JP2012178019A - Keyword applying device, content providing system, keyword applying method and program - Google Patents

Abstract

In a system for returning data based on a received keyword, it is possible to reduce a burden on a data input operator and to associate a keyword with data.
[Solution]
The correspondence acquisition unit 202 associates the records in the advertisement information master database 31 with the records in the product information master database 32 based on the similarity of the data stored in these records. Then, the keyword assigning unit 203 assigns the keyword stored in the product information master database 32 to the search key stored in the advertisement information master database 31 based on the association by the correspondence relationship obtaining unit 202, and the product The keyword stored in the advertisement information master database 31 is assigned to the search key stored in the information master database 32.
[Selection] Figure 1

Description

  The present invention relates to a keyword assigning device for assigning a keyword for searching for data provided to a client device, a content providing system including the keyword assigning device, a keyword assigning method for the keyword assigning device, and a program.

  Various systems for returning data associated with a keyword transmitted from a client device have been proposed. For example, in the advertisement information search system described in Patent Document 1, a master server system that stores advertisement information associated with a keyword ranks the advertisement information according to the keyword from the user terminal and ranks the advertisement information. Display (send) advertisement information to search users. This makes it easy for search users to select advertisement information, and to attract effective customers to advertisement information providers who are confident in their products.

JP 2004-38370 A

  The associating operation between keywords and data, which is a preparation operation for returning data associated with the keywords, is a burden on the data input operator who inputs the keywords and reply data to the system. In addition, there is a possibility that an input error or notation fluctuation occurs in the association work, and appropriate keyword association is not performed.

  For example, in the advertisement information search system described in Patent Document 1, the participating store inputs promotion information, keyword program information, or the like, or the service provider is acting on the basis of application documents from the participating store, It is considered that an information input operation performed by an employee or a service provider of a participating store includes an association operation between keywords and data. Such association work becomes a burden on the employee or service provider of the participating store who is the data input worker. In addition, in the association work, there is a possibility that an input error in which the association to be performed is not performed, or conversely, an input error in which the association should not be performed may occur. It will decline.

  In addition, for example, when the data input operator inputs “orange juice” to the keyword “orange juice” to be input, a keyword different from the keyword to be associated is associated. Will lead to a decrease in search accuracy. In addition, there is a high possibility that such notation fluctuations cannot be found in the input data check operation, as compared with the case of obvious mistakes.

  In particular, when the scale of the database is large, the number of data to be worked on becomes enormous, which increases the burden on the data entry worker, and as the work volume increases, there is a high risk of input errors and notation fluctuations. turn into.

  The present invention has been made in consideration of such circumstances, and its purpose is to reduce the burden on the data input operator and to provide a keyword assignment device capable of associating keywords with data, To provide a content providing system, a keyword assigning method, and a program.

  The present invention has been made to solve the above-described problems. The keyword assigning apparatus according to one aspect of the present invention can select each field in the second database when any field in the first database is selected. Calculating a data similarity with the selected field, and selecting a field of the second database determined to have the highest similarity, and the selected of the first database A record of the first database and a record of the second database including data determined to be similar by performing similarity determination between each data of the field and each data of the selected field of the second database To the correspondence acquisition unit, the information stored in the first database, and the record including the information. Characterized by comprising a keyword assignment unit for associating the obtained from Tagged record of said second database keyword, the.

  The keyword assigning apparatus according to an aspect of the present invention is the keyword assigning apparatus described above, wherein each data of the selected field of the first database and the selected field of the second database Each of the data is character string data, and the correspondence acquisition unit obtains each character string data of the selected field of the first database and of the selected field of the second database. Find the longest common substring with each character string data, and select either the character string data of the selected field of the first database or the character string data of the selected field of the second database. And the selection distance of the first database based on the edit distance in the approximate character string matching with the longest common subsequence. And each data field is, and performs similarity determination that the data of the selected field of the second database.

  A content providing system according to an aspect of the present invention stores a content database that stores content that is data provided to a client device, and a content search key that is a search key for the content in the content database. A content providing system comprising a database, a second database, a search key database, a keyword assigning device, and a content search device, wherein the keyword assigning device is any field of the first database. Is selected, the data similarity between each field of the second database and the selected field is calculated, and the field of the second database determined to have the highest similarity is selected. A field selector and the first data Performing a similarity determination between each data of the selected field of the base and each data of the selected field of the second database, and the record of the first database including the data determined to be similar to the record of the first database; From the correspondence acquisition unit for associating a record in the second database, the content search key stored in the first database, and the record in the second database associated with the record including the content search key A keyword assigning unit for associating the obtained keyword, and a correspondence information writing unit for writing information indicating the association by the keyword assigning unit to the search key database, and the content search device includes the content database. The search term for searching stored content is Based on the content search key acquired by the content search key acquiring unit, the content search key acquiring unit acquiring the content search key by searching the search key database based on the search term when transmitted from the client device A content acquisition unit that searches the content database to acquire content, and a communication unit that transmits the content acquired by the content acquisition unit to the client device.

  The keyword assigning method according to one aspect of the present invention is a keyword assigning method of the keyword assigning device, and when any field of the first database is selected, each field of the second database and the selection are selected. A field selection step of calculating the similarity of the data to the selected field and selecting the field of the second database determined to have the highest similarity, and each of the selected fields of the first database Similarity determination is performed between the data and each data of the selected field of the second database, and the record of the first database including the data determined to be similar is associated with the record of the second database A correspondence acquisition step, information stored in the first database, and the information Characterized by comprising a keyword assignment step for associating the keyword obtained from the record of said second database associated with the code, the.

