rescoring procedure for rescoring using a keyword accepted by the input acceptance procedure. Related to the program described in.

(Claim 17) The invention according to claim 17 is
14. The keyword creating procedure includes a procedure of extracting a new wording or context from the collected information based on the evaluation result of the content evaluation procedure and automatically generating a new keyword. To the program according to claim 16. (Claim 18) The invention according to claim 18 includes a sorting procedure for sorting the contents scored by the content evaluation procedure in the order of high scores, and a user output procedure for outputting the sorting result by the sorting procedure to the user. The program according to any one of claims 13 to 17 provided.

In the thirteenth to eighteenth inventions, the program can be stored in a recording medium and provided. Here, the "recording medium" is a medium that can carry a program that cannot occupy space by itself, and is, for example, a flexible disk, a hard disk, a CD-ROM, an MO (magneto-optical disk), or a D.
For example, VD-ROM.

[0029]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will now be described in more detail with reference to the embodiments and the drawings. The drawings used here are FIGS. 1 to 9. The “user” refers to a person (may be a corporation) who uses the search system according to the present embodiment.

(Overall Configuration) The overall configuration of the search system according to the embodiment operates in the background and acquires information from the Internet by using "SEIP (Search Engine with Intelligent Prediction).
s) ”and retrieves the acquired information from the database and
"CGI for displaying results" provided to users through eb browser
It is composed of two basic systems called "program".

The predictive analysis type search engine examines the contents of the contents on the Internet extracted according to the keyword and makes a score, and "contents evaluation means", and analyzes the user's preference, and searches for keywords that are the seeds of the search. A "keyword creating means" for creating and scoring the importance of the keyword, and a "URL collecting means" for collecting a URL that can predict that the content desired by the user will appear by analyzing the search result and the user's preference. "
It has and. The keyword creation means also has a function to automatically generate a new keyword based on the information of the content the user has shown interest in, based on the wording or context used in the information, based on the result of the content evaluation means. I have it.

(Predictive analysis type search engine) The process until the predictive analysis type search engine acquires information from the Internet and stores it in the database is the first process of acquiring a specified page given by the URL from the Internet,
It is roughly divided into a second process of analyzing the content of the acquired page, and a third process of storing the information in the database if the analysis result is the target information.

(Each function) Further, it has the following six functions, and these processing units are described in an interpreter language (Perl). First, the URL to be searched
Is a URL database processing unit (URL collecting unit) that adds, deletes, and updates URL information. Secondly, it is a page acquisition unit that acquires header information, checks the page update time, and acquires the page body. Thirdly, it is a tag information analysis means for removing tags and acquiring link information. Fourth, a page content analysis unit (content evaluation unit) that performs pattern matching processing using keywords and contexts and scores the pages acquired based on the result of the processing. Fifth, it is a keyword creating means for creating or extracting a new keyword based on content evaluation, or re-evaluating a keyword to review the score. Sixth is a database processing means for adding or deleting the acquired pages to the database and detecting the same message in the database.

(URL Database Processing Means) The predictive analysis type search system of this embodiment collects URLs by following the links of pages and examines the relevance of the information specified by the URLs. To expand. Expanding the search range based on the relevance of information means examining the preference level, and at the same time, determining the UR.
It also calculates the "expectation of information" that is updated in L. URLs are collected according to the degree of expectation. U collected
A new search will be performed based on the RL. The degree of expectation of information is calculated by "page content analysis means". By this means,
Based on this expectation, collection and deletion of URLs and ordering are performed as follows.

(Description of associative array) A method called associative array is adopted for managing URLs. FIG. 2 shows the structure of the associative array. In this embodiment, a dbm database that has been used as a standard for a long time in the UNIX (registered trademark) world was used. The URL is substituted for "key", and the page score and the number of searches are substituted for "value". When using an associative array, only one value can be stored for one key. However, since it is necessary to store the information on the score and the number of searches for one URL, two values are substituted with ":" as a delimiter. For example, when the value value = 12.5: 5, the page score is 12.5 and the number of searches is 5.

(Addition / deletion of URL) When a URL is newly added, the URL is set in the key and the value value is set to 0: 0. Deleting the registered URL is the valu
It was done only when e satisfied a certain condition. In the current version, when the number of searches is 10 or more and the score is 0, the deletion condition is set. However, this condition is provisionally set as an initial setting and can be changed.

