JP4253209B2 - Filtering apparatus, system, method and program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a highly accurate word filtering technique, that is, a filtering apparatus, system, method, and program.
[0002]
[Prior art]
In recent years, with the development of the Internet and computers, contents such as web pages provided on the Internet are increasing explosively. On the other hand, a lot of inappropriate content is provided to youth, and there are concerns about youth healthy development. In addition, privately use computer equipment in the workplace, and browse and obtain inappropriate content that is not related to work, such as buying and selling stocks and goods, etc. through personal Internet trade and Internet auctions, personal hobbies and interests It has been pointed out that there is a disadvantage that the work efficiency of work decreases due to activities such as doing. In addition, there is a problem that computers provided for the purpose of improving the convenience of using public facilities such as libraries are used inappropriately or illegally due to anonymity used by unspecified majority. This is also an aspect of inappropriate content usage that is not intended.
[0003]
The following technique called word filtering is known as a prior art of browsing regulation that prohibits or restricts browsing of inappropriate content. In other words, when connecting a LAN (Local Area Network) in a school or company to the Internet, it is considered inappropriate because it passes through a filtering device realized as a proxy server. A character string such as a short sentence or a word such as “Nasty” is registered in advance in the proxy server as a prohibited word called a black word or the like. Then, by analyzing a character string such as a sentence included in the content to be accessed from a computer on the LAN by computer processing in such a proxy server, it is automatically determined whether or not the content uses a black word, Such content is prohibited from browsing.
[0004]
By the way, if browsing is prohibited even if only one inappropriate term is detected in the content, not only inappropriate content but also contents such as newspaper articles and medical articles that do not hinder browsing as a whole are displayed. There are cases where over-regulation becomes impossible to browse, and such inconvenience is called mis-filtering or over-filtering. On the other hand, a technology has been proposed in which news terms, legal terms, medical terms, etc. are stored in advance as a meaningful term list, and even if there are prohibited words, browsing is allowed if there are meaningful terms in the same content (patents). Reference 1).
[0005]
[Patent Document 1]
JP 2001-282897
[0006]
[Problems to be solved by the invention]
However, in Japanese, there is a feature that characters are continuous with few spaces such as white space. If some character strings are extracted, they are prohibited words. However, depending on the combination with the preceding and following character strings, they may have completely different meanings. is there. This tendency is particularly strong in sentences written in hiragana and katakana. For this reason, when judging based on the character string match with the prohibited word, there is a problem that a part of an insignificant word coincides with the prohibited word, and is erroneously subject to browsing restriction.
[0007]
For example, if katakana's “inrun” is registered in the prohibited word list as a synonym for the kanji “nasty” and browsing of content is prohibited, content that includes “coin laundry” also includes “inrun” in part. Prohibited over-regulation. When such a problem occurs, there is a disadvantage that even if the above prior art (Patent Document 1) using a meaningful term list is applied, the problem can be solved only for contents in a specific field such as law and medical care.
[0008]
In addition, as a conventional technique for creating / updating data such as a prohibited word list used in filtering as described above, the proxy server is accessed from a computer in the same LAN to which the proxy server that performs such filtering is connected. And there was something that manipulated the forbidden word list. However, in such a conventional technology, in order to register and update terms including coined words that occur on the Internet every day, it is necessary to go to the installation location of the computer in the same LAN as the proxy server. was there. In addition, when there are LANs or proxy servers for each of the plurality of bases, there is an inconvenience that these cannot be managed collectively.
[0009]
In addition, there are some cautionary terms that are difficult to judge as black bans that are banned uniformly. there were.
[0010]
The present invention solves the problems of the prior art as described above, and an object thereof is to provide a word filtering technique with excellent accuracy, that is, a filtering apparatus, system, method and program. Another object of the present invention is to facilitate updating of data used for word filtering.
[0011]
[Means for Solving the Problems]
In order to achieve the above object, the invention of claim 1 is subject to browsing restriction in a filtering device that performs browsing restriction based on a character string designated in advance for content browsed via a communication network. Included in content First designation means for accepting designation of each restrictive word as a character string in advance; Content is Even including the restriction word , A string containing the regulated word Overall Content A second designation means for accepting in advance designation of each white word that is a character string to be excluded from the restriction target; About the content including the character string that matches the restriction word from the character string that matches the restriction word, except the character string that matches the restriction word and the character string that matches the white word. The computer realizes a restricting means for restricting browsing or performing evaluation judgment therefor.
