JP2018084983A - Content matching system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To match content expression of a sponsor with that of a content creator.SOLUTION: A content matching system includes: a recruiting-side expression item holding unit for holding a plurality of recruiting-side expression items which are provided in advance to make contents which is desired to recruit expressed to a content recruiting-side; an item-specific content recruiting-side expression information acquiring unit for acquiring item-specific content recruiting-side expression information by the content recruiting-side for each of the recruiting-side expression items; a creation-candidate-side appealing content item holding unit for holding a plurality of creation-candidate-side appealing content items which are provided in advance to make a creation-candidate-side express contents in which its own creation can be appealed; an item-specific content creation-candidate-side expression information acquiring unit for acquiring item-specific content creation-candidate-side expression information by the creation-candidate-side for each of the creation-candidate-side appealing content items; and a similarity degree calculation unit for calculating the degree of similarity and/or relevance with respect to the expression information acquired by both sides.SELECTED DRAWING: Figure 2

Description

The present invention relates to a technique for matching a content recruiting side such as an advertiser with a content creator side.

  When a content recruiting party such as an advertiser requests a content creator to produce, in recent years, content search processing or the like is performed on a network platform to determine the creator. In addition to searching for the content itself, the content of the content that the advertiser wants to produce is uploaded to the intermediary server, and the creator side refers to the content of the request and produces it. There has also been invented a system capable of matching processing in which an advertiser or the like selects a producer.

JP 2008-217423 A

  The technology of Patent Literature 1 relates to a matching system including an advertiser terminal, a creator terminal, and an intermediary server. The advertiser accumulates the contents of the advertisement production request on the mediator's server, and the creator refers to it. The creator uploads the produced work to the mediator's server, and the advertiser refers to this to examine the adopted work.

  As described above, when the advertiser accumulates the content of the advertisement production request in the mediator server, a plurality of producers can refer to the content of the request. In addition, by uploading content produced by a plurality of producers based on the requested content to the mediator server, it becomes possible for the advertiser to compare and examine the content and determine the producer to be employed.

  However, in the above-described technology, the creator first creates content based on the requested content, and the content recruitment side such as an advertiser compares the created content with a processing flow. That is, it is necessary for the producer to complete the content before the adoption is decided, and if it is not adopted, the created content is wasted, resulting in a problem that the producer's labor is lost.

  Based on the problems described above, the issue arises that the content recruitment side realizes an intermediary means that can perform selection or narrowing of content creators using existing information without waiting for content production. .

  In response to such a problem, the present invention provides a mediation server for appealing content items on the production candidate side, which is prior information such as what content the content creator can create in addition to the content requested by the content recruitment side. To accumulate. In addition, by calculating the similarity or / and relevance between the recruitment expression item that is the request content and the appealing content item on the production candidate side using a predetermined algorithm, the optimum content creator for the request content is determined or narrowed down. A content matching system capable of

  Specifically, this is a matching system between the content recruitment side and the content production candidate side, and the recruitment side expression that holds a plurality of recruitment side expression items provided in advance for the content recruitment side to express the content desired to be recruited An item holding unit, an item-specific content recruitment-side representation information acquisition unit that obtains item-specific content recruitment-side representation information, which is representation information made by the content recruitment side for each held recruitment-side representation item, and a content production candidate side In order to express content that can be appealed to itself, the content for each of the production candidate side appeal content items held by the production candidate side appeal content item holding unit that holds a plurality of production candidate side appeal content items is provided in advance. Itemized content system that is expression information made by the production candidate side The item-by-item content production candidate-side expression information acquisition unit that acquires the candidate-side expression information, the similarity or / and the degree of association between the acquired item-by-item content recruitment-side expression information and the acquired item-by-item content production candidate-side expression information. A content matching system having a similarity calculation unit for calculating is provided.

Further, in addition to the above feature, the similarity calculation unit may calculate the similarity between the word included in the acquired item-specific content recruitment-side expression information and the word included in the acquired item-specific content production candidate-side expression information, and / or Provided is a content matching system further comprising a brute force calculation means for calculating the degree of relevance as a brute force.