(Update of URL information) When a search is performed, a score is obtained for the URL. It is necessary to add this score as URL information. Also, the numerical value indicating the number of searches must be updated. In that case, the URL
Use as a key and change the value of that value. (Example) When the obtained score is 10 points and the value value is 0: 4 In this case, the search has been performed 4 times so far, and in each case, the score is 0 point. But this time 10
Since the points were obtained, the updated value value = 10:
It becomes 5.

(Page Acquisition Unit) FIG. 3 shows the concept of the page acquisition unit. The page acquisition unit acquires information on a page designated by the URL. In addition, at the time of acquisition, the page acquisition unit checks the update date and time of the page and acquires only the information updated within a predetermined period (for example, the past month). That is, the expiration date of information is assumed.
Usually, in information, one value is "freshness". Except for searches as materials, the value of information that has passed a certain period of time is often zero. Therefore, in this system, we focused on the effective period of information and set the effective period. The expiration date can be set arbitrarily.

Specifically, the update date and time is checked based on the header information obtained by the page obtaining means. Last-Modified in the header indicates the date and time when the page was last updated, and is used. Within the predetermined period, the entire page is taken in. This shortens the search time and the page acquisition time. Note that there is a case where the update date and time is not described in the header. In that case, whether the page can be acquired by reading the first fixed number of characters (several hundred bytes to several kilobytes) of the text described in HTML, searching for the item that describes the update date and time, and estimating the update date and time. Will be decided. This is because it is empirically understood that date / time data is often included in the writing portion of the document data.

(HTML Analysis Means) FIG. 4 shows HTML.
It shows the concept of the analysis means. The page acquired by the HTTP request is in a state that includes HTML tags,
It also includes information that is irrelevant to the content, such as links and layout information. This is removed by using a Perl regular expression, and a form suitable for performing matching processing using keywords is created. On the other hand, there is useful information that can be obtained from the state where the tags are written. For example, the title of the page and the link information provided from the page. The HTML analysis means acquires useful tag information and removes useless tag information to convert it into a format (for example, plain text) suitable for matching processing by keywords. The link information is handled as follows, for example. First, when the score of the corresponding page is equal to or higher than the threshold value, the URL of the link provided from within the page is acquired using the regular expression. The acquired URL is passed to the URL database processing unit and stored in the database. This is because there is a possibility that more valuable information may be obtained by examining the contents of that URL.

(Page Content Analyzing Means) This page content analyzing means performs pattern matching processing based on the context of keywords and text data created by a keyword generating unit, which will be described later. For example, the content is analyzed as follows. to grade. The matching processing using keywords and contexts is performed on the information that has become the plain text after the tags are removed by the HTML analysis means. Keywords and contexts are newly generated by the keyword generation means described later and scored.

It should be noted that default keywords and contexts are input by the user, and by the system designer in advance (for example, for retrieving slanderous remarks, for disseminating rumor about stock prices, and for specific product topics). Etc.)
Sometimes you need to enter it separately. In the present embodiment, a case will be described in which a public relations officer of a certain company (M company) checks whether there is an unwarranted statement, slander or the like regarding the company. Detect keywords and contexts appearing in the text of a page, add the scores (weights) for each, and find the total score of the page. It is assumed that the higher the score, the more likely the information is for the purpose. Regarding addition, simple addition and the position where keywords and contexts appear,
In some cases, weighted points are assigned based on the number of pieces. In this case, when a link is pasted on the page, it is preferable that the importance of the acquired link information is judged and weighted points are assigned. That is, this is a configuration for determining the importance of the page link structure. A page linking (quoting) a page (or URL) of high importance is reasonably high in importance, and a page (or URL) linking (quoting) to a page of high importance is also included in it. According to this, it is considered that the importance is high, and it is suitable for use as an index for evaluating the importance (score) of the page.

For example, assume that there is a page in which the following sentences are posted. In country A, which has severe corporate ethics, he has been hiding the fact of "deficiency" for over 20 years. The "responsibility" for this "defective" concealment is significant, and it seems certain that Company M has lost its credibility. As a keyword of the above sentence,
Set "defect" and "responsibility" and ask for the score of this sentence. The weight of these keywords is "defect = 5 points" and "responsibility =
2 points ". Therefore, the score of this sentence is 12 points (5 points × 2 + 2 points).

