JP5229504B2 - Advertisement presenting method, advertisement presenting system and program - Google Patents

Description

  The present invention relates to an advertisement presentation method, an advertisement presentation system, and a program.

  Affiliate advertising is a mechanism whereby the website operator pays the advertising fee to the website operator when the website operator directs the page viewer to the URL created by the advertiser (mainly the URL of the page where the product is posted). That's it. It is common for an advertiser to place an advertisement through an advertising agency, but the advertiser and the advertising agency may be the same. Typical examples of affiliate advertisements include Google adsense and Amazon affiliate advertisements.

  Such advertisements are mainly classified into two types according to the advertisement display method. Content-linked type and URL pasting type.

  In the content-linked type, the web component (mainly Java (registered trademark) Script) specified by the advertising agency is pasted on the web page operated by the website operator, and the advertisement that matches the content of the web page is automatically created. It is a method to display in the web component. Google's adsense is representative.

  The URL pasting type is a method in which an advertisement of a product recommended by an advertisement publisher is displayed by explicitly pasting a URL to a product advertisement by a website operator. Amazon affiliate ads are typical.

  In the former content-linked type, advertisements are automatically displayed, so Web site operators need only paste Web components. However, it is not possible to display only the advertisement that the page viewer wants to see, for example, the cheapest item in the same product group.

  In the latter URL pasting type, the recommended product of the website operator can be clearly displayed. Page viewers are not always interested in the product.

  For these reasons, a technique for selecting an advertisement using user attribute information of a page viewer has been disclosed.

  For example, Patent Document 1 discloses a technique for comparing advertisement content information with user attribute information and delivering an advertisement that matches the user information.

Patent Document 2 discloses a technique for displaying an advertisement that at least matches user attribute information by using an advertisement browsing history.
Japanese Patent Application No. 2001-314681 Japanese Patent Application No. 2002-327076

  Such a method has a problem that a large amount of user attribute information must be held in advance. Furthermore, even with the technique disclosed in Patent Document 2, there is a problem that a suitable advertisement cannot be displayed in an initial state where the user browsing history is not accumulated.

  Therefore, the present invention has been invented in view of the above problems, and an object of the present invention is to provide a technique capable of selecting an advertisement suitable for a user.

  The present invention that solves the above-described problem is an advertisement presenting method that calculates the degree of fitness of an advertisement related to an object with respect to the user based on reputation information including an expression for evaluating the object and user attribute information. The advertisement to be presented to the user is selected based on the fitness.

  The present invention that solves the above-described problem is an advertisement presentation system, which is adapted to calculate the degree of adaptation of an advertisement related to an object to the user based on reputation information including an expression for evaluating the object and user attribute information A degree calculation unit; and an advertisement selection unit that selects an advertisement to be presented to the user based on the degree of matching.

  The present invention that solves the above-described problem is a program for an information processing apparatus, and calculates a degree of fitness for an advertisement related to an object based on reputation information including an expression for evaluating the object and user attribute information. The information processing apparatus is configured to execute a fitness level calculation process to be performed and an advertisement selection process to select an advertisement to be presented to the user based on the fitness level.

  Since this invention is comprised so that an advertisement may be selected based on reputation information, the advertisement of the target object corresponding to the attribute and evaluation which the user is paying attention to can be selected.

FIG. 1 is a block diagram of an advertisement presentation system in the first embodiment. FIG. 2 is a diagram illustrating an example of the reputation information database 4. FIG. 3 is a diagram illustrating an example of the advertisement product database 5. FIG. 4 is a diagram showing an example of the click information database 6. FIG. 5 is a diagram showing an example displayed by a Web browser or the like. FIG. 6 is a flowchart of the operation of the advertisement presentation system in the first embodiment. FIG. 7 is a block diagram of an advertisement presentation system in the second embodiment. FIG. 8 is a diagram showing an example in which each advertisement in the advertisement product database 5 is weighted. FIG. 9 is a flowchart of the operation of the advertisement presentation system in the second embodiment. FIG. 10 is a block diagram of an advertisement presentation system in the third embodiment. FIG. 11 is a diagram illustrating an example of the product information database 20. FIG. 12 is a flowchart of the operation of the advertisement presentation system in the third embodiment. FIG. 13 is a block diagram of an advertisement presentation system in the fourth embodiment. FIG. 14 is a block diagram of an advertisement presentation system in the fifth embodiment. FIG. 15 is a diagram showing an example of the reputation information database 4 in the fifth embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Input part 2 Advertisement selection part 3 Advertisement display part 4 Reputation information database 5 Advertisement product database 6 Click information database

  The present invention selects reputation information including an expression to be evaluated with respect to attribute information of a user browsing a web page or the like from reputation information including an expression for evaluating the object, and selects the object corresponding to the selected reputation information. From the advertisement, an advertisement to be displayed on the Web page or the like is selected.

