JP6548684B2 - Advertisement distribution apparatus, advertisement distribution method, and advertisement distribution program - Google Patents

Description

  The present invention relates to an advertisement delivery apparatus, an advertisement delivery method, and an advertisement delivery program.

  In recent years, with the rapid development of the Internet, advertisement distribution via the Internet is actively performed. As advertisement delivery, advertisement information such as an image and a catch copy is displayed at a predetermined position of a web page, such as an image for displaying a company or a product etc., and when such advertisement information is clicked, a transition to a web page of an advertiser is there.

  Advertising expenses paid by an advertiser are often determined by the number of clicks on an advertisement posted on a web page, that is, the number of customers sent to a product page of the advertiser. Therefore, it is important that ad distributors click on more ads posted on web pages.

Therefore, in recent years, a technique for devising an advertisement to be effectively recognized by the user has been proposed. For example, Patent Document 1 discloses a terminal device capable of selecting an advertisement matching the user's taste preference and providing the advertisement to the user.
In addition, in order to display an advertisement that attracts the user's interest, it is also effective to provide an advertisement that is related in content to the article the user is viewing. In this case, for example, the advertisement database is searched using a keyword included in the article text, and an advertisement to be given is selected from the search.

JP, 2015-22561, A

  By the way, when the user tries to select an advertisement that is related to the article being browsed, there is a high possibility that the extracted keyword changes depending on the length and expression of the browsed article, and the similarity between the article and the advertisement is quantified. Difficult to assess. For this reason, when extracting an advertisement according to a keyword, there is a possibility that an advertisement which is inherently poorly related to the article may be included as noise.

In addition, since the CPC varies depending on the advertisement, there is a demand for providing the user with as high priority as possible the advertisement with a high CPC, from the business viewpoint of the advertisement distributor.
However, when an advertisement is selected based on the CPC height, the similarity decreases and the CTR decreases for the following reasons. That is, since the number of advertisements with high cost-per-click is relatively small and the variation of the advertisement type tends to be poor, it is highly likely that articles showing high similarity will be biased (for example, insurance companies advertise In the case of the Lord, "fire insurance" etc. for the article "fire" On the other hand, low cost-per-click ads tend to have relatively large numbers and variations in ad types, so similar ads for all types of articles (eg fire) (eg fire extinguishers, It is highly likely that it contains a lot of books on fire themes. Therefore, if priority is given only to similarity, ads with lower CPC will be selected, and if priority is given to only CPC height, the similarity of the ad will be lowered and the CTR will be lowered. X It is difficult to maximize the revenue from clicks.

  The present invention has been made in view of such circumstances, and it is an object of the present invention to provide an advertisement distribution apparatus, an advertisement distribution method, and an advertisement distribution program capable of performing advertisement distribution in consideration of advertisement revenue. .

In order to solve the above-mentioned subject, the present invention adopts the following means.
The present invention uses display probability calculation means for calculating display probabilities of a plurality of advertisements using a parameter related to the number of times an advertisement has been clicked and a parameter related to the price of the advertisement, and a plurality of browsed articles of the user Similarity calculation means for calculating the similarity to the first threshold, first threshold calculation means for calculating the first threshold by the statistical method from the similarity of the plurality of advertisements acquired by the similarity calculation means, and the similarity calculation In the case of receiving a second threshold calculating unit that calculates a second threshold that is a relative threshold to the maximum similarity using the maximum similarity among the similarity of the advertisement acquired by the unit, and a distribution instruction of the advertisement , based on the display probability computed by the display probability calculation unit from among the similarity is equal to or greater than the first threshold value or more and the second threshold value ad delivery pairs It said advertisement selection means for selecting an advertisement comprising, providing a Ru advertisement distribution device and an advertisement distribution unit configured to distribute advertisement content of the selected the advertisement as a distribution target.

The present invention relates to a display probability calculation process in which a computer calculates display probabilities of a plurality of advertisements using a parameter related to the number of times an advertisement is clicked and a parameter related to the price of the advertisement, and a plurality of users A similarity calculation step of calculating the similarity to the browsed article, a first threshold calculation step of calculating the first threshold by the statistical method from the similarity of the plurality of advertisements acquired by the similarity calculation step, and A second threshold calculation step of calculating a second threshold which is a relative threshold to the maximum similarity using the maximum similarity among the similarities of the advertisements acquired in the similarity calculation step; when receiving, the display probability computed in the display probability calculation step from among the similarity is equal to or greater than the first threshold value or more and the second threshold value ad Zui and provides the advertisement selection step of selecting the advertisement to be delivered subject, the advertisement distribution method of performing an advertisement delivery step for delivering advertising content to selected the advertisement as a distribution target.