Further, in addition to the above features, the similarity calculation unit may include a pair of each word included in the acquired item-specific content recruitment-side expression information and each word included in the acquired item-specific content production candidate-side expression information. A content matching system for calculating similarity or / and relevance using the number of search results on the web including a combination of words.

  By using the content matching system as described above, the content-representation-side representation information by item, which is the content of the recruitment-side representation item, which is the content requested by the advertiser (content recruitment side), and the appeal on the content producer side (content production candidate side) It is possible to calculate the similarity or / and the degree of association with the item-by-item content production candidate side expression information that is the content of the production candidate side appeal content item. The content recruiting side that is the advertiser has an effect that it is possible to narrow down the content creator candidates that are more suitable for the content requested by the advertiser by comparing the calculated similarity or / and relevance. Further, since the production candidate side appealing content item is an existing one, there is an effect that the production candidate side who is the content creator does not have a waste of creating new content each time an application is made.

  Further, in the content matching system as described above, item-specific content recruitment side expression information and item-specific content production candidate side expression information are handled as items. Therefore, the item-by-item content recruitment side expression information acquisition unit can acquire the item-by-item content recruitment side expression information while keeping the advertiser anonymous. Similarly, the item-by-item content recruitment-side expression information acquisition unit can acquire item-by-item content recruitment-side expression information without revealing a specific product name. In this way, the production candidates can be narrowed down without revealing highly confidential information.

The figure which shows the outline of processing of the contents matching system Diagram showing functional configuration of content matching system The figure for demonstrating the offer side expression item holding part and the production candidate side appeal content item holding part The figure for demonstrating the content offer side expression information acquisition part according to item The figure for demonstrating the content production candidate side expression information acquisition part according to item The figure which shows an example of the calculation method of the similarity of a numerical item in a similarity calculation part, and / or a relevance degree The figure which shows an example of the calculation method of the similarity of a text item in a similarity calculation part, and / or a relevance degree The figure which shows an example of a process at the time of using a corpus among the calculation methods of the similarity of a text item or / and a relevance degree in a similarity calculation part. The figure which shows an example of the calculation method of the similarity or / and relevance by the combination of the numerical item and text item in a similarity calculation part The figure which shows an example of the method of calculating the total score of the similarity or / and the relevance in the similarity calculator The flowchart which shows an example of the flow of processing of a content matching system The flowchart which shows an example of the flow of a process of the similarity calculation method of a content matching system The figure which shows an example of the hardware constitutions of a content matching system

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. The present invention should not be limited to these examples, and can be implemented in various modes without departing from the scope of the present invention.

<Overview>

  FIG. 1 is a diagram showing an outline of processing of the content matching system of this embodiment. As shown in this figure, the advertiser on the content recruitment side uploads the recruitment side expression item (0102) that is the request content to the content matching system (0101). On the other hand, a plurality of content production candidate sides upload production candidate side appeal content items (0103) for appealing their own production content. On the content matching system (0101), the similarity or / and relevance of the two sets of items is calculated by a predetermined algorithm. The calculated similarity or / and relevance are arranged in descending order of similarity or / and relevance. This is the end of the matching process in this system. Thereafter, the content recruiting side may select the content creator by discussing the constraint conditions and the like directly or online with the content production candidate side.

  In the following, the function and processing flow of the content matching system of this embodiment and the contents of the hardware will be described in detail.

<Functional configuration>
FIG. 2 is a diagram illustrating a functional configuration of the content matching system according to the present embodiment. Note that the functional blocks of the system described below can be realized as a combination of hardware and software. Specifically, if a computer is used, a CPU (Central Processing Unit), a main memory, a bus, or a secondary storage device (a storage medium such as a hard disk drive, a non-volatile memory, a CD, a DVD, and the like) Read drive, etc.), input devices used for information input, printing equipment and display devices, other hardware components such as external peripheral devices, and interfaces for external peripheral devices, communication interfaces, and control these hardware For example, a driver program, other application programs, and a user interface application. Then, by CPU processing according to the program developed on the main memory, the data input from the input device or other interface and held in the memory or hard disk is processed and stored, An instruction for controlling the software is generated. Alternatively, the functional blocks of this system may be realized by dedicated hardware.