(Keyword Generating Unit) As described above, “weight” is set for each keyword and context.
The “weight” is for reflecting the difference in the amount of information in meaning of a sentence. In the default (initial setting), the "weight" may be weighted by the user or may be given by the system designer together with the keyword and context as described above. However, it may not be said that it matches the preference of the user who uses this search system. It's no exception when users give keywords and contexts and their weightings by default.

FIG. 5 conceptually shows the page content analysis function. In the search system according to the present embodiment, the user may initially evaluate the information provided by the system, that is, the information searched for by the user, and the system may weight the keywords and the context. Further, even if the user is not aware of evaluation and weighting, this system has a function of automatically weighting according to how and what information is processed.

(Score Addition) The keyword the user is paying attention to is analyzed, and the score is calculated according to the number of appearances.
This is used as a user point, and is added to the score obtained in the usual way. The method is as follows: first click the title of the search result, the full text of the page is displayed, and at the same time, the keyword is detected and "0.1 x the number of appearances"
It is a method of giving the score obtained in. The score for each keyword obtained by the above method will be used when the same search is performed next time. That is, at the time of the first search, the keyword of the page body is analyzed in real time, and when the keyword of the user's attention is detected, the score obtained for each keyword is added to the normal score. The next search is based on this added value.

With this function, the higher the information including the keyword the user is paying attention to, the higher the score, and the user can quickly find the desired information from the database. If you select YES to the question "Is this information useful for you, is this information useful?", You will be awarded an additional 0.5 points in the same way. FIG. 6 shows how scores are added for each user. In the figure, "15.5
"3" represents the score obtained in the normal state, and "2.5" represents the score obtained for each user. The score obtained for each user constantly changes with repeated use, which enables a search more in line with the user's intention.

(Addition of Keyword) In order to deal with the case where the user wants to subdivide the search content or to improve the accuracy, the user can arbitrarily add the keyword. However, the addition of keywords imposes a heavy burden on the system, so the number of words that can be added is limited according to the performance of the system. In this embodiment, up to three additions are made.

(Addition of URL) The URL is optional by the user
Can be added as a search target. Unlimited number.
If you want to always search for a page found by chance, add a URL.

(New Generation of Keyword) When the user evaluates the information retrieved by the system, or when the system analyzes the behavior of the user and evaluates the analysis result, at that time In addition to updating the weight of registered keywords and contexts, if it is determined that the evaluation and analysis results are highly evaluated, new keywords and contexts are extracted from those information and stored in a database of keywords, etc. register.

(Adjustment for Duplication of Keyword) When a plurality of the same words appear in the same page, the score of the page is meaninglessly high. In order to avoid this and obtain a valid score, the score to be added is changed depending on the number of appearances of the word. Although various evaluation formulas can be considered, the following functions are used in this embodiment. Keyword score to be added = keyword normal score / (number of word appearances × 2) For example, if the normal score (weight) of the keyword is “4”, 4 points are given for the first recall and 2 points are given next.
The number of points to be added is halved, such as 1 point at the end and 1 point at the end.

(Keyword Thesaurus) For example, when there are a plurality of types of user names such as “X car, XXC, X self-engineering”, these are checked in advance or another name is recruited from the user. , Register in advance. And treat them as the same keyword. As the method of detecting the registered company name, the same means as the matching process using the keyword is used. When the company name registered in the page appears, the information of the page is stored in the database.

(Information Obtained from Page) Just before saving in the database, all the information to be saved in the database is complete. Information stored in the database according to the present embodiment includes a title, a text, a URL,
The acquisition time, the link information, the detected keyword, and the score. The user interface uses these to retrieve information from the database. Note that a sorting means is provided that can change the display order of the contents according to the score order, the acquisition time order, etc., for the convenience of the user.

When the score of the page exceeds the threshold value, the information is stored in the database. However, care must be taken when handling new information. For example, assume that the information obtained from the same URL is already in the database. In that case, as shown in FIG. 7, the information saved in the past is deleted from the database and the new information is saved.

As described above, depending on the page, Last-M
odified may not be described in the header. Then, since the update date and time becomes unknown, when the same information is detected, the system tries to acquire the information and add it to the database even though the page has not been updated. To prevent this, the same information in the database is detected immediately before saving. Then, when the same URL is detected, the information of the same URL acquired in the past is deleted from the database when new information is added to the database.