The present invention uses a parameter regarding the number of times an advertisement is clicked and a parameter regarding the price of the advertisement to calculate display probability of the plurality of advertisements, and a viewing article of the user for the plurality of advertisements. Similarity calculation processing for calculating the similarity to the first threshold, first threshold calculation processing for calculating the first threshold by the statistical method from the similarity of the plurality of advertisements acquired by the similarity calculation processing, and the similarity calculation In the case of receiving a second threshold calculation process of calculating a second threshold which is a relative threshold to the maximum similarity using the maximum similarity among the similarities of the advertisement acquired by the process, and a distribution instruction of the advertisement , based on the calculated the display probability in the display probability calculation processing from among the similarity is equal to or greater than the first threshold value or more and the second threshold value ad delivery pairs An advertisement selection process the selecting an advertisement to be, providing advertising distribution program for executing the advertisement distribution process on a computer that delivers advertising content selected the advertisement as a distribution target.

  According to the present invention, it is possible to perform advertisement delivery in consideration of advertisement revenue.

BRIEF DESCRIPTION OF THE DRAWINGS It is the figure which showed the whole structure of the advertisement delivery system which concerns on one Embodiment of this invention. It is the figure shown about the hardware constitutions of the media server concerning the embodiment of the present invention, an advertiser server, an advertising distribution server, and a user terminal. It is a functional block diagram showing an example of a function with which an advertisement distribution server concerning one embodiment of the present invention is provided. It is the figure which showed an example of the registration information stored in the memory | storage part shown in FIG. It is the figure which showed the structure of the similar advertisement extraction part shown in FIG. It is a figure for demonstrating the effect of extracting an advertisement using both a 1st threshold value and a 2nd threshold value. It is a figure for demonstrating the operation | movement of the advertisement delivery system which concerns on one Embodiment of this invention.

Hereinafter, an advertisement delivery apparatus, an advertisement delivery method, and an advertisement delivery program according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing an overall configuration of an advertisement distribution system having an advertisement distribution apparatus according to the present embodiment. As shown in FIG. 1, the advertisement distribution system 1 mainly includes a media server 10, an advertiser (advertising agency) server 11, and an advertisement distribution server (advertisement distribution apparatus) 12. Although one media server 10, one advertiser server 11, and one advertisement distribution server 12 are illustrated in FIG. 1, the number of these is not limited to the illustrated mode, and may be added according to the situation as appropriate. It is possible. The media server 10, the advertiser (advertising agency) server 11, and the advertisement distribution server (advertisement distribution apparatus) 12 are communicably connected to one another via the network 15, and are connected to the network 15 as well. A service (for example, delivery of an article or an advertisement, etc.) according to the browsing request from the user terminal 13 is performed on the server 13. An internet connection is given as an example of the network 15, but it may be another network connection.

  The media server 10 is a server possessed by a mass media such as newspaper, magazine, radio broadcast, television broadcast, or a portal site, and various articles are uploaded daily. When the media server 10 receives an article browsing request from the user terminal 13 via the network 15, the media server 10 distributes the article to be the target of the browsing request to the user terminal 13 via the network 15.

  The advertiser server 11 is a server that provides goods and services to the user terminal 13. For example, a server owned by an advertiser such as a corporation, such as a corporation providing a product or service to a consumer, or an advertisement for advertising a product or service of the corporation or the like in response to a request from the corporation or the like The server which an agency has is mentioned as an example. The advertiser server 11 uploads and stores various advertisements related to products and services of the corporation and the like.

  The user terminal 13 is a so-called information processing apparatus such as a personal computer and a tablet terminal, and a communication unit for performing communication such as the Internet, an input unit for the user to input an input operation, and a display unit for displaying information. Have. For example, the user terminal 13 accesses the media server 10 via the network 15 and browses articles (content) uploaded to the media server 10 using a web browser that the user terminal 13 has. Also, the web browser of the user terminal 13 interprets an article acquired from the media server 10, for example, article content described in HTML (HyperText Markup Language) etc. Send the information of the acquired article and request an advertisement.