  According to another aspect of the program of the present invention, when any field of the first database is selected by the computer, the degree of data similarity between each field of the second database and the selected field. A field selection step of selecting a field of the second database that has been determined to have the highest similarity, each data of the selected field of the first database, and the data of the second database A correspondence determination step of performing similarity determination with each data of the selected field and associating the record of the first database including the data determined to be similar with the record of the second database; Information stored in the database and the second data associated with the record including the information And keyword assignment step for associating the keywords obtained from over scan records a program for execution.

  According to this invention, it is possible to reduce the burden on the data input operator and to associate the keyword with the data.

It is a block diagram which shows schematic structure of the content provision system for advertisement in one Embodiment of this invention. In the same embodiment, it is explanatory drawing which shows the example of the data which the advertisement information master database has stored. In the same embodiment, it is explanatory drawing which shows the example of the data which the product information master database has stored. In the same embodiment, it is explanatory drawing which shows the example of the data which the advertisement information field database has stored. In the embodiment, it is explanatory drawing which shows the example of the data which the product information field database stores. In the embodiment, it is explanatory drawing which shows the example of the data which the basic database stores. In the embodiment, it is explanatory drawing which shows another example of the data which the basic database stores. In the embodiment, it is explanatory drawing which shows the example of the data which the keyword database has stored. In the embodiment, it is explanatory drawing which shows the example of the field of an advertisement information master database selected by the field selection part. It is explanatory drawing which shows the example of the similarity determination by the field selection part in the embodiment. In the same embodiment, it is explanatory drawing which shows the example of the field of the product information master database selected by the field selection part. In the same embodiment, it is explanatory drawing which shows the example of the longest common character string which a correspondence acquisition part produces | generates. In the same embodiment, it is a flowchart which shows the outline of the process sequence which a keyword provision apparatus provides a keyword to the search key of advertising content. In the embodiment, a field selection unit is a flowchart showing a processing procedure for selecting a field of an advertisement information master database and a field of a product information master database. In the embodiment, a correspondence relationship acquisition unit is a flowchart showing a processing procedure for generating data in an advertisement information field database and data in a product information field database. In the embodiment, a correspondence relationship acquisition unit is a flowchart showing a processing procedure for associating records in an advertisement information master database with records in a product information master database. In the embodiment, the keyword assignment unit is a flowchart showing a processing procedure for associating a keyword with a search key of an advertisement information master database, a search key of a product information master database, and a search key of a material file database. In the same embodiment, it is a flowchart which shows the process sequence in which a content search apparatus provides the content for advertisement.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a configuration diagram showing a schematic configuration of an advertisement content providing system according to an embodiment of the present invention. In the figure, an advertisement content providing system 1 includes a content search device 10, a keyword (KW) assigning device 20, an advertisement information master database (Database) 31, a product information master database 32, and a material file. A database 33, an advertisement information field database 34, a product information field database 35, a basic database 36, and a keyword database 37 are provided. The content search apparatus 10 includes a communication unit 101, a content search key acquisition unit 102, and a content acquisition unit 103. The keyword assignment device 20 includes a field selection unit 201, a correspondence relationship acquisition unit 202, and a keyword assignment unit 203. The content search device 10 is connected to the client device 9.

The advertisement content providing system 1 is an example of a content providing system. When a keyword is transmitted from the client device 9, the advertisement content providing system 1 returns an advertisement content serving as a material for generating an advertisement related to the keyword. The “content” here is a moving image, a still image, text, sound, or a combination thereof.
The client device 9 transmits a keyword to the advertisement content providing system 1 (content search device 10), and acquires the advertisement content. For example, when the client device 9 is a web server device that provides a web page search service and receives a search request including a keyword, the client device 9 returns a response including web page information of the search result and an advertisement related to the keyword. At that time, the client device 9 transmits (transfers) the received keyword to the advertising content providing system 1, generates an advertisement using the advertising content returned from the advertising content providing system 1, and includes the advertisement in the response.

Upon receiving the keyword transmitted from the client device 9, the content search device 10 acquires advertising content from the advertising information master database 31, the product information master database 32, or the material file database 33 based on the keyword, and The content is returned to the client device 9.
The communication unit 101 communicates with the client device 9 to transmit / receive various data. In particular, the communication unit 101 receives the keyword transmitted from the client device 9 and outputs the keyword to the content search key acquisition unit 102, and transmits the advertising content output from the content acquisition unit 103 to the client device 9.

When a keyword (search word) is transmitted from the client device 9, the content search key acquisition unit 102 searches the keyword database 37 based on the keyword, and the advertisement information master database 31, product information master database 32, or material file The search key of the database 33 is acquired and the search key is output to the content acquisition unit 103.
The “search key” here is an identifier that uniquely identifies a record in the database. This search key specifies a record for storing the advertisement content when the content search apparatus 10 searches the advertisement information master database 31, the product information master database 32, or the material file database 33 to acquire the advertisement content. Used for.
The “record” here is a row in the database. Hereinafter, an item included in a record is referred to as a “field”. Fields constitute columns in the database.

  In the following description, both the data configured to be searchable and the means for providing a process for the data are referred to as “database” unless otherwise required. Note that the database as the data processing providing means may be realized using database-dedicated hardware such as a database server. Alternatively, the memory included in the content search apparatus 10 stores data, and a CPU (Central Processing Unit) included in the content search apparatus 10 performs various processes on the data by reading and executing a predetermined program. Alternatively, it may be realized using hardware other than the database dedicated hardware.

  The content acquisition unit 103 searches the advertisement information master database 31, the product information master database 32, or the material file database 33 based on the search key acquired by the content search key acquisition unit 102, acquires the advertisement content, and Content for output to the communication unit 101.

  The keyword assigning device 20 stores the search keys of the advertisement information master database 31, the product information master database 32, and the material file database 33 that store advertisement content in the keyword database 37 in association with the keywords. More specifically, the keyword assigning device 20 associates the record of the advertisement information master database 31 with the record of the product information master database 32, and for each associated record, from the record, the advertisement information master database 31. , The search key of the product information master database 32, the search key of the material file database 33, and the keywords associated with these search keys, and writing them in the same record of the keyword database 37, Is associated.