  The present invention can be realized not only as a system but also as a method. A part of the invention can be configured as software. Furthermore, a program used for causing a computer to execute such software and a recording medium on which the program is fixed are naturally included in the technical scope of the present invention (the same applies throughout this specification).

  The “content matching system” (0200) according to the present embodiment includes a “recruitment-side expression item holding unit” (0201), an “item-specific content recruitment-side expression information acquisition unit” (0202), and a “production candidate side appeal content”. An “item holding unit” (0203), an “item-specific content production candidate side expression information acquisition unit” (0204), and a “similarity calculation unit” (0205).

  The “recruitment-side expression item holding unit” (0201) has a function of holding a plurality of recruitment-side expression items that are provided in advance in order to cause the content recruitment side to express content desired to be recruited. The “recruitment side” refers to a person who includes an advertiser or the like but selects a content creator with the intention of producing content such as video. For example, in TV program production, producers, advertising agencies, agents, etc. of TV stations that request program production from production companies are also included in this “recruitment side”. The “recruitment side expression item” is a specification item of content that the content recruitment side wants to create. That is, the contents of the request for content to the production candidate who is the content creator on the content recruiting side are listed by item. Here, the content of the recruitment side expression item is item-specific content recruitment expression information acquired by the item-by-item content recruitment expression information acquisition unit described later. In the recruitment side expression item (0102) of FIG. 1, desired content length, target product, appeal point, meta keyword, reward, production period, and constraint conditions are listed as examples. FIG. 3 is a diagram for explaining the recruitment side expression item holding unit and the production candidate side appeal content item holding unit. As shown in this figure, the recruitment-side expression items (1) to (7) are defined in the respective database tables in association with the production candidate-side appeal content items (a) to (g). . In the table, recruitment-side expression items and production candidate-side appeal content items become fields of each database, and item-specific content recruitment-side expression information (described later) and item-specific content production candidate-side expression information (described later) are held as records. The held offer-side expression item is output to the similarity calculation unit (0205). The processing of the headquarters is realized by RAM, which is a volatile storage medium, and storage such as HDD, which is a non-volatile storage medium, and a DBMS (Data Management Management System) that is expanded on the RAM and executed by the CPU. Is done.

  The recruitment-side expression item and the production candidate-side appeal content item are not limited to the items listed in (1) to (7) and (a) to (g) of FIG. For example, items relating to copyright handling may be set and used for matching processing. Specifically, the condition that the copyright is transferred to the content recruiting side, the condition that the content production candidate side does not exercise the author's moral rights, and the like can be listed as items and used for the matching process.

  Further, the production candidate side appeal content item may be input after referring to the recruitment side expression item, or may be input regardless of whether or not there is a recruitment and the item may be held in the database. Further, the above two modes may be mixed depending on the item.

  The “item-specific content recruitment-side expression information acquisition unit” (0202) has a function of acquiring item-specific content-recruitment-side expression information, which is expression information made by the content recruitment side for each held recruitment-side expression item. Specifically, the contents-side content recruitment side expression information for the recruitment side items defined by the recruitment side expression item holding unit (0201) is received from the user and stored in the recruitment side expression item holding unit (0201). FIG. 4 is a diagram showing an example of an input screen which is a part of the item-by-item content recruitment side expression information acquisition unit (0202). As shown in this figure, input is received for items (1) to (7). This input screen is merely an example, and input may be accepted for more items. The item-by-item content recruitment-side expression information acquisition unit (0202) may acquire a plurality of item contents as item-by-item content recruitment-side representation information (the same applies hereinafter). The processing of the headquarters is realized by an item-by-item content recruitment-side expression information acquisition program that is expanded on the CPU and RAM and executed by the CPU.

  The “production candidate appeal content item holding unit” (0203) has a function of holding a plurality of production candidate side appeal content items provided in advance in order to make the content production candidate side express content that can be appealed to the production. “Production candidate side” refers to a plurality of content production candidates. “Production candidate side appeal content item” refers to an item to be selected by the content recruiting side in order for the production candidate to receive an order for content production. As shown in the production candidate appeal content item (0103) of FIG. 1, the content of the production candidate side appeal content item is a content length that is good, a good product, a good appeal point, a meta keyword of an existing work, Examples include desired rewards, desired production period, and desired conditions. The production candidate-side appeal content item that is held is output to the similarity calculation unit (0205). The processing of the headquarters is realized by storage such as RAM and HDD, and a DBMS (Database Management System) that is expanded and executed on the RAM.