  Here, a specific product or service is called an object. For example, the product name “◯ × PC”, the service name “Δ □ service”, and the like. Reputation information is information including an expression indicating contents for evaluating the object, for example, information including expressions indicating evaluation contents such as “good”, “bad”, and “large”. In the following description, an expression for evaluating an object is referred to as an evaluation expression.

  The reputation information may include an attribute expression indicating the attribute of the object. An attribute expression is a word indicating the characteristics of an object. For example, if a certain personal computer (hereinafter also simply referred to as a personal computer) is an object, it is a word such as “screen” or “weight”. Attribute expressions can be hierarchically connected. For example, from an input sentence (natural text) that says “○ × Computer screen size is good”, [object “○ × PC”, attribute expression “screen”, attribute expression “size”, evaluation expression Reputation information “good”] is extracted.

  Further, the evaluation expression may be given an evaluation polarity indicating whether it is a positive expression or a negative expression. For example, since the evaluation expression “good” is a positive expression, a positive polarity value is given, and since the evaluation expression “bad” is a negative expression, a negative polarity value is given.

  Such reputation information can be collected by a blog or a bulletin board on the network. It should be noted that reputation information can be collected even if the object is not specified in the natural text text written on the blog or bulletin board, and the attribute expression is omitted. For example, on bulletin boards for cameras, even if the target object is not specified in the natural text, it can be seen that the target subject is a camera, and the attribute expression is omitted. It is because the sentence of can guess the attribute.

  Therefore, in the present invention, the information to be collected may be a combination of the object, the attribute expression, and the evaluation expression, or a combination of the attribute expression and the evaluation expression. A combination of the two expressions may be used, and it is only necessary that at least the object and the evaluation expression can be finally databased.

  Incidentally, Japanese Patent Application Laid-Open No. 2006-236379 discloses a technique for extracting reputation information from natural sentences.

  The user attribute information is information about the user, for example, not only the user's own information such as the user's age, gender, and address, but also information related to a specific product or service such as a computer of company A, or simply a computer. This is information on general products and services such as television, and things, services, and hobbies of interest to users such as travel and reading.

  The user attribute information is acquired from information already stored in the information processing apparatus, past search history by the user, reputation information written by the user, and the like.

  Next, as a method for selecting an advertisement to be presented to a user, reputation information including an expression for evaluating user attribute information is selected from the collected reputation information.

  For example, when the user's gender “female” and age “25 years” are obtained from the user's attribute information, reputation information including expressions to be evaluated for the gender “female” and age “25 years” is selected. At this time, the reputation information such as “A tour” is for women and “B book” is for 20s is selected as the reputation information. Then, an object corresponding to the selected reputation information, that is, an advertisement of “A tour” and “B book” is presented to the user. In addition, when there is a limit to the number of advertisements to be presented, advertisements are narrowed down according to predetermined conditions.

  In this way, an advertisement suitable for the user can be selected and presented.

  Embodiments of the present invention will be described below.

<First Embodiment>
A first embodiment will be described.

  In the first embodiment, an example will be described in which an advertisement suitable for a user is displayed on the Web page for a user who is browsing a Web page. In the first embodiment, description will be made assuming that user attribute information is acquired from the user's reputation information.

  Referring to FIG. 1, the advertisement presentation system in the first exemplary embodiment of the present invention includes an input unit 1, an advertisement selection unit 2, an advertisement display unit 3, a reputation information database 4, an advertisement product database 5, A click information database 6.

  The reputation information database 4 is a database that stores reputation information. An example of the reputation information database 4 is shown in FIG. The example shown in FIG. 2 is a database in which product names (objects), attribute expressions, evaluation expressions, polarities, and user IDs in which reputation information is written are associated. This is an example, and two product names and evaluation expressions, or two attribute expressions and evaluation expressions may be used. The degree of polarity is a value from “1” to “−1”. The closer to “1”, the more positive the reputation information, and the closer to “−1”, the negative the reputation information. This is only an example, and other values such as “100” to “0” may be used. Numerical values may be handled discretely and may be indicated by symbols such as “O” and “X”. It may be divided into A record with an unknown user ID may be included. Note that the reputation information stored in the reputation information database 4 is collected from a blog or a bulletin board on the network as described above.

  The advertisement product database 5 is a database that stores product information of advertisement targets. An example is shown in FIG. In the example shown in FIG. 3, an advertisement ID, a product name, a keyword, an advertisement text used for displaying an advertisement, and a link destination URL when an advertisement is clicked are associated with each other. This is an example, and the keyword may be automatically extracted from the advertisement text. Further, it may include advertisement display image data, moving image data, and the like.

  The click information database 6 is a database that stores how many times each advertisement is clicked on which site. Using this information, a financial procedure is performed such as paying an advertising fee to the Web page operator. An example is shown in FIG. In the example shown in FIG. 4, the site ID, the advertisement ID, and the number of clicks are associated with each other. This is an example, and the site ID may be an ID for each Web page operator.