[0012]
The invention of claim 4 captures the invention of claim 1 from the viewpoint of a method, and in a filtering method for restricting browsing based on a character string specified in advance for content browsed via a communication network, Subject to Included in content A first designation process for receiving in advance designation of each restriction word that is a character string; Content is Even including the restriction word , A string containing the regulated word Overall Content A second designation process for receiving in advance designation of each white word that is a character string to be excluded from the restriction target; About the content including the character string that matches the restriction word from the character string that matches the restriction word, except the character string that matches the restriction word and the character string that matches the white word. The computer realizes a browsing process or a regulation process for performing an evaluation judgment for the browsing.
[0013]
The invention of claim 7 captures the inventions of claims 1 and 4 from the viewpoint of a computer program. Based on a character string designated in advance by controlling the computer with respect to content browsed via a communication network. In a filtering program that regulates browsing, the program is subject to browsing restrictions on the computer. Included in content The specification of each restriction word that is a character string is accepted in advance, Content is Even including the restriction word , A string containing the regulated word Overall Content Each white word that is a character string to be excluded from the restriction is accepted in advance, About the content including the character string that matches the restriction word from the character string that matches the restriction word, except the character string that matches the restriction word and the character string that matches the white word. It is characterized by having browsing restrictions or evaluation judgments made for it.
[0014]
In these aspects, even in a language with a feature such as Japanese that has few spaces and many consecutive characters, the same character string as a restricted word such as “Inrun” is included, but as a whole, such as “Coin Laundry”. Can avoid over-regulation by specifying white words to be excluded from the restriction. In addition, reception of designation of restricted words and white words is not limited to input from the keyboard, but includes reception of data files.
[0015]
According to a second aspect of the present invention, in the filtering device according to the first aspect, the specifying unit is configured to be able to specify two types of black words and gray words as the restricting word, The content including at least one black word that is not included in the white word is prohibited from browsing, and the content including the gray word that is not included in any of the white words is predetermined as to whether or not to restrict browsing. It is configured to be a target of evaluation judgment based on a standard.
[0016]
The invention according to claim 5 captures the invention according to claim 2 from the viewpoint of a method. In the filtering method according to claim 4, the designation processing includes two types of black words and gray words as the restriction words. The content that includes the gray word that is not included in any of the white words is prohibited, and the restriction process prohibits browsing the content including the black word that is not included in any of the white words. Is characterized by being subject to evaluation judgment based on a predetermined standard as to whether or not to restrict browsing.
[0017]
The invention of claim 8 captures the inventions of claims 2 and 5 from the viewpoint of a computer program. In the filtering program of claim 7, the program sends the black word as the restriction word to the computer. Gray word and two types of designations are accepted, content including at least one black word that is not included in any of the white words is prohibited from browsing, and the gray word that is not included in any of the white words Content containing words is subject to evaluation judgment based on a predetermined standard as to whether or not to restrict browsing.
[0018]
In these aspects, apart from the black word that prohibits the browsing of content uniformly, it is suspected that the content is inappropriate, and a gray word requiring attention can also be designated. In addition, the evaluation of whether or not to prohibit browsing of content containing gray words is performed based on predetermined criteria such as the type and number of appearances of gray words existing in the content, and standard values based on them. Fine browsing restrictions according to the clarity are possible, and over-regulation can be avoided.
[0019]
According to a third aspect of the present invention, there is provided a filtering system in which the filtering device according to the first or second aspect is provided for each of one or more bases, and is connected to the control server through a communication network. For at least one of a restricted word and a white word, one or two or more filtering devices are subject to designation of contents to be changed and stored as update data, and the control based on the update data The computer realizes a reflecting means for reflecting the update of the restricted word or the white word from the server for the corresponding filtering device via the communication network.