  The item-by-item content production candidate side expression information acquisition unit (0204) acquires item-by-item content production candidate side expression information that is expression information made by the content production candidate side for each of the held production candidate-side appeal content items. It has the function to do. Specifically, for each production candidate side appeal content item defined by the production candidate side appeal content item holding unit (0203), item-specific content production candidate side expression information is received as input from the user, and the production candidate side appeal content item holding unit. (0203). FIG. 5 is a diagram showing an example of an input screen which is a part of the item-specific content production candidate side expression information acquisition unit (0204). As shown in this figure, input is accepted for items (a) to (g). This input screen is merely an example, and input may be accepted for more items. The item-by-item content production candidate side expression information acquisition unit may acquire item-by-item content production candidate side expression information for each of a plurality of content categories to be produced (still photos, moving images, graphics, etc.). The processing of the headquarters is realized by the CPU, the RAM, and the item-by-item content production candidate side expression information acquisition program developed on the RAM and executed by the CPU.

  The “similarity calculation unit” (0205) has a function of calculating the similarity or / and the relevance between the acquired item-by-item content recruitment-side expression information and the acquired item-by-item content production candidate-side expression information. Specifically, based on the correspondence between items in FIG. 3 ((1) and (a),... (7) and (g)), the item-by-item content recruitment expression information and the acquired item-by-item content production candidate. A set of side expression information is acquired, and a similarity or / and relevance is calculated. Here, the “similarity” is a relationship in which the closeness of concepts is the same in both directions like “mandarin orange” with respect to “orange”, and indicates the degree of closeness. Basically, it may indicate the closeness of both distances when a set of words is represented as a feature value in a multidimensional space. As will be described later, on a network based on a co-occurrence relationship, etc. It may indicate the closeness of the distance. “Relationship” does not have to be a two-way relationship, but is a one-way approach, such as a dependency relationship between sentences (for example, the relationship between “mountain” and “white” in “mountain is white”). Is a relationship that can be defined. As shown in the figure, the set of items is a numerical value, a text, or a combination of a numerical value and a text. Therefore, the calculation method of the similarity or / and the relevance is the optimum method for each. . A method for calculating similarity or / and relevance will be described below. The processing of the headquarters is realized by a CPU and RAM, and a similarity calculation program developed on the RAM and executed by the CPU.

  FIG. 6 shows an example of a method for calculating similarity or / and relevance of numerical items. As shown in this figure, the similarity or / and relevance of the numerical item is obtained by calculating the difference between the value on the content recruitment side and the value on the content production candidate side and dividing this by the value on the content recruitment side. . This is obtained for all the content production candidate sides that exist in plural, and the content production candidate side having the smallest value is adopted. The adopted value is substituted into a monotonically decreasing function of f (0) = 1 and f (1) = 0. Thereby, if the value calculated by substitution takes a positive value, this becomes the similarity or / and the relevance. On the other hand, if the calculated value takes a negative value, the degree of similarity and / or the degree of association may be zero. The case of taking a negative value represents a case where the desired reward amount on the content production candidate side exceeds the reward amount on the content recruitment side. In this case, a value of 0 may be adopted on the assumption that there is no similarity or / and relevance to indicate that the content recruiting side is in an advantageous position.