  The advertisement distribution server 12 is a configuration that forms the core of the advertisement distribution system in the present embodiment, and when a request for article information and an advertisement is received from the user terminal 13, selects a suitable advertisement similar to the content of the article. And transmit to the user terminal 13. As a result, for example, on the display unit of the user terminal 13, an article suitable for the article is displayed together with the article requested to be browsed by the user.

Next, hardware configurations of the media server 10, the advertiser server 11, and the advertisement distribution server 12 according to the present embodiment will be described with reference to FIG. As shown in FIG. 2, the media server 10, the advertiser server 11, and the advertisement distribution server 12 are so-called computer systems, and for example, a CPU 21 and a ROM (Read Only Memory) for storing programs executed by the CPU 21 and the like. 22, a RAM (Random Access Memory) 23 functioning as a work area at the time of execution of each program, a hard disk drive (HDD) 24 as a mass storage device, a communication interface 25 for connecting to a network, a keyboard, a mouse, etc. And a display unit 27 including a liquid crystal display device or the like for displaying data. These units are connected via a bus 28.
The storage medium for storing the program and the like executed by the CPU 21 is not limited to the ROM 22, and may be another storage medium including a magnetic disk, a magneto-optical disk, a semiconductor memory or the like.
The ROM 22 stores various programs, and the CPU 21 reads the programs from the ROM 22 to the RAM 23 and executes the programs to realize various functions.
The media server 10, the advertiser server 11, and the advertisement distribution server 12 do not have to have the display unit 27. The media server 10, the advertiser server 11, and the advertisement distribution server 12 may be connected to the input unit 26 via a network or the like, and may be configured to allow remote operation.

  FIG. 3 is a functional block diagram showing an example of the function of the advertisement delivery server 12. As shown in FIG. 3, the advertisement distribution server 12 includes, for example, a storage unit 31, a display probability calculation unit 32, a similar advertisement extraction unit 33, an advertisement selection unit 34, an advertisement distribution unit 35, a registration information update unit 36, and an advertisement database. It has 37. For example, various processes to be described later implemented by these units are implemented by the CPU 21 reading out an advertisement distribution program stored in the ROM 22 to the RAM 23 and executing the program.

  The storage unit 31 includes a parameter relating to the number of clicks and a parameter relating to the price of the advertisement for each of the plurality of advertisements that can be distributed by the advertisement distribution server 12 or clusters classified according to predetermined information such as a category. The registration information associated with is stored.

  The parameter relating to the number of clicks is an index value reflecting the number of clicks, and for example, the click rate CTR (Click Through Rate) is taken as an example. The click rate CTR is a value indicating the number of display times at which the advertisement is displayed on the user terminal 13 or the ratio of the number of clicks to the number of distribution times at which the advertisement is distributed to the user terminal 13. Further, the parameter related to the number of clicks may be represented as a function including the click rate CTR or the like as a variable.

  Further, the parameter relating to the price of the advertisement is an index value reflecting the price of the advertisement, and, for example, the cost per click CPC (Cost Per Click) of the advertisement and the like can be mentioned as an example. The parameters relating to the price of the advertisement may be expressed by the cost-per-click itself, or two or more types of parameters may be used depending on the situation, such as the display price and the cost-per-click. In addition, it may be expressed as a function including one or more types of parameters (for example, the cost per click CPC) as a variable.

  The registration information stored in the storage unit 31 is information that is updated as needed. For example, as for the parameter related to the click rate, the value is obtained by the registration information updating unit 36 described later by the advertisement being distributed to the user terminal 13, the advertisement being displayed on the display unit of the user terminal 13, or the advertisement being clicked. It will be updated. Also, the value of the CPC is updated, for example, when the CPC is changed with the advertiser.

  FIG. 4 shows an example of registration information in the case where the click rate CTR is exemplified as a parameter relating to the click function, and the cost per click CPC is exemplified as a parameter relating to the price of the advertisement. FIG. 4A shows registration information in the initial state, and FIG. 4B shows registration information when display is performed a predetermined number of times. The display probability and the definite display probability described in FIG. 4A and FIG. 4B are information to be calculated and stored by a display probability calculation unit 32 described later. The display probability and the fixed display probability may not be associated with the registration information, but may be associated with only the advertisement ID and registered in the storage unit 31. In the initial state shown in FIG. 4A, an example is shown in which the same value (100 [%] / 3 = 0.33 [%]) is registered as the click rate CTR, but the initial value is appropriately set. It is possible.