  The advertisement selection unit 2 receives from the input unit 1 the user ID of the user who is browsing the page, the site ID that operates the page, and the number of display advertisements. Then, the user's reputation information corresponding to this is retrieved from the reputation information database 4 from the user ID of the user who is browsing, and the attribute expression that the user is paying attention to is extracted from the user's reputation information as user attribute information get. Then, as for the acquired attribute information of the user, an advertisement of a product having a positive reputation and a low negative reputation is selected as many as the display number, the advertisement of the selected product is searched from the advertisement product database 5, and the advertisement ID of the advertisement is selected. To get.

  Here, although it is an advertisement selection method, the adaptability of the advertisement with respect to the browsing user is calculated and used. The advertisement suitability is a large value for a product that has a lot of positive reputation information and a small amount of negative reputation information for the attribute that the user is paying attention to. In order to calculate this degree of conformity, attribute information, evaluation expression, and polarity of reputation information are used.

For example, the following advertisement adaptability is calculated for each product advertisement.
Goodness of advertisement of product p = Σ Aku * Akp * Ako
Here, Aku is the number that the attribute expression Ak appears in the user's reputation information, Akp is the number that the attribute expression Ak appears in the total reputation information of the product p (including the user's own reputation information), and Ako Is the average value of the polarity of the attribute expression Ak in the overall reputation information of the product p. Then, the sum of those calculated for each attribute expression Ak is used as the fitness of the advertisement of the product p.

  In this way, the advertisement suitability can be calculated. This is because, if there is a lot of Aku, it indicates that the degree of attention of the user of the attribute expression is high. Further, by considering the degree of polarity with Ako, the positive attribute expression has a higher fitness level, and the negative attribute expression has a lower fitness level.

  For example, when the degree of matching between the user “user B” and the advertisement of the product “product A” shown in FIG. 2 is calculated, the number that the attribute expression Ak “screen” appears in the reputation information (attribute information) of the user B Is 1 (Aku = 1), the number of attribute expression Ak “screen” appearing in all the reputation information of product A is 2 (Akp = 2), and the polarity average value of the attribute expression Ak “screen” of product A is 1 (Ako = 2/2 = 1). Accordingly, Aku * Akp * Ako (1 * 2 * 1) of the attribute expression Ak “screen” of the product A of the user B is 2.

  Similarly, the attribute expression Ak “brightness” of the product A is Aku = 1, Akp = 0, Ako = 0, and thus 0.

And, for all attributes, adding the calculated values together, the ad fitness of product A is
Σ Aku * Akp * Ako = 2 + 0 = 2
It becomes.

  This calculation formula is an example, and may be divided by the number of attributes of each product, may be aggregated for each pair of attribute expression and evaluation expression, or may be aggregated for each evaluation expression.

  The advertisement selection unit 2 selects advertisements corresponding to the number of display advertisements from the calculated advertisement with a large degree of matching, and passes the advertisement ID and site ID of the selected advertisement to the advertisement display unit.

  The advertisement display unit 3 receives the advertisement ID and the site ID from the advertisement selection unit 2, reads the advertisement text corresponding to the advertisement ID from the advertisement product database 5, and advertises the read advertisement text in the advertisement display area of the site ID page. Display a sentence. FIG. 5 is an example displayed by a Web browser or the like. When the user clicks on the displayed advertisement, the number of clicks is written or updated in the click information database based on the site ID and the advertisement ID.

  Next, operation | movement of the advertisement presentation system in this Embodiment is demonstrated in detail using the flowchart of FIG.

  First, the user ID and the number of display advertisements are acquired from the input unit 1 (Step 100).

  Next, the reputation information database 4 is referred to, and reputation information corresponding to the user ID is acquired (Step 101).

  Next, the fitness of each advertisement is calculated based on the reputation information of the user and the reputation information of other products (Step 102).

  Finally, advertisements corresponding to the number of displayed advertisements are selected from advertisements with a high degree of fitness (Step 103), and the advertisement ID and site ID of the advertisement are passed to the advertisement display unit 3 (Step 104). At that time, advertisements of products evaluated by the user may be excluded.

  The advertisement display unit 3 displays an advertisement based on the advertisement ID and the site ID (Step 105).

In this embodiment, since the advertisement selection unit 3 is configured to select an advertisement based on reputation information, it is possible to select an advertisement corresponding to the attribute expression or evaluation expression that the user is paying attention to. .
<Second Embodiment>
A second embodiment will be described.

  In the second embodiment, in addition to the first configuration, an advertisement that matches the browsing page is selected by weighting an advertisement that includes the keyword of the browsing page that the user browses. .

  Referring to FIG. 7, in addition to the first embodiment, the second embodiment of the present invention includes a keyword selection unit 10 and an advertisement weighting unit 11 that weights advertisements based on the extracted keywords. The point that is added is different.

  The keyword selection unit 10 receives the web page text viewed from the input unit 1, extracts keywords from the web page, and passes the extracted keywords and the number of appearances of the keywords to the advertisement weighting unit 11.