  FIG. 7 shows an example of a method for calculating the similarity or / and relevance of text items. In the calculation of the similarity or / and relevance of the text item, the similarity between the word included in the acquired item-specific content recruitment expression information and the word included in the acquired item-specific content production candidate expression information or / And a process of calculating the degree of relevance by brute force. As shown in this figure, it is compared whether there is a matching item content (word) between the item content on the content recruitment side and the item content on the content production candidate side. Here, “match” may include a case where a synonym exists. First, the contents of both items are compared. (I) If there is even one matching word, the similarity or / and relevance is set to 1. (Ii) When there is no matching word, for example, the similarity or / and the relevance are calculated as follows. A corpus is prepared, and calculation is performed based on the distance in the corpus from each word on the content recruitment side to the word on the content production candidate side. Here, the corpus is a linguistic database that has a sufficient number of words by statistically processing words contained in a large number of sentences, and the co-occurrence relations in the enormous number of sentences used for statistical processing. It is desirable that the word is determined to be relatively strong or structured so as to be relatively close to each other in the virtual space. It is assumed that the similarity or / and the relevance are higher as the distance in the corpus is shorter and lower as the distance in the corpus is longer. Specifically, the distance in the corpus of each combination of the content recruitment side word (m) and the content production candidate side word (n) is calculated, and the smallest one is adopted as the representative value. The similarity or / and relevance is a monotonically decreasing function of f (0) = 1 and f (1) = 0, which is a value obtained by dividing the representative value by the average distance to the word on the content production candidate side corresponding to the representative value. It is obtained by substituting.

  FIG. 8 is a diagram illustrating an example of calculation of similarity or / and relevance at a distance within a corpus among text item similarity or / and relevance calculation methods. For example, let us consider a case where the content recruiting side is (3) “Rice milling degree 80%” as the appeal point and the content production candidate side is (c) “Food” as the appeal point. “Rice milling degree 80%” can be divided into the words “milled rice”, “degree” and “80%”. The words in the corpus can form a network as shown in FIG. 8 by statistically processing the co-occurrence relationship or the dependency relationship. Specifically, for example, a sentence prepared in advance as a material is subjected to morphological analysis, and a Bigram indicating a co-occurrence relationship is created. A method of arranging the created bigrams in descending order of frequency and forming a network in which words having a frequency equal to or greater than a certain value are connected by links indicating co-occurrence relations as nodes can be considered. In the network in FIG. 8 created in this way, it is possible to reach the “food” node from the “rice mill” node in 3 hops. At this time, the value of the distance from “milled rice” to “food” can be “3”.

  It is also possible to determine whether the words (words) have a similar relationship by using a graph index. A graph index is a structured graph in which words are arranged in an n-dimensional space. For example, all words are defined in n dimensions. And it is judged that the word arrange | positioned in the position approximated in n-dimensional space has a close meaning. Then, the nearest words or the nearest k words in each word (node) are connected by a link. The closeness between words is represented by the number of links to reach each other on this link. This space can be configured as a virtual space of a computer, and the number of links can be obtained by a computer. The degree of similarity can be determined by setting a threshold value that determines that the number of links is within a maximum of the number of links.

  In addition to the above, as an example of a method for calculating the similarity or / and relevance of a text item, each word included in the acquired item-by-item content recruitment-side expression information and included in the acquired item-by-item content production candidate-side expression information The following may be mentioned in which similarity or / and relevance is calculated using the number of search results on the web including a pair of words with each word. That is, the text that is the item-by-item content recruitment side expression information on the content recruitment side is decomposed into words (m), and the text that is the item-by-item content production candidate side expression information is decomposed into words (n) Is calculated for how many combinations occur in the corpus for each combination. Then, this is divided by the number of words (m) in the item-by-item content recruitment expression information to obtain the co-occurrence rate. If no co-occurrence word is found in the corpus, the number of hits is calculated by performing a near search on the m words on the content recruitment side and the n words on the content production candidate side using a search engine. The co-occurrence rate may be calculated by dividing the number of search hits for each of the m words on the content recruiting side.

  FIG. 9 shows an example of a similarity or / and relevance calculation method based on a combination of numerical items and text items. This method mainly calculates the degree of similarity or / and relevance between (4) meta-keyword related to content and (d) meta-keywords of existing works (including cases where there are multiple pieces) and the number of accesses. Used for. First, for the text part of the item contents, the similarity or / and the relevance are calculated by the method described above. Regarding the numerical item, that is, (d) the number of accesses to the existing work uploaded so far, the weight is calculated using the total number of reproductions of the existing work on the moving image site or SNS (Social Networking Service). The similarity or / and the relevance calculated in the text item are multiplied by a weight wf set in advance according to the total number of reproductions. For example, the case where wf = 0.3 is set for 5000 or less and 5000 or less and 50,000 or less is set to wf = 0.6, and wf = 1.0 is set for 50,000 or more. Thus, the similarity or / and relevance is calculated by multiplying the similarity or / and relevance of the text item by the wf value. When there are a plurality of existing works that have been uploaded, the maximum value among the values calculated for each work is adopted as the similarity or / and the relevance.