The field selection unit 201 selects a field to be a similarity determination target when the correspondence acquisition unit 202 associates a record in the advertisement information master database 31 with a record in the product information master database 32.
The correspondence acquisition unit 202 associates the record of the advertisement information master database 31 with the record of the product information master database 32 based on the field data selected by the field selection unit 201.

  For each record associated by the correspondence acquisition unit 202 from the advertisement information master database 31 and the product information master database 32, the keyword assigning unit 203 and the search key of the advertisement information master database 31 and the product information master database 32 The search key, the search key of the material file database 33, and the keywords associated with these search keys are read out. Then, the keyword assignment unit 203 writes the read search key and keyword into the keyword database 37. At that time, the keyword assigning unit 203 uses the search key and keyword read from one record of the advertisement information master database 31 and the search key and keyword read from the record of the product information master database 32 associated with the record. , The same record in the keyword database 37 is written.

The advertisement information master database 31 is an example of a first database, and stores data related to various advertisements in advance.
FIG. 2 is an explanatory diagram illustrating an example of data stored in the advertisement information master database 31. In the figure, each record in the advertisement information master database 31 stores a series of data relating to one advertisement, and each record includes an advertisement ID (Identifier) field, an advertiser field, a product name field, and a product. It has a name kana field, a medium field, and a material ID field.

  The advertisement ID field stores an advertisement ID that is a search key of the advertisement information master database 31. The advertiser field stores an advertiser name. The product name field stores the product name to be advertised. The product name Kana field stores the product name written in kana. The medium field stores advertisement medium type information. The material ID field stores a material ID that is a search key of the material file database 33.

  Of the data stored in the advertisement information master database 31, the advertisement ID and the material ID are used as a content search key that is a search key used when the content search apparatus 10 acquires content for advertisement. The advertiser name, product name, and the like are written into the keyword database 37 by the keyword assigning device 20 as keywords associated with the search key.

The product information master database 32 is an example of a second database, and stores data related to various products in advance.
FIG. 3 is an explanatory diagram showing an example of data stored in the product information master database 32. In the figure, each record of the product information master database 32 stores a series of data relating to one product, and each record includes a product ID field, a manufacturer field, a product name field, and a product name Kana field. , A type field, a form field, and a release date field.

  The product ID field stores a product ID that is a search key for the product information master database 32. The manufacturer field stores the manufacturer name of the product. The product name field stores the product name. The product name Kana field stores a product name written in kana. The type field stores product type information. The form field stores product form (“can”, “bin”, etc.) information. The release date field stores product release date information.

  Of the data stored in the product information master database 32, the product ID is used as a content search key that is a search key when the content search device 10 acquires advertising content. The manufacturer name, product name, and the like are written in the keyword database 37 by the keyword assigning device 20 as keywords associated with the search key.

The material file database 33 stores in advance content files that store advertising content, such as moving image files that store television advertisement images, in association with material IDs that are search keys of the material file database 33.
The advertisement information master database 31, the product information master database 32, and the material file database 33 are all examples of content databases, and the content search apparatus 10 can store content files stored in the material file database 33 and advertisement information masters. Data relating to advertisements stored in the database 31 and data relating to products stored in the product information master database 32 are provided (returned) to the client device 9 as advertisement contents.

The advertisement information field database 34 stores the data stored in the field selected by the field selection unit 201 among the data stored in the advertisement information master database 31 in association with the advertisement ID.
FIG. 4 is an explanatory diagram showing an example of data stored in the advertisement information field database 34. In the figure, each record in the advertisement information field database 34 corresponds to each record in the advertisement information master database 31, and the advertisement information field database 34 includes the product name kana among the data stored in the advertisement information master database 31. The field data is stored in association with the advertisement ID.
As will be described later, the data stored in the advertisement information field database 34 is written by the field selection unit 201.

  In the product information field database 35, the correspondence acquisition unit 202 generates a common partial sequence generated from the data in the advertisement information master database 31 and the data in the product information master database 32, the search key of the advertisement information master database 31, and the product information master database. Stored in association with 32 search keys. The “partial string” here is a character string obtained by extracting some characters from the character string and arranging them in the same order as the original character string. Further, the “common partial sequence” is a partial sequence common to a plurality of character strings. Hereinafter, the longest common subsequence is referred to as a “longest common subsequence”.

FIG. 5 is an explanatory diagram showing an example of data stored in the product information field database 35. In the drawing, the common partial sequence generated by the correspondence acquisition unit 202 is stored in the common partial sequence field. Note that “(null)” indicates an empty character string (a character string having a character string length of 0).
The advertisement ID field stores a search key (advertisement ID) of a record that stores data that is a source of common partial sequence generation in the advertisement information master database 31. The product ID field stores a search key (product ID) for a record that stores another data that is the original data for generating the common partial sequence in the product information master database 32.
The common partial sequence generated by the correspondence relationship acquisition unit 202 will be described later.

The basic database 36 stores data indicating the correspondence between the records in the advertisement information master database 31 and the records in the product information master database 32.
FIG. 6 is an explanatory diagram showing an example of data stored in the basic database 36. In the figure, each record of the basic database 36 stores data indicating a record of the advertisement information master database 31 and a record of the product information master database 32 that are associated by the correspondence acquisition unit 202. Has an integrated ID field, an advertisement ID field, a product name Kana field, a product ID field, a product name Kana field, and a common subsequence field.