  The advertisement weighting unit 11 receives the keyword extraction result from the keyword selection unit 10, and if these keywords are in the keyword or the advertisement text of the advertisement product database 5, the number of appearances of the keyword with respect to the advertisement corresponding to the keyword. Weighting is performed. For example, as a keyword extraction result from the keyword selection unit 10, information on the number of appearances of the keyword “screen” is 2, the number of appearances of the keyword “brightness” is 3, and the number of appearances of the keyword “ecology” is 1 is received. If the keyword “screen” is included in the keyword or ad text, a weight of 2 is applied, and if the keyword “brightness” is included in the keyword or ad text, the weight is 3 and the keyword “ecology” is applied. A weight of 1 is applied to an advertisement included in a keyword or an advertisement text. When the weight of each advertisement is finished, the final weight of each advertisement is normalized by a value from 0 to 1 by performing a process such as dividing the maximum weight value. FIG. 8 is a diagram showing an example in which each advertisement in the advertisement product database 5 is weighted. The normalization process described above is not always necessary.

  The advertisement selection unit 2 receives from the input unit 1 the user ID of the user who is browsing the page, the site ID that operates the page, and the number of display advertisements. Then, the user's reputation information corresponding to this is retrieved from the reputation information database 4 from the user ID of the browsing user, the attribute expression that the user is paying attention to is extracted from the user's reputation information, and the advertisement of each product is displayed. The fitness is calculated in the same manner as in the first embodiment described above. Furthermore, in the second embodiment, the advertisement adaptability calculated in the same manner as in the first embodiment is multiplied by the weight of each advertisement, and the product is used as the advertisement adaptability. Select as many ads as the number of displayed ads.

  Specifically, the following formula is calculated.

Goodness of advertisement of product p = Wp * (Σ Aku * Akp * Ako)
Here, Wp is the weight of the advertisement.

  This calculation formula is an example, and for example, the following values may be used.

Goodness of advertisement of product p = Wp * (Σ Aku * Akp * Ako + Σ Kku * Kkp)
Here, Kku is the number of keywords Kk in the advertisement product database 5 for the products evaluated by the user, and Kkp is the number of keywords Kk in the advertisement product database for the product p. Thereby, it becomes a big value with respect to the goods which have the same keyword as the goods which the user evaluated.

  Finally, an advertisement of a product corresponding to the number of display advertisements is selected from those having a high advertisement adaptability, and the advertisement ID and site ID of the advertisement are passed to the advertisement display unit 3. At that time, advertisements of products evaluated by the user may be excluded.

  The operation of the advertisement presenting system in the present embodiment will be described with reference to the flowchart of FIG. Detailed description of the same parts as those in the first embodiment will be omitted.

  First, the keyword selection unit 10 receives a Web page from the input unit 1 and extracts a keyword from the Web page (Step 200). As an extraction method, keyword extraction is performed by using an existing method such as TF-IDF or by preparing a keyword candidate database in advance and selecting from candidates.

  Next, the advertisement weighting unit 11 receives the keyword extraction result from the keyword selection unit 10, and weights the advertisement corresponding to the keyword if the keyword is in the keyword or the advertisement text of the advertisement product database 5. (Step 201).

  The advertisement selection unit 2 refers to the advertisement product database 5 and calculates the fitness of each advertisement based on the user's reputation information, the reputation information of other products, and the weight of the advertisement (Step 202).

  Finally, advertisements corresponding to the number of displayed advertisements are selected from advertisements with a high degree of fitness (Step 103), and the advertisement ID and site ID of the advertisement are passed to the advertisement display unit 3 (Step 104). At that time, advertisements of products evaluated by the user may be excluded.

In the present embodiment, the keyword selection unit 10 selects a keyword from the Web page, the advertisement weighting unit 11 weights the advertisement based on the keyword, and the advertisement selection unit 2 considers the advertisement weight. Since the calculation is performed, an advertisement closer to the content of the Web page can be selected.
<Third Embodiment>
A third embodiment will be described.

  In the third embodiment, in addition to the first configuration, an advertisement for a product having the same specifications as the product evaluated by the user is selected.

  Referring to FIG. 10, the third embodiment of the present invention is different from the first embodiment in that a product information database 20 is further included.

  The product information database 20 is a database in which product names, attributes, and attribute values are associated with each other, and is a database that stores detailed information about products. An example is shown in FIG. Here, the attribute is a word indicating a component or feature of the product, and is, for example, “CPU”. The attribute value is a specific value of the component, such as “1.2 GHz”. These pieces of information may be input by an advertiser or automatically acquired from a web page using a technique such as web wrapper induction.

  The advertisement selection unit 2 receives the user ID, the site ID, and the number of displayed advertisements from the input unit 1, refers to the reputation information database 4, the advertisement product database 5, and the product information database 20, and calculates the advertisement suitability.

In particular,
Goodness of advertisement of product p = ΣAku * Akp * Ako + ΣZku * Zkp * Zko
Calculate from

Where the formula, Σ
Aku * Akp * Ako is the degree of fitness described in the first embodiment.