  The degree of similarity or / and the degree of association calculated for each item associated in this way is calculated as a total score of the degree of similarity or / and degree of association by summing up. FIG. 10 is a diagram illustrating an example of a calculation method. As indicated by 1 in this figure, the sum total may be simply used as a total point, or a combination of items associated with each other may be weighted (represented by w) and multiplied. Further, as shown in 2., a cs value assigned to the content production candidate side may be included as a coefficient. The cs value is the degree of satisfaction that the content recruiting side gives to the content production candidate side after the production is completed. For example, it is conceivable to set any value of (0.2, 0.4, 0.6, 0.8, 1.0) as the five-level evaluation. This value is stored for each content production candidate side, used in the next selection, and updated every time the selection is made.

  The similarity production unit may have means for displaying on an output device such as a display on the content recruitment side by arranging the content production candidates in order from the highest overall score. Furthermore, the similarity production unit may have a notification means for notifying the narrowed production candidate side about the details of the contract conditions. Furthermore, the content matching system automatically generates draft contracts for content production using the narrowed-down production candidate-side item-specific content production candidate-side representation information and the content-recruitment-side content recruitment-side representation information as input. You may have a means to output or a means to perform creation assistance.

<Process flow>
<< Process flow of the entire system >>
FIG. 11 is a flowchart for illustrating an example of a processing flow when the content matching system according to the present embodiment is used. First, item-by-item content recruitment-side expression information is acquired (step S1101). Next, item-specific content production candidate side expression information is acquired (step S1102). Next, similarity or / and relevance are calculated (step S1103).
<< Flow of processing for calculating similarity or / and relevance >>
FIG. 12 shows in detail a process for calculating the degree of similarity and / or relevance among the processes when the content matching system according to the present embodiment is used. As shown in this figure, first, the similarity or / and relevance of a numerical item is calculated (step S1201). Next, the similarity or / and relevance of the text item is calculated (step S1202). Next, the similarity of the combination item of a numerical value and a text is calculated (step S1203). The processes so far may be interchanged or may be performed simultaneously. Next, a total point is calculated (step S1204). Finally, they are arranged and displayed in descending order of the total score (step S1205).

<Hardware configuration>
FIG. 13 is a diagram illustrating a hardware configuration of the content matching system according to the present embodiment. Hereinafter, description will be made with reference to this figure.

  As shown in this figure, the content matching system according to the present embodiment includes a “CPU (central processing unit)” (1301) that performs various arithmetic processes and a “storage” (1302) that is a nonvolatile storage such as an HDD. "RAM" (1303). Also, a “network I / F” (1304) for connecting to the Internet is provided. Then, they are connected to each other through a data communication path such as a “system bus” (1305) to perform transmission / reception and processing of information.

  The “RAM” (1303) provides a work area that is also a work area of the program at the same time as reading a program for performing various processes to be executed by the “CPU” (1301). A plurality of addresses are assigned to the “RAM” (1303), and programs executed by the “CPU” (1301) exchange data with each other by specifying and accessing the addresses. , Processing can be performed (the same applies throughout this specification).

  First, 1. An item-by-item content recruitment-side expression information acquisition program is expanded on the RAM (1303) and executed by the CPU (1301). The program acquires item-specific content recruitment-side expression information from the terminal via the network I / F (0604). The acquired content is temporarily stored in the RAM (0603), and is stored in the recruitment side expression item DB on the storage (1302) as the recruitment side expression item. Next, 2. An item-specific content production candidate side expression information acquisition program is expanded on the RAM (1303) and executed by the CPU (1301). The program acquires item-specific content production candidate side expression information from the terminal via the network I / F (1304). The acquired content is temporarily stored in the RAM (0603), and is stored in the production candidate side appeal content item DB on the storage (1302) as a production candidate side appeal content item. Next, the similarity calculation program is expanded in the RAM (1303) and executed by the CPU (1302). The program is based on a recruitment-side expression item DB and a production candidate-side appeal content item DB on storage, and a. Item-by-item content recruitment expression information and b. The item-by-item content production candidate side expression information is read, and the similarity or / and relevance between the two is calculated by the above-described arithmetic processing. The calculated similarity or / and relevance is temporarily stored in the RAM (1303), and when the similarity or / and relevance is calculated for all items, the total score is calculated by the above method. Do. This is executed for all item-by-item content production candidate side expression information stored in the production candidate side appeal content item DB, rearranged in descending order of total score, and stored on the RAM (1303). This is referred to from a terminal connected via the network I / F (1804).