  The integrated ID field stores an integrated ID that is a search key of the basic database 36. The advertisement ID field stores an advertisement ID indicating a record on the advertisement information master database 31 side that is associated by the correspondence acquisition unit 202. The product name Kana field stores the product name by kana writing used in the association. The product ID field and the product name Kana field are respectively an advertisement ID indicating a record on the product information master database 32 side associated with the correspondence relationship acquisition unit 202, and a product by kana written in the record. Stores a name. The common partial sequence field stores the common partial sequence used for the association.

The data stored in the basic database 36 is not limited to the one-to-one correspondence between the records in the advertisement information master database 31 and the records in the product information master database 32.
FIG. 7 is an explanatory diagram showing another example of data stored in the basic database 36. As shown in rows L11 to L12 in the figure, a plurality of product IDs may be associated with one advertisement ID. That is, a plurality of records in the product information master database 32 may be associated with one record in the advertisement information master database 31.
For example, there are multiple products corresponding to one product, such as products with 350 ml (milliliter) and products with 1.5 L (liter) sold under the same brand name “apple juice” In this case, the basic database 36 stores a row corresponding to each product. As a result, keywords for each product such as “350 ml” and “1.5 L” can be associated with the product (advertisement), and the hit rate can be increased.

Further, in the row L13 and the row L14 in the figure, the same product name Kana (D12 and D22) is associated with different advertisement IDs (D11 and D21). Here, since the product name is only given as a guide indicating the content of the advertisement, a relatively rough name (with a small number of characters) is used. For this reason, as shown in line L13 and line L14, the same product name can be associated with different advertisements.
On the other hand, in different rows L13 and L14, different product IDs (D13 and D23) and different product names Kana (D14 and D24) are stored. Since the product name is used to identify the product, a relatively detailed name (having a large number of characters) is used. For this reason, different product names are associated with each product as shown in the rows L13 and L14.
For the data of a database (for example, the advertisement information master database 31) that stores rough keywords such as the product name Kana, the product name Kana such as the product name Kana is obtained from another database (for example, the product information master database 32). By acquiring and associating detailed keywords, it is possible to increase the number of keywords and increase the hit rate.

  In addition, as shown in a row L15 in the figure, there may be an advertisement ID that is not associated with any product ID. That is, there may be a record in the advertisement information master database 31 that is not associated with any record in the product information master database 32.

The keyword database 37 stores search keys of the advertisement information master database 31, the product information master database 32, and the material file database 33, which are content databases, and keywords associated with these search keys.
FIG. 8 is an explanatory diagram showing an example of data stored in the keyword database 37. In the figure, each record in the keyword database 37 corresponds to each record in the basic database 36, and each record includes an integrated ID field, an advertisement ID field, a product ID field, a material ID field, and an advertiser field. A manufacturer field, a product name field, a product name field, a product name Kana field, a product name Kana field, and a type field.

The integrated ID field stores the integrated ID in the basic database 36. This integrated ID is also used as a search key for the keyword database 37.
The advertisement ID field and the product ID field respectively store the advertisement ID and the product ID stored in the record of the basic database 36 specified by the integrated ID. The advertisement ID and the product ID indicate the correspondence between the fields of the advertisement information master database 31 and the fields of the product information master database 32.

The material ID field, the advertiser field, the product name field, and the product name Kana field are respectively stored in the record of the advertisement information master database 31 specified by the advertisement ID, and the advertiser ID. A name, a product name, and a product name written in kana are stored.
The manufacturer field, the product name field, the product name kana field, and the type field are respectively stored in the record of the product information master database 32 specified by the product ID, Stores the product name by kana writing and product type information.

  The content search device 10 stores the advertiser name, manufacturer name, product name, product name, product name by kana writing, product name by kana writing, and product type information stored in the keyword database 37. When searching for advertising content, it is used as a keyword associated with the search key of the content database. When a keyword is transmitted from the client device 9, the content search device 10 acquires a keyword similar to the keyword from the keywords stored in the keyword database 37, and uses the search key associated with the acquired keyword. Get advertising content.

Next, field selection by the field selection unit 201 will be described with reference to FIGS. The field selection unit 201 selects a field as preprocessing for associating the search key with the keyword performed by the keyword assigning unit 203.
FIG. 9 is an explanatory diagram illustrating an example of fields in the advertisement information master database 31 selected by the field selection unit 201. FIG. 10 is an explanatory diagram illustrating an example of similarity determination by the field selection unit 201. FIG. 11 is an explanatory diagram showing an example of fields in the product information master database 32 selected by the field selection unit 201.

  The field selection unit 201 first selects one of the fields of the advertisement information master database 31 as shown in FIG. For example, the field selection unit 201 is set in advance to select a specific field among the fields of the advertisement information master database 31 and selects the field. In the present embodiment, the field selection unit 201 selects a product name kana field in the advertisement information master database 31. Hereinafter, the field selected by the field selection unit 201 is referred to as “selected field”.

  Next, the field selection unit 201 determines, for each data in the selection field of the advertisement information master database 31, whether the data is similar to each data in the product information master database 32. For example, in the example of FIG. 10, the field selection unit 201 determines the similarity between the data D111 “Mokamto” of the product name kana field that is the selection field of the advertisement information master database 31 and each data of the product information master database 32. To do. Then, the field selection unit 201 obtains a determination result that the data D121 “Ichiban Mocamou 200ml” is similar to the data D111 “Mocamou”, and the other data is dissimilar to the data D111. Furthermore, the field selection unit 201 performs the same determination for other data of the product name Kana field such as the data D112 to D114.

Here, for example, the field selection unit 201 determines that a character string (data) whose edit distance is equal to or smaller than a predetermined threshold in similar character string matching is similar, and determines that a character string whose edit distance is larger than the threshold is dissimilar. The “approximate character string collation” here refers to the number of necessary editing operations when the character string B is generated from the character string A by performing only one character insertion, one character deletion, and one character replacement as editing operations. In this process, the edit distance between the character string A and the character string B is obtained.
For example, from the data D112 “Ichiban Mocamou 200ml”, 1 character is deleted 11 times, “I”, “Chi”, “B”, “N”, “” (blank), “” (blank), “ By deleting each of the characters “2”, “0”, “0”, “m”, and “l”, the data D111 “Mokamto” can be generated. Therefore, the edit distance between the data D112 and the data D111 is 11.
Various algorithms for approximate character string matching have been proposed, and the field selection unit 201 can perform an approximate character string query using a known algorithm.