  Σ Zku * Zkp * Zko indicates the degree to which the specification of the product p is similar to the product group evaluated by the user, and Zku is an attribute-attribute value pair Zk in the product information database 30 for the product evaluated by the user. Zkp is the number of attribute-attribute value pairs Zk in the product information database 30 for the product p, and Zko is the average of the polar values of the products having the attribute-attribute value pair Zk among the products evaluated by the user It is. Here, the attribute-attribute value pair Zk refers to the attribute-attribute value pair of each record in the product information database 30. For example, if the record has an attribute “CPU” -attribute value “1.2 GHz”, These two “CPU”-“1.2 GHz” are one attribute-attribute value pair.

  For example, if user B is evaluating product A and product B, the degree of similarity between the attribute-attribute value pair of product C and the attribute-attribute value pair of product A and product B is calculated. The following is calculated and added for each attribute-attribute value of product A and product B.

(Attributes in product A and product B-number of attribute values) x (attributes in product C-number of attribute values) x (average polarity of products evaluated by users with attribute-attribute values)
For example, in the example of FIG. 11, “CPU-1 GHz” of the product A is 1 * 1 * 1. Similarly, other attribute-attribute value pairs are calculated and added.

  Thereby, it becomes a big value with respect to the goods which have the same attribute as the goods which the user evaluated positively.

  The operation of the advertisement selection unit 2 will be described with reference to the flowchart of FIG. Detailed description of the same parts as those in the first embodiment will be omitted.

  The advertisement selection unit 2 receives the user ID, the site ID, and the number of displayed advertisements from the input unit 1, refers to the reputation information database 4, the advertisement product database 5, and the product information database 20, and uses the same specifications as the products evaluated by the user. The degree of adaptation of the advertisement for selecting the advertisement of the product is calculated (Step 300).

In the present embodiment, the advertisement selection unit obtains the product detail information from the product information database, and products similar to the products positively evaluated by the user have a higher fitness, so the displayed advertisement is also clicked. It becomes easy.
<Fourth embodiment>
A fourth embodiment will be described.

  Referring to FIG. 13, the fourth embodiment of the present invention is different from the first embodiment in that a user preference complementing unit 30 is included.

  The user preference complementing unit 30 receives the user ID from the input unit 1, and acquires information about a user similar to the user from the reputation information database 4. Furthermore, information on similar users is referred to, and attribute expressions and evaluation expressions that the user is not paying attention are complemented. Accordingly, it is possible to use the reputation information that is attracting attention by users with similar tastes, and to expand the range of advertising products.

  Specifically, the user complementing unit 30 operates as follows.

  First, a user ID is received from the input unit 1. Next, the reputation information database 4 is referred to and a user similar to the user is specified. A similar user is a user who is similar in how to evaluate a product, for example, a user who is performing the same evaluation on the same attribute for the same product. The product is specified based on the product name, attribute expression, and evaluation expression in the reputation information database 4.

Specifically, for each user, the user similarity indicating the degree of similarity with other users is calculated as in the following equation.
User similarity (u, v) = Σ (number of reputation information of products evaluated by both user u and user v) + Σ (number of reputation information of attribute expressions evaluated by both user u and user v) + Σ ( Number of reputation information of evaluation expressions evaluated by both user u and user v)

  For example, when calculating the user similarity between the user A and the user B, in the example of the reputation information database 4 in FIG. 2, the product evaluated by both the user A and the user B is “product A”, and its reputation The number of information is 2 (reputation information of “product A” evaluated by user A = 1, reputation information of “product A” evaluated by user B = 1). Also, the attribute expression evaluated by both the user A and the user B is “screen”, and the number of reputation information is 2 (reputation information of “screen” evaluated by the user A = 1, “screen evaluated by the user B” "Reputation information = 1). Moreover, there is no reputation information number of the evaluation expression which both the user A and the user B evaluated.

  Therefore, the user similarity between user A and user B is 2 + 2 + 0 = 4.

  By calculating in this way, the degree of similarity of users having the same product name, attribute expression, and evaluation expression records is increased. However, this calculation method is an example, and the number of records of combinations of product names and attribute expressions may be included.

  As a result, several similar users with high user similarity are specified. The number of persons to be specified is preset by the system.

  Next, the system sends the specified similar user's reputation information and the user's reputation information to the advertisement selection unit 2.

  The advertisement selection unit 2 receives the specified similar user's reputation information and the user's reputation information, and uses these reputation information to calculate the fitness as in the first embodiment.

In the present embodiment, the user preference complementing unit specifies a user similar to the browsing user, and the advertisement selection unit calculates the degree of matching including the reputation information of the similar user, thereby expanding the range of the advertisement.
<Fifth embodiment>
Referring to FIG. 14, the fifth embodiment of the present invention is different from the first embodiment in that a time weighting unit 40 is included. Another difference is that a time field can be added to the reputation information database 4.