  The display probability calculation unit 32 calculates the display probability of each advertisement by using a parameter related to the number of clicks of each advertisement stored in the storage unit 31 and a parameter related to the price of the advertisement in a predetermined arithmetic expression. This computing equation is, for example, a computing equation such that the display probability is higher as the value of the parameter related to the number of clicks is larger or as the parameter related to the price of the advertisement is larger.

  For example, the display probability calculation unit 32 calculates the display probability using the following arithmetic expression (1).

Display probability = eCPM / eCPM _total (1)

Here, eCPM = CTR × CPC × 1000. Also, eCPM _total is the sum of eCPM of all advertisements.

  For example, when the registration information as shown to Fig.4 (a) is stored in the memory | storage part 31, the display probability of each advertisement becomes as follows.

ID "0001": eCPM = 0.33 x 10 x 1000 = 3300
ID "0002": eCPM = 0.33 x 50 x 1000 = 16500
ID "0003": eCPM = 0.33 x 100 x 1000 = 33000

ID “0001”: display probability = 3300/52800 × 100 ≒ 6.3
ID “0002”: display probability = 16500/52800 × 100 ≒ 31.2
ID “0003”: display probability = 33000/52800 × 10010062.5

The display probability of each advertisement calculated by the display probability calculation unit 32 is, for example, associated with identification information of each advertisement and stored in the storage unit 31 (see FIG. 3).
In the above example, the display probability is calculated by simply multiplying the parameter related to the number of clicks and the parameter related to the price of the advertisement. However, the calculation formula is not limited to the above example, for example, a calculation formula using weighting values It is also good.
In addition, when there is a past value of display probability for each advertisement, the display probability calculation unit 32 may calculate the latest display probability using the past value. As described above, it is possible to suppress a sudden change in display probability by calculating the latest display probability using a past value.
For example, when using the previous value of the display probability, the display probability is calculated as an average value of the display probability (hereinafter referred to as “display probability (current value)”) calculated using the above equation (1) and the previous value. Be done. It is as follows when it represents as a formula. Moreover, in order to distinguish the display probability calculated by the following equation (2) from the display probability (the current value), it is referred to as “certain display probability” for convenience.

    Fixed display probability = {display probability (previous value) + display probability (current value)} / 2 (2)

For example, when each display probability (current value) is calculated based on the registered state of FIG. 3B, the following results are obtained.
ID "0001": Display probability (current value) 2 26.0
ID "0002": Display probability (current value) 3 37.0
ID "0003": Display probability (current value) 3 37.0

Therefore, the finalized display probability of each advertisement is as follows.
ID “0001”: definite display probability = (6.3 + 26.0) /2=16.2
ID “0002”: definite display probability = (31.2 + 37.0) /2=34.1
ID “0003”: definite display probability (62.5 + 37.0) /2=49.7

  In addition, the method of calculating the latest display probability using the past display probability is not limited to the above example, for example, various calculation methods such as calculating statistically using the display probability for several times in the past, etc. It is possible to adopt.

When the media server 10 receives a browsing request from the user terminal 13, the similar advertisement extracting unit 33 receives information on an article for which the browsing request has been received from the media server 10, and extracts an advertisement similar to the content of the article. .
For example, as shown in FIG. 5, the similar advertisement extracting unit 33 includes a similarity calculating unit 41, a first threshold calculating unit 42, a second threshold calculating unit 43, and an extracting unit 44.
The similarity calculation unit 41, for example, extracts a keyword from the content of the article received from the media server 10, and calculates the similarity of each advertisement based on the keyword. Note that the calculation of the advertisement similarity may be performed for each cluster created by classifying the advertisement according to the category and appearance tendency. Here, the degree of interest of the user may be used instead of the degree of similarity.
A number of known methods have been proposed as a technique for extracting keywords from a read article and a method for calculating the degree of similarity (degree of interest) between a read article and an advertisement, and these methods may be adopted as appropriate. For example, it is realizable by employ | adopting suitably the method disclosed by Unexamined-Japanese-Patent No. 2015-22561, 2016-62102.