The similarity determination method performed by the field selection unit 201 is not limited to the above-described method, and various methods that can determine the similarity of two character strings can be used. For example, the field selection unit 201 makes a pair of the data in the selection field of the advertisement information master database 31 and the data in the product information master database 32, such as selecting one character string having the smallest distance in the approximate character string matching. You may make it select by 1 correspondence (it determines with similarity).
Alternatively, the field selection unit 201 determines that a character string having the longest common partial string length equal to or greater than a predetermined threshold is similar, and determines that a character string less than the predetermined threshold is dissimilar, etc. You may make it perform the similarity determination of a character string. Various algorithms such as the Hunt-Szymanski LCS algorithm have been proposed for obtaining the longest common substring, and the field selection unit 201 can obtain the longest common character string length using a known algorithm.

  Next, the field selection unit 201 selects, from the fields of the product information master database 32, the field that contains the most data similar to any of the selected fields of the advertisement information master database 31 based on the similarity determination result. To do. In the example of FIG. 11, the field that contains the most data similar to any data in the product name Kana field that is the selection field of the advertisement information master database 31 is the product name Kana field. Therefore, the field selection unit 201 sets this product name kana field as a selection field of the product information master database 32.

As described above, the field selection unit 201 selects a highly common field by selecting a field containing a lot of similar data as the selection field of the advertisement information master database 31 and the selection field of the product information master database 32. It is expected. For example, in the example of FIG. 11, a product name Kana field and a product name Kana field indicating the name of an article by kana writing are both selected.
Using this highly common field data, the correspondence acquisition unit 202 associates the records of the advertisement information master database 31 with the records of the product information master database 32, and the keyword assigning unit 203 according to the association. However, it is expected that an appropriate keyword can be assigned by assigning a keyword to the search key.

  For example, in the example of FIG. 11, data D111 and data D121 indicate the same type of article (mocha-free sugar produced by Ichiban). Accordingly, the search key D211 for the advertisement content of the article includes the maker name data D221 “Ichiban”, the product name data D222 “Ichiban Mocha Unsweetened 200 ml”, and the product name data D121 “Kana”. A keyword related to the advertisement target product can be given by giving “Ichiban Mokamuto 200 ml” or the type information data D223 “canned coffee” as a keyword. Therefore, when the content search apparatus 10 acquires advertising content based on the keyword from the client apparatus 9, it is expected that appropriate content can be acquired.

Next, generation of data of the product information field database 35 by the correspondence acquisition unit 202 will be described with reference to FIG.
FIG. 12 is an explanatory diagram illustrating an example of the longest common character string generated by the correspondence acquisition unit 202.
The correspondence acquisition unit 202 obtains the longest common partial sequence as a common partial sequence for each combination of the data of the selection field of the advertisement information master database 31 and the data of the selection field of the product information master database 32. Write in the product information field database 35 in association with the ID. As described above, the longest common partial sequence is the longest of the partial sequences common to a plurality of character strings. Here, the data in any of the selection fields of the advertisement information master database 31 and the product information master database This is the longest of the partial sequences common to any data in the 32 selection fields.

In the example of FIG. 12, the longest common subsequence of data D113 “bean taprino coffee” and data D322 “bon bean tap reno coffee 200 ml” is “bean taprino coffee”. The correspondence relationship acquisition unit 202 associates the longest common partial sequence, the search key (advertisement ID) D311 of the record including the data D113, and the search key (product ID) D321 of the record including the data D322, thereby obtaining product information. Write to the field database 35.
The correspondence relationship acquisition unit 202 performs the same processing for other combinations of the data of the selection field of the advertisement information master database 31 and the data of the selection field of the product information master database 32. As a result, the data shown in FIG. 5 is stored in the product information field database 35.

The data stored in the product information field database 35 is used when the correspondence acquisition unit 202 associates the records in the advertisement information master database 31 with the records in the product information master database 32.
For example, the correspondence acquisition unit 202 performs similarity determination between the product name Kana field data in FIG. 2 and the common partial sequence field data in FIG. 5 by approximate character string matching. Then, the correspondence relationship acquisition unit 202 stores the data “Mokamutou” in the product name Kana field of the record with the advertisement ID “M1012” in FIG. Since the edit distance with “” is equal to or less than a predetermined threshold, it is determined that both data are similar. Then, the correspondence relationship acquisition unit 202 sets the advertisement ID “M1012”, the product name Kana field data “Mokamto”, the product ID “PID0034”, and the common subsequence field data “Mokamto” in the basic database 36. Write to record. The correspondence acquisition unit 202 performs the same processing for other combinations of the product name Kana field data in FIG. 2 and the common subsequence field data in FIG. 5, and further integrates them as search keys in the basic database 36. Write the ID. As a result, the data shown in FIG. 6 is stored in the basic database 36.

Here, the correspondence acquisition unit 202 directly uses the data of the selection field of the advertisement information master database 31 and the data of the selection field of the product information master database 32 to directly record the records of the advertisement information master database 31 and the product information master database. It is also possible to associate with 32 records. However, since the product name Kana field data is added with a manufacturer name and content information by kana writing, there is a possibility that the edit distance becomes large and appropriate association cannot be performed.
For example, the edit distance between the product name Kana field data “Mokamuto” (data D111 in FIG. 11) and the product name Kana field data “Ichiban Mocamou 200ml” (data D121 in FIG. 11) is 11. Therefore, when the similarity determination threshold is set to 10 or less, the correspondence acquisition unit 202 determines that these data are dissimilar and no longer associates records.