  Reputation information may change over time. The reputation of the product is expected to change gradually over time. For example, the reputation of a company at a certain time varies depending on the previous news, settlement information, etc. Therefore, when dealing with reputation information, it is necessary to consider the passage of time, such as giving weight to the polarity in recent reputation information. The purpose of this embodiment is to perform reputation information and advertisement selection in consideration of time.

  The reputation information database 4 has a time field added as compared to the first embodiment. An example is shown in FIG. The time field indicates the time when reputation information was written.

  The time weighting unit 40 weights each record in the reputation information database using time. For example, the weight is set to 1 for the reputation information written in a certain period (such as the last three months), and the weight is set to 0 for others. For example, if the product is a personal computer, the model is changed every quarter, so only the reputation information evaluated within the latest three months is adopted. This is an example, and for example, a weight may be given every month, a difference between the current time and reputation information writing time may be calculated, and the reciprocal of the obtained time difference may be used as the weight.

The advertisement selection unit 2 calculates the fitness level as in the first embodiment. However, the weight of each record in the time weighting unit 40 is taken into consideration.
Goodness of fit of product p = Σ Aku '* Akp' * Ako '
Here, Aku 'is obtained by adding the weight of each record of the attribute expression Ak in the user's reputation information. Akp ′ is obtained by adding the weight of each record of the attribute expression Ak in the reputation information of the product p. Ako ′ is an average value of (polarity * weight) of the attribute expression Ak in all reputation information of the product p.

  In the present embodiment, the time weighting unit weights the reputation information so that the freshness of the reputation information can be considered.

  In the above-described embodiment, an example in which each unit is configured by hardware has been described. However, main components such as the advertisement selection unit 2 may be configured by an information processing apparatus that operates by a program.

  In the embodiment described above, an advertisement presenting method, an advertisement presenting system, and a program having the following characteristic configuration are shown.

  The first advertisement presenting method is an advertisement presenting method, which calculates the degree of suitability of an advertisement related to an object to the user based on reputation information including an expression for evaluating the object and user attribute information, The advertisement presenting method is characterized in that an advertisement to be presented to the user is selected based on the fitness.

  A second advertisement presenting method is characterized in that, in the advertisement presenting method, user attribute information is acquired from user reputation information.

  A third advertisement presenting method is characterized in that, in the above-described advertisement presenting method, the degree of suitability of the advertisement related to the object to the user is calculated using reputation information other than the user.

  The fourth advertisement presenting method is characterized in that reputation information is collected from a bulletin board or a blog on the network.

  According to a fifth advertisement presenting method, in the advertisement presenting method described above, the degree of conformity is that the degree of conformity of an advertisement related to an object of reputation information including a positive evaluation expression with respect to the attribute information of the user is increased. Features.

  A sixth advertisement presenting method is the advertisement presenting method described above, wherein the fitness is related to an object of reputation information that includes many positive evaluation expressions and few negative evaluation expressions with respect to the user attribute information. It is characterized by a high degree of advertisement adaptability.

  The seventh advertisement presenting method is the above-described advertisement presenting method, wherein the number of attribute expressions Ak indicating the characteristics of the object p appears in the user's reputation information is Aku, and the number of attribute expressions Ak appears in all the reputation information of the object p. Is Akp, and the average value of the degree of polarity representing the degree of affirmation or denial of the attribute expression Ak in all reputation information of the product p is Ako, Aku * Akp * Ako for each attribute expression Ak of the object p And Σ Aku * Akp * Ako, which is the sum of those calculated for each attribute expression Ak of the object p, is used as the fitness of the advertisement of the object p.

  An eighth advertisement presenting method is the advertisement presenting method according to the above advertisement presenting method, wherein weighting is applied to an advertisement including a keyword of a page on which an advertisement is presented, and an advertisement to be presented to the user based on the weighted advertisement conformity. It is characterized by selecting.

  A ninth advertisement presenting method is the above-described advertisement presenting method, wherein the object A and the object B are based on reputation information obtained by evaluating the object A among user attribute information and information on the object B. The similarity of the advertisement for the object B is calculated based on the similarity of the object between the object A and the object B and the fitness of the advertisement for the object B. It is characterized by calculating.

  A tenth advertisement presenting method is the above-described advertisement presenting method, wherein a user similarity between users is calculated based on the reputation information, and based on the user similarity, the advertisement information using similar user reputation information is calculated. It is characterized by calculating the fitness.

  An eleventh advertisement presenting method is characterized in that, in the advertisement presenting method, weighting related to the written time is performed on the reputation information, and the weighting is reflected on the fitness of the advertisement.

  The first advertisement presenting system is an advertisement presenting system, and calculates the degree of suitability of the advertisement related to the object to the user based on the reputation information including the expression for evaluating the object and the attribute information of the user. It has a fitness level calculation means and an advertisement selection means for selecting an advertisement to be presented to the user based on the fitness level.

  The second advertisement presenting system is characterized in that, in the advertisement presenting system, the fitness level calculating means obtains user attribute information from user reputation information.

  The third advertisement presenting system is characterized in that, in the advertisement presenting system, the suitability calculation means calculates the suitability of the advertisement related to the object to the user using reputation information other than the user.