The first threshold calculator 42 calculates a first threshold from the similarity of the advertisement calculated by the similarity calculator 41 by a statistical method. For example, the first threshold is calculated by calculating the average, the root of the square root, or the like of all the advertisements or the advertisements having a similarity that exceeds the predetermined threshold or the similarity per cluster.
The second threshold calculating unit 43 specifies the maximum similarity among the similarity of the advertisement calculated by the similarity calculating unit 41, and calculates the second threshold using the maximum similarity. For example, the second threshold calculation unit calculates a value obtained by multiplying the maximum similarity by a predetermined constant (for example, 0.4) as a second threshold. Thus, the second threshold is a threshold relative to the maximum similarity.
The extraction unit 44 extracts an advertisement having a similarity of at least a first threshold and at least a second threshold from all advertisements or all clusters, and creates an advertisement list in which the extracted advertisements (clusters) are described.

  Here, the effect of extracting an advertisement using both the first threshold and the second threshold will be described with reference to FIG. FIG. 6 is a diagram showing the advertisement possessed by each of the advertisers A, B and C for each cluster classified according to the category, and the vertical axis shows the degree of similarity. Here, the ads handled by the advertisers A and B have relatively high CPCs and are concentrated in categories, while the ads handled by the advertiser C have relatively low CPCs and the categories have a variety Because of their richness, the dynamic range of the degree of similarity is shown quite widely compared with the advertisers A and B.

  As shown in FIG. 6, when the dynamic range (variation) of the degree of similarity of advertisements is large, a large number of advertisements not related to articles are included. Therefore, due to the influence of these ads having low similarity, the value of the first threshold calculated using a statistical method (for example, the mean or the square root) may be calculated low. For this reason, when the advertisement is extracted using only the first threshold, the extracted advertisement includes an advertisement having a relatively low degree of similarity, such as the advertisement of the advertiser A in FIG. There is a fear. Then, when the cost per click of this relatively low advertisement (the advertisement of the advertiser A in FIG. 6) is high, the display probability becomes a high value, and it may be selected as the advertisement to be provided to the user.

  However, since such an advertisement with a relatively low degree of similarity (degree of interest) is not originally an advertisement that attracts the user's interest, it is desirable to exclude it from distribution candidates as much as possible. Therefore, in the present embodiment, the second threshold, which is a relative threshold calculated using the maximum similarity, is also used to extract an advertisement. Thus, by using the second threshold, for example, as shown in FIG. 6, it is avoided that the advertisement of the advertiser A whose similarity is relatively low among the advertisements having the first threshold or more is extracted. be able to.

The advertisement selecting unit 34 selects a predetermined number of advertisements from the advertisement list extracted by the similar advertisement extracting unit 33 in the order of advertisements having a high (provisional) display probability calculated by the display probability calculating unit 32. The predetermined number changes, for example, according to the configuration of the browsed article received from the media server 10.
The advertisement distribution unit 35 extracts advertisement content corresponding to the predetermined number of advertisements selected by the advertisement selection unit 34 from the advertisement database 37, and transmits the extracted advertisement content to the user terminal 13. Here, in the advertisement database 37, advertisement content transmitted by the advertisement distribution server to the user terminal 13 is stored in association with the advertisement ID. In addition, the advertisement database 37 may be provided in each advertiser server 11 which is an advertiser. That is, the storage format is not particularly limited as long as the advertisement distribution server 12 can obtain the advertisement content.

  For example, when the advertisement is distributed to the user terminal 13, the advertisement is displayed on the display unit 27 of the user terminal 13, or the advertisement is clicked by the user, the registration information updating unit 36 uses the information from the user terminal 13 The registration information stored in the storage unit 31 is updated based on the acquired information. Moreover, when the contract etc. of the CPC is changed, the registration information is updated based on the information after the change. The registration information update unit 36 also counts the total number of clicks.

  Next, the operation of the advertisement distribution system 1 according to an embodiment of the present invention will be described with reference to FIG. In addition, although the advertisement delivery system 1 according to the present embodiment delivers articles and advertisements to many user terminals 13 via the network 15, one user terminal 13 is illustrated for convenience of explanation. The transmission and reception of information with the user terminal 13 will be described.