  Therefore, as described above, the correspondence relationship acquisition unit 202 performs similarity determination using the longest common subsequence. Accordingly, the manufacturer name and content information can be deleted in advance from the data of the product name Kana field, and the correspondence acquisition unit 202 can perform more appropriate similarity determination and perform more appropriate association.

  The similarity determination method performed by the correspondence acquisition unit 202 is not limited to the above-described method, and various methods that can determine the similarity of two character strings can be used. For example, in the same manner as described above for the field selection unit 201, the correspondence relationship acquisition unit 202 selects one character string having the smallest distance in the approximate character string collation, and the record of the advertisement information master database 31 and the product The records in the information master database 32 may be selected in a one-to-one correspondence (determined as similar). Alternatively, the correspondence relationship acquisition unit 202 determines that the character string having the longest common partial string length equal to or greater than a predetermined threshold is similar, and determines that the character string that is less than the predetermined threshold is dissimilar, and the like, The character string similarity determination may be performed as follows.

Next, the operation of the advertisement content providing system 1 will be described with reference to FIGS.
FIG. 13 is a flowchart showing an outline of a processing procedure in which the keyword assigning device 20 assigns a keyword to a search key for advertising content.
The keyword assigning device 20 performs the process shown in FIG. 1 in advance (as a prior process for the content search device 10 to receive the keyword from the client device 9 and acquire the advertising content).

  First, the field selection unit 201 selects one of the fields of the advertisement information master database 31 and one of the fields of the product information master database 32 (step S101). Next, the correspondence acquisition unit 202 stores the advertisement information field database 34 described with reference to FIG. 4 based on each data of the selection field of the advertisement information master database 31 and each data of the selection field of the product information master database 32. Data and data of the product information field database 35 described with reference to FIG. 5 are generated (step S102). The correspondence acquisition unit 202 then associates the records in the advertisement information master database 31 with the records in the product information master database 32 based on the similarity between the data in the advertisement information field database 34 and the data in the product information field database 35. The data indicating the correspondence is written in the basic database 36 (step S103). Next, the keyword assigning unit 203 searches the advertisement information master database 31, the product information master database 32, and the material file database 33, based on the association performed by the correspondence acquisition unit 202. In addition, the keyword read from the advertisement information master database 31 and the keyword read from the product information master database 32 are associated with each other, and data indicating this association is written in the keyword database 37 (step S104).

FIG. 14 is a flowchart showing a processing procedure in which the field selection unit 201 selects a field in the advertisement information master database 31 and a field in the product information master database 32. The field selection unit 201 performs the process shown in FIG. 13 in step S101.
First, the field selection unit 201 selects one of the fields of the advertisement information master database 31 (Step S201). For example, the field selection unit 201 is set in advance to select a specific field among the fields of the advertisement information master database 31 and selects the field.

Next, the field selection unit 201 starts a processing loop L11 for each data of the field selected in step S201 (step S211). Further, the field selection unit 201 starts a process loop L12 for each piece of data in the product information master database 32 (step S212). Then, the field selection unit 201 performs similarity determination between each data in the field selected in step S201 and each data in the product information master database 32 (step S213).
Then, the field selection unit 201 performs similarity determination for all combinations of each data in the field selected in step S201 and each data in the product information master database 32, and then ends the loops L12 and L11 (steps). S214, S215).

Next, the field selection unit 201 starts a process loop L13 for each field of the product information master database 32 (step S221). Then, the field selection unit 201 determines the number of data determined to be similar to any data in the selection field of the advertisement information master database 31 in step S213 for each field of the product information master database 32. The similarity is calculated (aggregated) between the field and the selected field of the advertisement information master database 31 (step S222).
Then, the field selection unit 201 ends the loop L13 after calculating the similarity for all the fields in the product information master database 32 (step S223).

Next, the field selection unit 201 selects a field determined to have the highest similarity calculated in step S222 among the fields of the product information master database 32 (step S231). Then, the field selection unit 201 outputs information indicating the selection field of the advertisement information master database 31 in step S201 and the selection field of the product information master database 32 in step S231 to the correspondence acquisition unit 202.
Thereafter, the process of FIG.

FIG. 15 is a flowchart illustrating a processing procedure in which the correspondence relationship acquisition unit 202 generates data in the advertisement information field database 34 and data in the product information field database 35. The correspondence acquisition unit 202 performs the process in FIG. 13 in step S102 of FIG.
First, the correspondence acquisition unit 202 sets the product information field database 35 to a blank line (number of records 0) as initialization of the product information field database 35 (step S301).

Next, the correspondence relationship acquisition unit 202 starts a processing loop L21 for each data in the selection field of the advertisement information master database 31 (step S311). Further, the correspondence acquisition unit 202 starts a processing loop L22 for each data in the selection field of the product information master database 32 (step S312). Then, the correspondence acquisition unit 202 generates the longest common partial sequence of each data of the selection field of the advertisement information master database 31 and each data of the selection field of the product information master database 32 (step S313). Then, the correspondence acquisition unit 202 adds one row (one record) to the product information field database 35, and the longest common partial sequence generated in step S313 and the longest common partial sequence are generated in the added row. The advertisement ID of the field including the data of the advertisement information master database 31 and the product ID of the field including the data of the product information master database 32 from which the longest common subsequence is generated are written (step S314).
Then, the correspondence acquisition unit 202 performs the process of step S314 on all combinations of each data of the selection field of the advertisement information master database 31 and each data of the selection field of the product information master database 32, and then the loop L22 and L21 is terminated (steps S315 and S316).