  The fourth advertisement presenting system is characterized in that, in the advertisement presenting system, the suitability calculation means collects reputation information from a bulletin board or a blog on a network.

  In a fifth advertisement presenting system, in the above advertisement presenting system, the fitness level calculation means increases the fitness level of an advertisement related to an object of reputation information including a positive evaluation expression for the attribute information of the user. It is characterized by calculating as follows.

  In a sixth advertisement presenting system, in the advertisement presenting system, the fitness level calculating means includes a negative rating expression including a lot of positive rating expressions for the user attribute information. It is characterized in that the calculation is performed so that the fitness of the advertisement related to the object having a small reputation information is high.

  In a seventh advertisement presenting system, in the above advertisement presenting system, the fitness calculation means sets Aku as the number of attribute expressions Ak indicating the characteristics of the object p appearing in the user's reputation information, and the attribute expression Ak is the object p. Akp is the number that appears in all the reputation information of the product, and Ako is the average value of the degree of polarity that represents the positive or negative degree of the attribute expression Ak in the entire reputation information of the product p. For each attribute expression Ak of the object p Calculating Aku * Akp * Ako, and calculating Σ Aku * Akp * Ako, which is the sum of those calculated for each attribute expression Ak of the object p, as the fitness of the advertisement of the object p. Features.

  In an eighth advertisement presentation system, in the advertisement presentation system, the fitness level calculation unit weights the fitness level of an advertisement including a keyword of a page on which an advertisement is presented, and the advertisement selection unit includes the weighted advertisement. The advertisement to be presented to the user is selected on the basis of the degree of fitness.

  According to a ninth advertisement presenting system, in the above advertisement presenting system, the fitness level calculating means is configured based on reputation information obtained by evaluating the object A among user attribute information and information on the object B. The object similarity between A and the object B is calculated, and the new object is calculated based on the object similarity between the object A and the object B and the fitness of the advertisement related to the object B. It is characterized by calculating the fitness of the advertisement related to the object B.

  In a tenth advertisement presentation system according to the advertisement presentation system, the matching degree calculation means calculates a user similarity between users based on the reputation information, and based on the user similarity, It is characterized by calculating the fitness of advertisement using reputation information.

  In an eleventh advertisement presentation system, in the advertisement presentation system, the fitness level calculation means weights reputation information related to the written time, and reflects the weighting in the fitness level of the advertisement. Features.

  The first program is a program for an information processing device, and a fitness level for calculating a fitness level for an advertisement related to an object based on reputation information including an expression for evaluating the object and user attribute information. The information processing apparatus is caused to execute a calculation process and an advertisement selection process for selecting an advertisement to be presented to the user based on the fitness.

  In the above program, the second program calculates the fitness level calculation processing so that the fitness level of the advertisement related to the object of reputation information including a positive evaluation expression with respect to the attribute information of the user is increased. It is characterized by that.

  In the third program, the relevance calculation process weights the relevance of the advertisement including the keyword of the page presenting the advertisement, and the ad selection process performs the weighting based on the relevance of the weighted advertisement. An advertisement to be presented to the user is selected.

  Although the present invention has been described above with reference to preferred embodiments and aspects, the present invention is not necessarily limited to the above-described embodiments and aspects, and various modifications may be made within the scope of the technical idea. I can do it.

  This application claims the priority on the basis of Japanese application Japanese Patent Application No. 2007-287091 for which it applied on November 5, 2007, and takes in those the indications of all here.

  The present invention can be used in an advertisement distribution system that uses reputation information and product detail information. In addition to advertisements, it can also be applied to various product recommendation services.

Claims (15)