(User Interface) In implementing the user interface, Perl was adopted as the language used in this embodiment. This is because it is the most used CGI language on the Internet. The information stored in the database is the title, body, URL, acquisition time, detected keyword, and score for each page. By using these, the maximum purpose is to extract the information desired by the user from the database and display it. The user interface implemented in this embodiment enables the database to be searched in various ways. FIG. 8 shows an example of an interface using web browser software.

(Search Using Information Acquisition Time) Display information acquired within 3 days, display information acquired within one week, display all and search using acquisition date and time. Of course, it has a user interface that sorts and outputs the data in the order of acquisition date and time.

(Search by score obtained by HTML analysis section page) The evaluation value (risk level) of the page is set by the score of the page. In this embodiment, the risk level is High, M
It is divided into three stages, middle and low. What is the high risk level?
The information is judged to have the highest possibility of being extremely harmful, and the score is 16 or more. The risk level Middle is information that is judged to be highly harmful, and the score is 8 or higher. The risk level Low is information that may be harmful, and the score is 3 or more.
By using these, only the risk level High is displayed, and the risk level Middle
Show above, Show risk level Low or above, Show all 4
It is possible to search in stages.

(Sort by score, new arrival) For users, the pages can be sorted and viewed in the descending order of score. It is also possible to sort and display the pages in the descending order of acquisition time. Furthermore, it is possible to search by using these functions together. Also, the full text of the page can be browsed by clicking the title part of the search result. FIG. 9 shows a state in which the entire text of the page is displayed.

(Backup of Database) The search system according to the present embodiment uses U recorded in the database.
The RLs are cycled and searched according to their importance. Therefore, the contents of the database are very important, and if the contents of the database are destroyed due to a sudden system abnormality, it becomes a serious problem in the operation of the system. Therefore, before starting the tour, we will automatically create a backup of the database.

[0061]

According to the inventions of claims 1 to 6, it is possible to provide a search system capable of accurately collecting information desired by a user from among rebellious information on the Internet. Further, claim 7 to claim 1
According to the invention described in 2, it is possible to provide a search method capable of accurately collecting the information desired by the user from the information that rebels on the Internet. Further, according to the inventions of claims 13 to 18, it is possible to provide a search program capable of accurately collecting the information desired by the user from the information revolting on the Internet.

[Brief description of drawings]

FIG. 1 is a conceptual diagram of an embodiment of the present invention.

FIG. 2 is a diagram showing a structure of an associative array.

FIG. 3 is a diagram showing a concept of a page acquisition unit.

FIG. 4 is a diagram showing the concept of HTML analysis means.

FIG. 5 is a diagram conceptually showing a page content analysis function.

FIG. 6 is a diagram showing a screen display on which a score and a user's evaluation selection can be made according to the present embodiment.

FIG. 7 is a diagram illustrating functions of a database processing unit.

FIG. 8 is a diagram showing an example of an interface using web browser software.

FIG. 9 is a diagram showing an example of full text browsing.

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) G06F 13/00 540 G06F 13/00 540E (71) Applicant 501247223 Kanda Toyotomi 1-10 Nakamachi, Machida-shi, Tokyo 21 (72) Inventor Masakatsu Morii 326 H-term, Tokushima City, Tokushima Prefecture F-term (reference) 5B075 ND03 NK02 NR05 NR15 PP22 PR04 PR08

Claims (18)