[0020]
The invention of claim 6 captures the invention of claim 3 from the viewpoint of a method, and each filtering device that executes the filtering method of claim 4 or 5 for each of one or more bases is subjected to predetermined control. In the filtering method using the filtering system connected to the server for communication by the communication network, the control server changes at least one of the restricted word or the white word of the filtering device for one or more filtering devices. The specification of the content to be received is received and stored as update data, and based on the update data, the control server reflects the update of the restriction word or white word to the corresponding filtering device via the communication network. It is characterized by.
[0021]
In these aspects, the update contents such as new addition, change, and deletion of the restriction word such as black word and white word used in each filtering device at each base are shared by a plurality of devices in the server at the headquarters away from the base. Data can be updated as the contents of updates for individual devices, and can be remotely updated from anywhere to be reflected in each device at each desired location, so new terms including coined words that are born one after another can be dealt with easily and quickly. It becomes.
[0022]
DETAILED DESCRIPTION OF THE INVENTION
Next, an embodiment of the present invention (hereinafter referred to as “embodiment”) will be specifically described with reference to the drawings. The present embodiment can be realized by controlling a computer such as a server system provided with necessary peripheral devices by a program. In this case, various implementation modes of hardware and programs can be changed. Further, the present invention can be grasped as an apparatus, a system, a method and a program, and a computer-readable recording medium recording such a program. As described above, since various implementation modes of the present embodiment are conceivable, virtual circuit blocks that implement the functions of the present invention and the present embodiment are used in the following description.
[0023]
[1. Constitution〕
In the present embodiment, as shown in the functional block diagram of FIG. 1, the filtering device 1 is provided for each of a plurality of bases that are schools, that is, A school to D school, and for control provided in an organization management base such as a school board. The filtering system is connected to the server 2 via the Internet 3 which is a communication network. Each filtering device 1 regulates browsing of content such as web pages browsed via the Internet 3 based on a character string designated in advance, and is configured as a proxy server, for example.
[0024]
In other words, a plurality of student PCs 15 as personal computers (hereinafter referred to as “PCs”) as client systems are connected to the campus LAN for each base, and the content on the web server on the Internet 3 is transferred from each student PC 15. All accesses are mediated by the Internet 3 through ISP (Internet Service Provider) after word filtering by the filtering device 1.
[0025]
In addition to each student PC 15, a base administrator PC 16 is connected to each school LAN, and the filtering device 1 performs word filtering only for access from the base administrator PC 16 among client systems on the same LAN. It is set in advance to pass through without performing.
[0026]
In the filtering device 1 and the control server 2, each element shown in FIG. 1 is realized by the action of a program for controlling each of them, and each element plays a role as means for acting as follows.
[0027]
[2. Action)
[2-1. Outline of action)
That is, the local setting unit 4 of the filtering device 1 is a first designation means for preliminarily designating each restricted word that is a character string that is subject to browsing restriction. For example, a new addition, change, or deletion of the restricted word is performed. The designation is accepted and reflected in the restricted word data 5. Further, the local setting unit 4 is a second designation means for pre-designating each white word that is a character string to be excluded from the regulation target even if the regulation word is included. The designation of change, deletion, etc. is received and reflected in the white word list 6 as data. Moreover, the control part 7 is a control means which performs browsing restrictions or the evaluation judgment for the content including the said restricted word which is not contained in any of the said white words.
[0028]
In addition, the local setting unit 4 accepts two types of designations of black words and gray words as restriction words, and reflects the contents of each in the black word list 51 and the gray word list 52 constituting the restriction word data 5. . In addition, the restricting unit 7 prohibits browsing the content including the black word that is not included in any of the white words, and browsing the content including the gray word that is not included in any of the white words. Whether or not to regulate is subject to evaluation judgment based on a predetermined standard.
[0029]
Further, the remote setting unit 8 of the control server 2 accepts designation of contents to be changed for the restriction word and white word of the filtering device 1 for one or two or more filtering devices 1 and also serves as update data 9. Means for storing. In order to realize this function, the control server 2 is configured as a web server and database server for the administrator, and the administrator provides the control server 2 from the CGI function within the same LAN from the predetermined organization administrator PC 17. If a change content is specified using GUI parts such as an input field, a selection box, a check box, and a radio button on the web page, the update data 9 is reflected.