Next, the correspondence acquisition unit 202 blanks the advertisement information field database 34 as an initialization process of the advertisement information field database 34. Then, the correspondence acquisition unit 202 extracts the advertisement ID field and the selection field from the data in the advertisement information master database 31, and writes the data in which these are associated with each other in the advertisement information field database 34 (step S321).
Thereafter, the process of FIG.

FIG. 16 is a flowchart illustrating a processing procedure in which the correspondence acquisition unit 202 associates a record in the advertisement information master database 31 with a record in the product information master database 32. The correspondence acquisition unit 202 performs the process in FIG. 13 in step S103 in FIG.
First, the correspondence acquisition unit 202 blanks the basic database 36 as an initialization process of the basic database 36 (step S401).

Next, the correspondence relationship acquisition unit 202 starts a process loop L31 for each record in the advertisement information field database 34 (step S411). Here, the record to be processed in the loop L31 is represented by a record i.
The correspondence relationship acquisition unit 202 extracts a record storing the advertisement ID of the record i from the product information field database 35 (step S412).

Furthermore, the correspondence acquisition unit 202 starts a process loop L32 for each record extracted in step S412 (step S413). Here, a record to be processed in the loop L32 is represented by a record j.
The correspondence acquisition unit 202 calculates the similarity between the data of the selected field of the record i and the data of the common partial sequence field of the record j (step S414), and performs similarity determination (step S415).

When it is determined that they are similar (step S415: YES), the correspondence acquisition unit 202 adds one line to the basic DB, and the advertisement ID stored in the record i and the data in the selection field of the record i are added to the added line. Then, the product ID stored in the record j and the data of the common split field of the record j are written, and an integrated ID for identifying the added row is generated and written in the added row (step S416).
Thereafter, the process proceeds to step S417.

On the other hand, when it is determined in step S415 that the data in the selected field of record i is not similar to the data in the common subsequence field of record j (step S415: NO), the correspondence acquisition unit 202 stores the data in the basic database 36. The process proceeds to step S417 as it is without adding a line or writing data.
Then, after completing the processing for all the records extracted in step S412, the correspondence acquisition unit 202 ends the loop L32 (step S417), and completes the processing for all the records in the advertisement information field database 34. The loop L31 is terminated (step S418).
Thereafter, the process of FIG.

FIG. 17 is a flowchart showing a processing procedure in which the keyword assigning unit 203 associates keywords with the search key of the advertisement information master database 31, the search key of the product information master database 32, and the search key of the material file database 33. . The keyword assigning unit 203 performs the processing shown in FIG. 13 in step S104.
First, the keyword assigning unit 203 blanks the keyword database 37 as an initialization process of the keyword database 37 (step S501).

Next, the keyword assigning unit 203 starts a processing loop L41 for each record in the basic database 36 (step S511). Here, a record to be processed in the loop L41 is represented by a record k.
The keyword assigning unit 203 reads the advertisement ID from the record k. And the keyword provision part 203 searches the advertisement information master database 31 based on read advertisement ID, and reads a search key and a keyword from the predetermined field of the record specified by the said advertisement ID. In the present embodiment, the keyword assigning unit 203 reads an advertisement ID and a material ID as a search key, and reads an advertiser name, a product name, and a product name by kana writing as keywords.
Then, the keyword assigning unit 203 adds one line to the keyword database 37 and writes the read search key and keyword (step S512).

Further, the keyword assigning unit 203 reads the product ID from the record k. Then, the keyword assigning unit 203 searches the product information master database 32 based on the read product ID, and reads a search key and a keyword from a predetermined field of a record specified by the product ID. In the present embodiment, the keyword assigning unit 203 reads a product ID as a search key, and reads a manufacturer name, a product name, a product name by kana writing, and product type information as keywords.
Then, the keyword assigning unit 203 writes the read search key and keyword in the row added in step S512 (step S513).
Then, after completing the processing for all records in the basic database 36, the keyword assignment unit 203 ends the loop L41 (step S514).
Thereafter, the process of FIG.

FIG. 18 is a flowchart illustrating a processing procedure in which the content search apparatus 10 provides advertising content. The content search apparatus 10 performs the process shown in FIG. 13 after the process of the keyword assigning apparatus 20 described in FIGS. 13 to 17 is completed and the data of the keyword database 37 is generated.
First, the communication unit 101 waits for a keyword transmitted from the client device 9 (step S601). When the communication unit 101 receives a keyword from the client device 9, the communication unit 101 outputs the keyword to the content search key acquisition unit 102.

Next, the content search key acquisition unit 102 starts a processing loop L101 for each record in the keyword database 37 (step S611). Here, the record to be processed in the loop L101 is represented by a record m.
The content search key acquisition unit 102 calculates the similarity between each keyword stored in the record m and the keyword transmitted from the client device 9 (step S612), and performs similarity determination (step S613). For example, as described above for the field selection unit 201 and the correspondence relationship acquisition unit 202, the content search key acquisition unit 102 performs similarity determination based on the edit distance in the approximate character string matching. Alternatively, the similarity search may be performed by other methods such as the content search key acquisition unit 102 determining similarity based on the longest common character string length.

When it is determined that any of the keywords stored in the record m is similar to the keyword transmitted from the client device 9 (step S613: YES), the content search key acquisition unit 102 reads the search key from the record m, The content acquisition unit 103 searches the database based on the search key to acquire advertising content, and outputs the acquired content to the communication unit 101. In this embodiment, it is predetermined that the product name and the file stored in the material file database 33 are returned to the client device 9. Therefore, the content search key acquisition unit 102 reads the product ID from the record m, and outputs the read product ID to the content acquisition unit 103. The content acquisition unit 103 searches the product information master database 32 based on the product ID output from the content search key acquisition unit 102, acquires the product name, and outputs the product name to the communication unit 101 (step S614). In addition, the content search key acquisition unit 102 reads the material ID from the record m, and outputs the read material ID to the content acquisition unit 103. The content acquisition unit 103 searches the material file database 33 based on the material ID output from the content search key acquisition unit 102, acquires the content file, and outputs the content file to the communication unit 101 (step S615).
Thereafter, the process proceeds to step S616.