[Claims] 1. A keyword creating means having a function of creating a keyword for content search on the Internet and scoring the importance of the keyword, and scoring text data in the content according to the keyword or the like. Content evaluation means for performing the content in the search results obtained based on the keyword created by the keyword creation means, the content evaluation means to evaluate and score, and the content scored. By comparing the keywords with each other, the importance of the keywords is scored, and the keyword creating means creates the keywords to be used in the next search based on the importance of the keywords that are scored. Predictive analysis type search system characterized by updating . 2. A keyword creating means for creating a keyword for content search on the Internet and scoring the importance of the keyword, and a content evaluating means for scoring text data in the content according to the keyword or the like. A URL that collects a URL that can predict that the content desired by the user will appear using the importance of the keyword scored by the keyword creating means and the content scored by the content evaluation means.
A predictive analysis type search system including a collecting means. 3. The content evaluation means comprises tag information analysis means for analyzing tag information in the content.
Alternatively, the predictive analysis type search system described in 2. 4. The keyword creating means comprises input accepting means for accepting a keyword input by a user, and the keyword creating means uses the keyword accepted by the input accepting means to score again. Item 4. The predictive analysis type search system according to any one of Items 3. 5. The keyword creating means has a function of automatically generating a new keyword by extracting a new word or context from the collected information based on the evaluation result of the content evaluation means. The predictive analysis type search system according to any one of claims 1 to 4. 6. A sort means for sorting the contents scored by the content evaluation means in descending order of the score, and a sort result by the sort means can be output to a user interface for providing a search result to a user. The predictive analysis type search system according to any one of claims 1 to 4. 7. A keyword creation procedure having a function of creating a keyword for content search on the Internet and scoring the importance of the keyword, and scoring text data in the content according to the keyword or the like. And a content in a search result obtained based on the keyword created by the keyword creation procedure, and scored by evaluating the content by the content evaluation procedure. By comparing the keywords, the importance of the keywords is scored, and the keywords used in the next search are created based on the importance of the keywords scored by the keyword creation procedure. A predictive analysis type search method characterized by updating. 8. A keyword creating procedure for creating a keyword for content search on the Internet and scoring the importance of the keyword, and a content evaluation procedure for scoring text data in the content according to the keyword or the like. A URL that collects a URL at which it can be predicted that the content desired by the user will appear using the importance of the keyword scored by the keyword creation procedure and the content scored by the content evaluation procedure.
A predictive analysis type search method comprising a collecting procedure. 9. The content evaluation procedure comprises a tag information analysis procedure for analyzing tag information in the content.
Alternatively, the predictive analysis type search method according to item 8. 10. The keyword creation procedure comprises an input acceptance procedure for accepting a keyword input by a user, and a re-scoring procedure for re-scoring using the keyword accepted by the input acceptance procedure. Item 10. The predictive analysis type search method according to any one of Items 9. 11. The keyword creating procedure comprises a procedure for automatically generating a new keyword by extracting a new wording or context from the collected information based on the evaluation result of the content evaluation procedure. The predictive analysis type search method according to any one of claims 7 to 10. 12. The method according to claim 7, further comprising a sorting procedure for sorting the contents scored by the content evaluation procedure in descending order of the score, and a user output procedure for outputting the sorting result by the sorting procedure to the user. 11
The predictive analysis type search method described in any one of 1. 13. A program for realizing a predictive analysis type search method for a computer, the program creating a keyword for content search on the Internet and scoring the importance of the keyword. It has a keyword creation procedure with a function to perform, and a content evaluation procedure for scoring text data in the content according to keywords, etc., and a search result obtained based on the keywords created by the keyword creation procedure The content is evaluated and scored by the content evaluation procedure, the scored content is compared with the keyword, and the importance of the keyword is scored, and is used at the next search by the keyword creation procedure. Keywords are scored based on the importance of the keywords Program for executing create, that thereby continue to update the keywords to be used during the search on the computer Te. 14. A program for causing a computer to realize a predictive analysis type search method, which program creates a keyword for content search on the Internet and scores the importance of the keyword. Using the keyword creation procedure, the content evaluation procedure for scoring the text data in the content according to the keyword, etc., the importance of the keyword scored by the keyword creation procedure and the content scored by the content evaluation procedure , A URL that collects URLs from which the user can expect that the desired content will appear
A program that causes a computer to perform the collection procedure. 15. The program according to claim 13, wherein the content evaluation procedure includes a tag information analysis procedure for analyzing tag information in the content. 16. The method according to claim 13, wherein the keyword creation procedure comprises an input acceptance procedure for accepting a keyword input by a user, and a rescoring procedure for rescoring using the keyword accepted by the input acceptance procedure. Item 16. The program according to any one of Items 15. 17. The keyword creating procedure includes a procedure for automatically generating a new keyword by extracting a new wording or context from the collected information based on the evaluation result of the content evaluation procedure. The program according to any one of claims 13 to 16. 18. The method according to claim 13, further comprising a sorting procedure for sorting the contents scored by the content evaluation procedure in descending order of the score, and a user output procedure for outputting the sorting result by the sorting procedure to the user. 1
7. The program according to any one of 7.