[0030]
Further, the remote update unit 10 of the control server 2 and the remote update unit 11 of the filtering device 1 are connected to the filtering device 1 from the control server 2 to the corresponding filtering device 1 based on the update data 9 via the Internet 3. This is a reflection means for reflecting the word update.
[0031]
[2-2. Access recommendations and handling of prohibited URLs]
Next, the processing procedure of the word filtering in the present embodiment is shown in the flowchart of FIG. 2, but the following filtering based on the URL may be performed prior to the word filtering. That is, when browsing content on the Internet, a URL (Uniform Resource Locator) of a site that provides the content is designated. Before attempting to access the site based on the URL, the filtering device 1 determines whether the specified URL is an access recommended URL or an access prohibited URL, and a URL list prepared in advance representing them. And URL filtering processing is performed.
[0032]
As a result, when the URL to be accessed is the recommended access URL, the word filtering is omitted, and the filtering device 1 accesses the URL through the Internet 3 to obtain the content, and the student PC 15 as the access source. Forward to. On the other hand, if the URL to be accessed is an access-prohibited URL, the fact that the access is prohibited due to the browsing action on the access-prohibited site is displayed instead of the content on the screen of the student PC 15 as the access source. To do.
[0033]
When the URL to be accessed is neither the recommended access URL nor the prohibited access URL, the filtering apparatus 1 temporarily sends a request for the URL to the access destination based on the HTTP (HyperText Transfer Protocol) procedure, and returns a reply. While receiving data in packet units, the data is subjected to word filtering as described later. At this time, since the content of the web page and the like is described in HTML (Hypertext Markup Language), only the information of the display character string is extracted from the content based on the description rule, and the following from the beginning of the character string: Perform the following process.
[0034]
[2-3. (Word filtering procedure)
That is, as shown in FIG. 2, the restriction unit 7 of the filtering device 1 first compares the received packet with each white word in the white word list 6, and whether a completely matching character string exists in the packet. Check (step 1). If there is a matching character string at this time, the corresponding character string portion in the content is designated as a non-searchable character string for black words and gray words (step 2). This designation may be performed by converting these character string portions into special characters that mean non-searchable character strings. This portion is excluded from the subsequent search target of black words and gray words.
[0035]
In the subsequent step 3, it is searched whether there is any character string that matches any black word in the black word list 51. At this time, if there is a special character in the packet indicating that it is not to be searched, the character string is ignored. When there is a character string that matches the black word in the black word list 51 in the packet (step 3), the subsequent packet constituting the content is ignored and browsing of the content is prohibited (step 10). Specifically, a web page that tells the user that browsing of the content is prohibited is passed to the student PC 15 that requested the browsing of the content, and the processing for the content is terminated.
[0036]
If there is no character string that matches the black word in the packet (step 3), then it is searched whether there is a character string that matches any gray word in the gray word list 52 (step 3). 4). Also at this time, if the character string in the packet includes the special character that means that it is not subject to search, the character string is ignored. When there is a character string that matches any gray word in the gray word list 52 in the packet (step 4), the gray word is registered in the gray word determination reference table (step 5) by repeating the process. (Step 6) All gray words detected in the packet are registered in the determination criterion table.
[0037]
In the following step 7, it is checked whether or not all the packets constituting the content have been examined. If not completed yet, the process returns to step 1, but if completed, the process proceeds to step 8. In this step 8, all the gray words registered in the gray word determination criterion table are classified by type, and at the same time, the number of exposures, that is, the number of appearances is counted. Then, comparison is made with a predetermined standard such as the limit number for each type of gray word, the number of exposure times of the gray word, etc., and when the limit is reached, the process proceeds to the prohibition of browsing (step 10). If the limit is not reached, browsing of the content is permitted (step 9). Specifically, it is determined that the content satisfies the browsing permission criteria, and each packet constituting the content is sequentially sent to the student PC 15 that has requested browsing of the content, and the processing of the content is terminated.
[0038]
[2-4. Update each list)
In order to improve the accuracy of word filtering in the present embodiment described above, not only the creation of each list, that is, the black word list 51, the gray word list 52, and the white word list 6 (collectively referred to as “each list”) is maintained. is important. That is, there is a need for means for promptly absorbing new terms, including coined words that are born every day in various contents of the Internet, in each of the aforementioned lists. Moreover, if each administrator directly manages these lists for each filtering device 1 installed at a base such as a school or a company, the management efficiency is poor. In other words, it is desirable that the updated content based on the management policy unified by the business organization, such as the local school board or the corporate headquarters organization, can be quickly reflected in these lists at each site by the organization manager or business site manager. At the same time, it is desirable to be able to flexibly manage the site based on the site policy.