  First, when a predetermined article or the like is selected by the user on the web browser displayed on the user terminal 13, a browsing request corresponding to the article is transmitted to the media server (step SA1 in FIG. 7). The media server 10 extracts the article for which the browsing request has been made, and transmits the extracted article to the user terminal 13 (step SA2 in FIG. 7). When receiving an article from the media server 10, the user terminal 13 transmits the received article information to the advertisement distribution server 12 and requests an advertisement (step SA3 of FIG. 7).

  When the advertisement delivery server 12 receives an advertisement delivery request together with article information, first, the similar advertisement extraction unit 33 extracts an advertisement list similar to the content of the received article (step SA4 in FIG. 7). Specifically, first, the similarity calculation unit 41 (see FIG. 5) extracts a keyword from the content of the article, and calculates the similarity of each advertisement based on the keyword. Next, using the similarity calculated by the similarity calculation unit 41, the first threshold calculation unit 42 calculates a first threshold that is a statistical threshold. Further, the second threshold value operation unit 43 specifies the maximum similarity among the similarities calculated by the similarity degree operation unit 41, and calculates the second threshold using the maximum similarity. When the first threshold and the second threshold are calculated, the extraction unit 44 extracts, as an advertisement list, advertisements of which the degree of similarity is equal to or higher than the first threshold and equal to or higher than the second threshold among all the advertisements.

Subsequently, the advertisement selection unit 34 (see FIG. 3) refers to the latest determined display probability stored in the storage unit 31 and selects a predetermined number of advertisements listed in the advertisement list in descending order of display probability. An advertisement is selected (step SA5 in FIG. 7). When an advertisement is selected by the advertisement selection unit 34, the advertisement distribution unit 35 acquires the content of the selected advertisement from the advertisement database 37 (step SA6 in FIG. 7), and transmits the extracted advertisement content to the user terminal 13 (FIG. 7) Step SA7 of FIG. 7).
Thus, the display unit 27 of the user terminal 13 displays the article information transmitted from the media server 10, and displays the advertisement content transmitted from the advertisement distribution server 12 at a predetermined position of the article. (Step SA8 in FIG. 7).

  In the display screen, when the user clicks (selects) the advertisement content, a request for viewing advertisement information associated with the advertisement content is transmitted to the advertisement distribution server 12 (step SA9 in FIG. 7). In the advertisement distribution server 12, the number of clicks is measured by the registration information update unit 36, and the registration information stored in the storage unit 31 is updated based on the reception information (step SA10 in FIG. 7). That is, the parameters related to the click corresponding to the clicked advertisement content are updated. In addition, a browsing request for advertisement information from the user terminal 13 is transmitted to the advertiser server 11 (step SA11 in FIG. 7), and advertisement data corresponding to the browsing request is transmitted from the advertiser server 11 to the user terminal 13 (FIG. 7 steps SA12). Thereby, the display unit 27 of the user terminal 13 displays advertisement data distributed from the advertiser server 11, for example, a web page on which product information is posted.

  In addition, the display probability calculation unit 32 (see FIG. 3) calculates the finalized display probability based on the registration information stored in the storage unit 31 at a predetermined timing, and stores the latest (finalized) display probability in the storage unit 31. Store. The predetermined timing is, for example, each time the total number of clicks measured by the registration information updating unit 36 reaches a predetermined number of times (for example, 1000 times), or a predetermined time has elapsed from the time of the last update of the definite display probability Occasionally, etc. may be mentioned as an example. Also, each time the registration information is updated, the fixed display probability of the updated advertisement may be calculated.

  As described above, according to the advertisement distribution server (the advertisement distribution device), the advertisement distribution method, and the advertisement distribution program according to the present embodiment, the advertisement using the parameter related to the price of the advertisement and the parameter related to the number of clicks of the advertisement The display probability of is calculated, and the advertisement to be provided to the user is selected using the calculated display probability. Here, since the display probability is calculated using an expression such that the higher the ad price, the higher the number of clicks on the ad, the higher the display probability, so both the number of clicks on the ad and the price are reflected. It is possible to select suitable advertisement content and provide it to the user.

  Furthermore, the advertisement delivered to the user terminal 13 is limited to advertisements whose similarity to the content of the article for which the user has requested browsing is higher than the first threshold and higher than the second threshold. Transmission to the user terminal 13 can be avoided (see FIG. 6). In other words, only advertisements with relatively high similarity can be extracted as selection candidates, and among them, advertisements with high display probability can be provided to the user. As a result, it is possible to realize advertisement delivery in consideration of advertisement revenue while eliminating the delivery of advertisements having a low degree of similarity.