0201: Recruitment-side expression item holding unit 0202: Item-specific content recruitment-side expression information acquisition unit 0203: Production candidate appealing content item holding unit 0204: Item-specific content production candidate-side expression information acquisition unit 0205: Similarity calculation unit

Claims (7)

A matching system between the content recruitment side and the content production candidate side,
A recruitment-side expression item holding unit that holds a plurality of recruitment-side expression items provided in advance in order to cause the content recruitment side to express the content desired to be recruited;
An item-specific content recruitment-side representation information acquisition unit that obtains item-specific content recruitment-side representation information that is representation information made by the content recruitment side for each held recruitment-side representation item;
A production candidate side appeal content item holding unit that holds a plurality of production candidate side appeal content items provided in advance to express content that can appeal to the content production candidate side;
An item-by-item content production candidate-side expression information acquisition unit for acquiring item-by-item content production candidate-side expression information that is expression information made by the content production candidate side for each of the held production candidate-side appeal content items;
A similarity calculation unit that calculates the similarity or / and the relevance between the acquired item-specific content recruitment-side expression information and the acquired item-specific content production candidate-side expression information;
A content matching system.   The similarity calculation unit calculates the degree of similarity or / and relevance between the words included in the acquired item-specific content recruitment-side expression information and the words included in the acquired item-specific content production candidate-side expression information The content matching system according to claim 1, further comprising a brute force calculation means.   The similarity calculation unit includes a pair of words including a combination of each word included in the acquired item-specific content recruitment-side expression information and each word included in the acquired item-specific content production candidate-side expression information. The content matching system according to claim 1 or 2, wherein the similarity or / and relevance is calculated using the number of search results above. The operation method of the matching system between the content recruitment side and the content production candidate side,
An item-specific content recruitment-side representation information acquisition step for obtaining item-specific content recruitment-side representation information that is representation information made by the content recruitment side for each held recruitment-side representation item;
Item-by-item content production candidate-side expression information acquisition step for acquiring item-by-item content production candidate-side expression information that is expression information made by the content production candidate side for each held production candidate-side appeal content item;
A similarity calculation step for calculating a similarity or / and a relevance between the acquired item-by-item content recruitment-side expression information and the acquired item-by-item content production candidate-side expression information;
A method of operating a content matching system.   In the similarity calculation step, the similarity or / and the relevance between the word included in the acquired item-by-item content recruitment-side expression information and the word included in the acquired item-specific content production candidate-side expression information is calculated brute force The operation method of the content matching system according to claim 4, further comprising a brute force calculation substep.   The similarity calculation step includes a pair of words including a combination of each word included in the acquired item-by-item content recruitment-side expression information and each word included in the acquired item-based content production candidate-side expression information. The operation method of the content matching system according to claim 4 or 5, wherein the similarity or / and the relevance are calculated using the number of search results above. An operation program for the matching system between the content recruitment side and the content production candidate side,
An item-specific content recruitment-side representation information acquisition step for obtaining item-specific content recruitment-side representation information that is representation information made by the content recruitment side for each held recruitment-side representation item;
Item-by-item content production candidate-side expression information acquisition step for acquiring item-by-item content production candidate-side expression information that is expression information made by the content production candidate side for each held production candidate-side appeal content item;
A similarity calculation step for calculating a similarity or / and a relevance between the acquired item-by-item content recruitment-side expression information and the acquired item-by-item content production candidate-side expression information;
Is a program that causes the content matching system, which is a computer, to execute.

Images