[0039]
Therefore, in the present embodiment (FIG. 1), in addition to reflecting the desired update contents from the local setting unit 4 to the restricted word data 5 and the white word list 6 for each school that is a base, a place away from those bases Even from this, the headquarters corresponding to the application field can apply the desired update content to the desired object.
[0040]
That is, in the educational field as in this embodiment, the remote setting unit 8 of the control server 2 from the school board that manages all schools in the school district, the administrator of the head office that manages a plurality of offices in the workplace, etc. For at least one of the restricted word or the white word of the filtering device 1, the designation of the content to be created or changed is accepted and stored as the update data 9 for one or more filtering devices 1. If a specific base is specified and each list is managed by update data for the base, it is possible to cope with a case where a small base does not have an administrator.
[0041]
Then, the remote update unit 10 on the control server 2 side and the remote update unit 11 on the filtering device 1 side communicate with each other via the Internet 3, so that the corresponding A is received from the control server 2 based on the update data 9. The update of the regulation word or the white word is reflected on the filtering device 1 of the school D. Such an update may be reflected periodically based on a predetermined time interval set in the filtering device 1 or the control server 2 or a designated time such as midnight, or may be updated by an organization manager or a site manager. An update that has not been reflected in the instruction may be detected and reflected.
[0042]
Note that the local setting unit 4 of the filtering apparatus 1 is accessed from a computer such as a PC used by the site managers of the schools A to D, or the remote setting unit 8 of the control server 2 from the PC such as a PC used by the organization manager. As long as access to the network is accepted not only from within the LAN but also from the outside via the Internet N, the list is updated anytime and anywhere based on the policy of the school or the common policy of the entire school board, etc. Can be added, providing very good convenience.
[0043]
[3. effect〕
As described above, in the present embodiment, even in a language with a feature such as Japanese that has few spaces and many consecutive characters, it includes the same character string as a restricted word such as “inrun”, but “coin laundry” As a whole, over-regulation can be avoided by designating whitewords to be excluded from the restriction.
[0044]
In addition, in the present embodiment, apart from a black word that prohibits the browsing of content uniformly, a gray word requiring attention can be designated because it is suspected to be inappropriate. In addition, the evaluation of whether or not to prohibit browsing of content containing gray words is performed based on predetermined criteria such as the type and number of appearances of gray words existing in the content, and standard values based on them. Fine browsing restrictions according to the clarity are possible, and over-regulation can be avoided.
[0045]
Also, in this embodiment, update contents such as new addition, change, and deletion of restricted words such as black words and white words used in each filtering device at each site are transferred to a plurality of devices at the headquarters server away from the site. Since it is possible to remotely update the data to be updated as data related to common or individual devices and reflect them on each device at the desired timing from anywhere, new terms including coined words that are born one after another on the Internet can be easily and quickly It becomes possible to cope.
[0046]
[4. Other embodiments]
In addition, this invention is not limited to the said embodiment, Other embodiments which are illustrated next are included. For example, the content to which the present invention is applied is not limited to a web page, and the type of content such as newsgroup posting or streaming distribution is free. Further, the applicable language is not limited to Japanese, and may be Chinese, Korean, or the like.
[0047]
In addition, the present invention is realized by installing an application program on a PC directly connected to an ADSL modem, router, etc. at home, etc., instead of being a proxy server separate from the PC as shown in FIG. May be. As described above, when the present invention is realized as an application program resident in the client system, the content such as an HTML file can be decomposed and assembled with the IP packet in a layer lower than the application layer in the OSI basic reference model. Unlike the form, the word filtering process may be performed not on a packet basis but on a content basis.
[0048]
【The invention's effect】
As described above, according to the present invention, it is possible to provide a highly accurate word filtering technique, that is, a filtering device, a system, a method, and a program.