  On the other hand, when it is determined in step S613 that none of the keywords stored in the record m is similar to the keyword transmitted from the client device 9 (step S613: NO), the content search key acquisition unit 102 performs the advertisement The process proceeds directly to step S616 without acquiring the content.

Then, the content search key acquisition unit 102 ends the loop L101 after completing the processing for all the records in the keyword database 37 (step S616).
Then, the communication unit 101 transmits (replies) the product name output from the content acquisition unit 103 in step S614 and the content file output from the content acquisition unit 103 in step S615 to the client device 9 (step S621). ).
Thereafter, the process returns to step S601, and the process of FIG.

  As described above, the keyword assigning device 20 associates the search key for the advertisement content with the keyword. That is, the keyword assigning device 20 assigns the keyword stored in the product information master database 32 to the search key stored in the advertisement information master database 31, and the search key stored in the product information master database 32. The keywords stored in the advertisement information master database 31 are assigned. This eliminates the need to manually associate the search key with the keyword. Therefore, it is possible to reduce the burden on the data input operator and to associate the keyword with the data (advertisement content).

  In addition, the correspondence acquisition unit 202 performs the data similarity determination using the longest common substring of the selection field data of the advertisement information master database 31 and the selection field data of the product information master database 32. Even when a character string is added only to the data of the selected field, the similarity determination can be performed appropriately. As a result, the keyword assigning unit 203 can appropriately associate the search key with the keyword, and the content search device 10 can return an appropriate advertising content to the keyword transmitted from the client device 9. .

  Note that the scope of application of the present invention is not limited to the advertising content described above. The present invention can be applied to various data that can be associated with keywords.

A program for realizing all or part of the functions of the advertising content providing system 1 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system and executed. The processing of each unit may be performed as necessary. Here, the “computer system” includes an OS and hardware such as peripheral devices.
Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” refers to a portable medium such as a flexible disk, a magneto-optical disk, and a compact disk, and a storage device such as a hard disk built in the computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design changes and the like within a scope not departing from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 Advertisement content provision system 10 Content search apparatus 101 Communication part 102 Content search key acquisition part 103 Content acquisition part 20 Keyword assignment apparatus 201 Field selection part 202 Correspondence acquisition part 203 Keyword assignment part 31 Advertisement information master database 32 Product information master database 33 Material File Database 34 Advertising Information Field Database 35 Product Information Field Database 36 Basic Database 37 Keyword Database

Claims (5)

When any field of the first database is selected, the similarity of data between each field of the second database and the selected field is calculated, and the first is determined to have the highest similarity. A field selection section for selecting a field of the database of 2;
The first database including data determined to be similar by performing similarity determination between each data of the selected field of the first database and each data of the selected field of the second database A correspondence acquisition unit that associates the record of the second database with the record of the second database;
A keyword assignment unit that associates information stored in the first database with a keyword obtained from a record in the second database associated with a record including the information;
A keyword assigning device comprising: Each data of the selected field of the first database and each data of the selected field of the second database are character string data,
The correspondence acquisition unit obtains the longest common substring of each character string data of the selected field of the first database and each character string data of the selected field of the second database; Approximate character string collation between each character string data of the selected field of the first database or each character string data of the selected field of the second database and the longest common substring The similarity determination between each data of the selected field of the first database and each data of the selected field of the second database is performed based on the editing distance in the first database. Item 2. The keyword assigning device according to Item 1. A content database for storing content that is data provided to the client device;
A first database storing a content search key that is a search key for the content in the content database;
A second database;
A search key database;
A keyword assigning device;
A content search device;
A content providing system comprising:
The keyword assigning device includes:
When any field of the first database is selected, data similarity between each field of the second database and the selected field is calculated, and it is determined that the similarity is the highest. A field selection unit for selecting a field of the second database;
The first database including data determined to be similar by performing similarity determination between each data of the selected field of the first database and each data of the selected field of the second database A correspondence acquisition unit that associates the record of the second database with the record of the second database;
A keyword assigning unit that associates the content search key stored in the first database with a keyword obtained from the record of the second database associated with the record including the content search key;
A correspondence information writing unit for writing information indicating the association by the keyword assigning unit to the search key database;
Comprising
The content search device includes:
When a search term for searching for content stored in the content database is transmitted from the client device, a content search key acquisition unit that acquires the content search key by searching the search key database based on the search term When,
A content acquisition unit that acquires content by searching the content database based on the content search key acquired by the content search key acquisition unit;
A communication unit that transmits the content acquired by the content acquisition unit to the client device;
A content providing system comprising: A keyword assignment method for a keyword assignment device,
When any field of the first database is selected, the similarity of data between each field of the second database and the selected field is calculated, and the first is determined to have the highest similarity. A field selection step for selecting a field of the database of 2;
The first database including data determined to be similar by performing similarity determination between each data of the selected field of the first database and each data of the selected field of the second database A correspondence acquisition step for associating the record of the second database with the record of the second database;
A keyword assigning step for associating information stored in the first database with a keyword obtained from a record in the second database associated with a record including the information;
A keyword assigning method comprising: On the computer,
When any field of the first database is selected, the similarity of data between each field of the second database and the selected field is calculated, and the first is determined to have the highest similarity. A field selection step for selecting a field of the database of 2;
The first database including data determined to be similar by performing similarity determination between each data of the selected field of the first database and each data of the selected field of the second database A correspondence acquisition step for associating the record of the second database with the record of the second database;
A keyword assigning step for associating information stored in the first database with a keyword obtained from a record in the second database associated with a record including the information;
A program for running

Images