As mentioned above, although this invention was demonstrated using the said embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. Various changes or improvements can be added to the above-described embodiment without departing from the scope of the invention, and a form to which the changes or improvements are added is also included in the technical scope of the present invention. Also, the above embodiments may be combined as appropriate.
In the above embodiment, the procedure of processing in the advertisement distribution server has been described, but these procedures are an example, and unnecessary processing steps may be deleted or new processing may be performed without departing from the scope of the present invention. Steps may be added or processing steps may be switched.

1: Advertisement distribution system 10: Media server 11: Advertiser server 12: Advertisement distribution server 13: User terminal 27: Display unit 31: Storage unit 32: Display probability calculation unit 33: Similar advertisement extraction unit 34: Advertisement selection unit 35: Advertisement distribution unit 36: Registration information update unit 37: Advertisement database 41: Similarity calculation unit 42: First threshold calculation unit 43: Second threshold calculation unit 44: Extraction unit

Claims (5)

Display probability calculation means for calculating display probabilities of a plurality of advertisements using a parameter related to the number of times an advertisement is clicked and a parameter related to the price of the advertisement;
Similarity calculation means for calculating the similarity to the browsing article of the user for a plurality of the advertisements;
First threshold value calculation means for calculating a first threshold value by a statistical method from the similarity degrees of the plurality of advertisements acquired by the similarity degree calculation means;
Second threshold value calculation means for calculating a second threshold value which is a relative threshold value to the maximum similarity using the maximum similarity among the similarity of the advertisement acquired by the similarity calculation means;
When a distribution instruction of an advertisement is received, it becomes a distribution target based on the display probability calculated by the display probability calculation unit from among the advertisements whose similarity is equal to or higher than the first threshold and equal to or higher than the second threshold. Advertisement selection means for selecting the advertisement;
Ru advertisement distribution device and an advertisement distribution unit configured to distribute advertisement content of the selected the advertisement as a distribution target.   The advertisement distribution device according to claim 1, wherein the display probability calculation means calculates the display probability using an arithmetic expression in which the display probability is calculated to be higher as the price of the advertisement is higher.   The advertisement distribution apparatus according to claim 1, wherein the display probability calculation unit calculates the latest display probability using the display probability in the past of each of the advertisements. The computer is
A display probability calculation step of calculating display probabilities of a plurality of advertisements using a parameter related to the number of times an advertisement is clicked and a parameter related to the price of the advertisement;
A similarity calculation step of calculating similarity to a user's browsed article for a plurality of the advertisements;
A first threshold value calculating step of calculating a first threshold value by a statistical method from the similarity of the plurality of advertisements acquired in the similarity calculating step;
A second threshold calculation step of calculating a second threshold which is a relative threshold to the maximum similarity using the maximum similarity among the similarity of the advertisement acquired in the similarity calculation step;
When a distribution instruction of an advertisement is received, it becomes a distribution target based on the display probability calculated in the display probability calculation step from among the advertisements whose similarity is the first threshold or more and the second threshold or more. An advertisement selection step of selecting the advertisement ;
Advertisement distribution method of performing an advertisement delivery step for delivering advertising content to selected the advertisement as a distribution target. Display probability calculation processing for calculating display probabilities of a plurality of advertisements using a parameter related to the number of times an advertisement has been clicked and a parameter related to the price of the advertisement;
Similarity calculation processing for calculating the similarity to the user's browsed articles for a plurality of the advertisements;
A first threshold value calculation process of calculating a first threshold value by a statistical method from the similarity of the plurality of advertisements acquired by the similarity calculation process;
A second threshold value calculation process of calculating a second threshold value that is a relative threshold value to the maximum similarity using the maximum similarity among the similarity of the advertisement acquired by the similarity calculation process;
When a distribution instruction of an advertisement is received , it becomes a distribution target based on the display probability calculated in the display probability calculation process from among the advertisements whose similarity is the first threshold or more and the second threshold or more. An advertisement selection process of selecting the advertisement ;
An advertisement distribution program for causing a computer to execute an advertisement distribution process of distributing advertisement content of the advertisement selected as a distribution target.

Images