[Brief description of the drawings]
FIG. 1 is a functional block diagram showing a configuration of an embodiment of the present invention.
FIG. 2 is a flowchart showing a processing procedure of word filtering in the embodiment of the present invention.
[Explanation of symbols]
1 ... Filtering device
2 ... Control server
3 ... Internet
4 ... Local setting section
5 ... Regulated word data
51 ... Black word list
52 ... Gray word list
6 ... White Word List
7 ... Regulatory Department
8. Remote setting part
9 ... Update data
10. Remote update unit
11 ... Remote update part
15, 16, 17 ... PC

Claims (8)

For content that is browsed via a communication network, in a filtering device that regulates browsing based on a character string specified in advance,
First designation means for accepting in advance designation of each restricted word that is a character string included in the content subject to browsing restriction;
Even if the content includes the restriction word, as a whole character string including the restriction word, a second designation means for receiving in advance the designation of each white word that is a character string excluding the content from the restriction target;
The content including the character string that matches the restriction word from the character string that matches the restriction word is excluded from the search target as the restriction word and the character string that matches the restriction word is excluded from the search restriction or Regulatory means to make evaluation judgment for that,
Is realized by a computer. The designation means is configured to be able to designate two types of black words and gray words as the restriction words,
The regulating means is
Any content that contains any of the black words that are not included in any of the white words is prohibited from browsing.
The content including the gray word that is not included in any of the white words is configured to be subject to evaluation judgment based on a predetermined criterion as to whether or not to restrict browsing. Filtering device. In the filtering system which provided the filtering device according to claim 1 or 2 for every one or two or more bases, and connected to the control server via a communication network,
In the control server, for at least one of the restriction word or white word of the filtering device, for one or two or more filtering devices, accepting designation of contents to be changed and storing as update data;
Reflecting means for reflecting the update of the restricted word or white word via the communication network from the control server to the corresponding filtering device based on the update data;
A filtering system characterized by realizing a computer. In a filtering method for restricting browsing based on a character string specified in advance for content browsed via a communication network,
A first designation process for accepting in advance designation of each restriction word, which is a character string included in content subject to browsing restriction;
Even if the content includes the restriction word, as a whole character string including the restriction word, a second designation process for receiving in advance the designation of each white word that is a character string that excludes the content from the restriction target;
The content including the character string that matches the restriction word from the character string that matches the restriction word is excluded from the search target as the restriction word and the character string that matches the restriction word is excluded from the search restriction or Regulation processing to make evaluation judgment for that,
A filtering method characterized in that a computer is realized. The designation process accepts two types of designations, black words and gray words, as the restriction words,
The regulation process is
Any content that contains any of the black words that are not included in any of the white words is prohibited from browsing.
5. The filtering method according to claim 4, wherein the content including the gray word that is not included in any of the white words is subjected to evaluation determination based on a predetermined criterion as to whether or not to restrict browsing. A filtering method using a filtering system in which each filtering device that executes the filtering method according to claim 4 or 5 for each of one or more bases is connected to a predetermined control server via a communication network.
In the control server, for at least one of the restriction word or white word of the filtering device, for one or two or more filtering devices, accepting designation of contents to be changed and storing it as update data,
A filtering method characterized in that, based on the update data, an update of a restricted word or a white word is reflected from the control server to the corresponding filtering device via the communication network. In a filtering program that controls browsing based on a character string specified in advance by controlling a computer for content browsed via a communication network,
The program is stored in the computer,
The specification of each restriction word that is a character string included in the content subject to browsing restrictions is accepted in advance,
Even if the content includes the restriction word, the entire character string including the restriction word is allowed to accept in advance the designation of each white word that is a character string that excludes the content from the restriction target,
The content that includes the character string that matches the restriction word from the character string that matches the restriction word is excluded from the search target as the restriction word and the character string that matches the restriction word is excluded from the search restriction or A filtering program characterized by making an evaluation judgment for this purpose. The program is stored in the computer.
As the restriction word, two types of designations, black word and gray word, are accepted,
Browsing any content that contains at least one black word that is not included in any of the white words,
The filtering program according to claim 7, wherein content including the gray word that is not included in any of the white words is subjected to evaluation determination based on a predetermined criterion as to whether or not to restrict browsing.

Images