An advertisement presentation method,
Ad presentation system
Get the user ID of the user browsing the site,
Reputation information including an object, an attribute expression of the object and an evaluation expression for the attribute expression of the object, a degree of polarity that takes a larger value as the degree of affirmation is higher, and a user who has written the reputation information From the evaluation information database associated with the user ID, the number of each attribute expression of each object associated with the user ID, the total number of each attribute expression associated with each object, Search the evaluation information database for the degree of polarity associated with each attribute expression,
The number of each attribute expression of each object associated with the retrieved user ID, the total number of each attribute expression associated with each retrieved object, and the retrieved attribute expressions And the degree of suitability for the user browsing the site of the advertisement for each object based on the polarity associated with
A method for presenting an advertisement, wherein a predetermined number of advertisements are selected in order from the advertisement of an object with a high degree of fitness as an advertisement to be presented to a user browsing the site . The advertisement presentation method according to claim 1 , wherein the advertisement presentation system collects the reputation information from a bulletin board or a blog on a network. The advertisement presentation system includes:
Wherein the number of attributes representing Ak object p that is associated with the user ID and Aku, the number of attributes representing Ak associated with the object p and Akp,
If the Ako the average value of the polarity of the attribute expression Ak associated with the object p, with respect to each attribute representation Ak of the object p,
Aku * Akp * Ako
Calculate
The object above p is the sum of those calculated for each attribute representing Ak Σ Aku * Akp * Ako
The advertisement presentation method of claim 1 or claim 2, characterized in that calculated as the ad relevance of the object p for user viewing the site. The advertisement presentation system includes:
Performs weighting ads to fit including keyword pages presenting advertisements, based on the weighted advertisements fit, claim from claim 1, characterized in that selecting an advertisement to be presented to the user 4. The advertisement presenting method according to any one of 3 . The advertisement presentation system includes:
Based on the evaluation information including an evaluation expression that evaluates the reputation information including an evaluation expression for evaluating an attribute representation of the object A and the object A, the attribute representation of the object B and the object B, the object Calculating the object similarity between A and the object B;
The advertisement adaptability for the new object B is calculated based on the object similarity between the object A and the object B and the advertisement adaptability for the object B. The advertisement presentation method according to any one of claims 1 to 4 . The advertisement presentation system includes:
Calculate user similarity between users based on the reputation information,
The advertisement presentation method according to any one of claims 1 to 5 , characterized in that, based on the user similarity, advertisement suitability is calculated using reputation information of similar users. The advertisement presentation system includes:
The reputation information is weighted in relation to the written time,
The advertisement presentation method according to any one of claims 1 to 6 , wherein the weighting is reflected in the fitness of the advertisement. An ad presentation system,
Reputation information including an object, an attribute expression of the object and an evaluation expression for the attribute expression of the object, a degree of polarity that takes a larger value as the degree of affirmation is higher, and a user who has written the reputation information An evaluation information database associated with the user ID of
Means for obtaining a user ID of a user browsing the site;
The number of each attribute expression of each object associated with the user ID, the total number of each attribute expression associated with each object, and the degree of polarity associated with each attribute expression Search from the evaluation information database, the number of each attribute expression of each object associated with the searched user ID, the total number of each attribute expression associated with each searched object A fitness calculation means for calculating a fitness for a user browsing the site of the advertisement related to each object based on the polarities associated with the searched attribute expressions ;
Advertisement presentation system characterized by having an advertisement selection means for selecting a predetermined number of advertisements from the advertisement of the goodness of fit is large object, as an advertisement to be presented to the user viewing the site. The advertisement presentation system according to claim 8 , wherein the fitness level calculating unit collects reputation information from a bulletin board or a blog on a network. The fitness calculation means includes:
Wherein the number of attributes representing Ak object p that is associated with the user ID and Aku, the number of attributes representing Ak associated with the object p and Akp,
If the Ako the average value of the polarity of the attribute expression Ak associated with the object p, with respect to each attribute representation Ak of the object p,
Aku * Akp * Ako
Calculate
The object above p is the sum of those calculated for each attribute representing Ak Σ Aku * Akp * Ako
The advertisement presentation system according to claim 8 or 9 , characterized in that is calculated as a degree of suitability of an advertisement of the object p with respect to a user browsing the site . The fitness calculation means weights the fitness of the advertisement including the keyword of the page presenting the advertisement,
The advertisement presentation system according to any one of claims 8 to 10 , wherein the advertisement selection unit selects an advertisement to be presented to the user based on the fitness of the weighted advertisement. The fitness calculation means includes:
Based on the reputation information including reputation information including a representation of evaluating the attribute representation of the object A and the object A, the expression for evaluating an attribute representation of the object B and the object B, a the object A Calculating the object similarity with the object B;
Claims wherein the object A and the object similarity between the object B on the basis of the fitness of the advertisement for the object B, and characterized by calculating the fit of the ad to the new the object B The advertisement presentation system according to any one of claims 8 to 11 . The fitness calculation means includes:
Calculate user similarity between users based on the reputation information,
The advertisement presentation system according to any one of claims 8 to 12 , wherein the advertisement suitability is calculated using reputation information of similar users based on the user similarity. The fitness calculation means includes:
The relative reputation information, performs weighting associated with the written time, advertisement presentation system according to claim 13 claim 8, characterized in that to reflect the weighting advertisement fitness. An information processing apparatus program,
A user ID acquisition process for acquiring a user ID of a user browsing the site;
Reputation information including an object, an attribute expression of the object and an evaluation expression for the attribute expression of the object, a degree of polarity that takes a larger value as the degree of affirmation is higher, and a user who has written the reputation information From the evaluation information database associated with the user ID, the number of each attribute expression of each object associated with the user ID, the total number of each attribute expression associated with each object, A search process for searching the evaluation information database for the degree of polarity associated with each attribute expression;
The number of each attribute expression of each object associated with the retrieved user ID, the total number of each attribute expression associated with each retrieved object, and the retrieved attribute expressions A fitness calculation process for calculating a fitness for a user browsing the site of the advertisement related to each object based on the polarity associated with
An information processing apparatus is configured to execute an advertisement selection process for selecting a predetermined number of advertisements as an advertisement to be presented to a user browsing the site in order from an advertisement of an object having a high fitness. program.

Images