JP7275216B2 - ADVERTISING PRODUCTION SUPPORT PROGRAM, ADVERTISING PRODUCTION SUPPORT DEVICE, AND ADVERTISING PRODUCTION SUPPORT METHOD - Google Patents

Description

The present invention relates to an advertisement production support program , an advertisement production support device, and an advertisement production support method.

BACKGROUND ART The importance of advertisements in business, public activities, etc. has been recognized for some time, and techniques for predicting the effects of advertisements have been developed. As an example of such technology, there is a generation device disclosed in Patent Document 1. This generation device includes an acquisition unit and a generation unit. The acquisition unit acquires user information about a user. The generation unit generates a model for predicting an evaluation value related to the type in the data format of the advertisement information based on the user information acquired by the acquisition unit. Further, the type of the data format of the advertisement information referred to here indicates a classification based on the data format of the advertisement information, such as character information, image information, or a combination of character information and image information.

JP 2017-054261 A

However, the generation device described above only predicts an evaluation value related to the type of advertising content to be distributed. For this reason, the generation device described above determines whether the effect of the undelivered advertisement, which is a candidate for the advertisement to be newly delivered, exceeds the effect of the delivered advertisement that has already been delivered, and whether there is merit in delivering the undelivered advertisement. It is not possible to present the grounds for judging Further, even when the data format of the undelivered advertisement and the data format of the delivered advertisement are different, the above-described generating device determines that the effect of the undelivered advertisement exceeds the effect of the delivered advertisement, and the undelivered advertisement is It is not possible to present grounds for judging whether or not there is merit in distribution.

In addition, if the data format differs for each ad due to different ad formats, a dedicated machine learning model is used for each data format to estimate the delivery performance of each ad and calibrate the estimated delivery performance of each ad. It can also be envisaged that However, the calibration is difficult.

The present invention has been made in view of the circumstances described above, and determines whether or not there is merit in distributing an undelivered advertisement even if the advertisement format of the undelivered advertisement and the advertisement format of the delivered advertisement are different. It is intended to provide an advertisement production support program , an advertisement production support device, and an advertisement production support method capable of presenting the grounds for making an advertisement.

One aspect of the present invention provides a computer with first advertising data indicative of a first advertisement having a first advertising format and second advertising data indicative of a second advertisement having a second advertising format different from the first advertising format. and an advertisement data acquisition function that acquires the first advertisement data and the second advertisement data to cause a machine learning device to learn, and based on the advertisement data that indicates the advertisement, the distribution indicating the estimation result of the delivery performance of the advertisement and a machine learning execution function for causing the machine learning device to generate a distribution performance estimation program for outputting performance estimation data.

One aspect of the present invention is the above machine learning execution program, wherein the advertisement data acquisition function has the same data format as the second advertisement data, and the part related to the second advertisement format is masked. Obtaining first advertising data and obtaining the second advertising data having the same data format as the first advertising data and masking a portion related to the first advertising format.

According to one aspect of the present invention, a computer has a condition specification data acquisition function for acquiring condition specification data specifying a condition for estimating a delivery record of an undelivered advertisement; A delivered advertisement data acquisition function for acquiring delivered advertisement data indicating a delivered advertisement having a delivery performance satisfying a predetermined standard; and delivery performance data showing the delivery performance of the delivered advertisement indicated by the delivered advertisement data. Shows a distribution performance data acquisition function to be acquired, an undelivered advertisement data acquisition function for acquiring undelivered advertisement data indicating the undelivered advertisement having a data format different from that of the delivered advertisement, and a first advertisement having a first advertisement format. using a delivery performance estimation program generated using first advertisement data and second advertisement data indicating a second advertisement having a second advertisement format different from the first advertisement format, according to the condition specifying data a delivery performance estimation data generating function for estimating the delivery performance of the undelivered advertisement under the indicated conditions and generating delivery performance estimation data indicating an estimation result of the delivery performance of the undelivered advertisement; and the delivered advertisement data. , a data output function for outputting the delivery performance data, the undelivered advertisement data, and the delivery performance estimation data.

One aspect of the present invention is the advertisement production support program described above, wherein the distribution performance estimation data generation function has the same data format as the second advertisement data, and the portion related to the second advertisement format is masked. and the second advertisement data having the same data format as the first advertisement data and masking the portion related to the first advertisement format, the delivery performance estimation generated using A program is used to generate the estimated delivery performance data.

One aspect of the present invention obtains first advertising data indicative of a first advertisement having a first advertising type and second advertising data indicative of a second advertisement having a second advertising type different from the first advertising type. and an advertisement data acquisition unit, which causes a machine learning device to learn using the first advertisement data and the second advertisement data, and based on the advertisement data indicating the advertisement, delivery performance estimation data indicating an estimation result of the delivery performance of the advertisement. and a machine learning execution unit that causes the machine learning device to generate a distribution performance estimation program that outputs a.

According to one aspect of the present invention, a condition designating data acquisition unit that acquires condition designating data that designates a condition for estimating the delivery performance of an undelivered advertisement; a delivered advertisement data acquisition unit that acquires delivered advertisement data indicating a delivered advertisement that has a delivery performance that satisfies the conditions; A performance data acquisition unit, an undelivered advertisement data acquisition unit for acquiring undelivered advertisement data indicating the undelivered advertisement having a data format different from that of the delivered advertisement, and a first advertisement indicating the first advertisement having the first advertisement format. and second advertisement data indicating a second advertisement having a second advertisement format different from the first advertisement format, using a delivery performance estimation program generated using the said indicated by said condition specifying data a delivery performance estimation data generation unit for estimating the delivery performance of the undelivered advertisement under a condition and generating delivery performance estimation data indicating an estimation result of the delivery performance of the undelivered advertisement; and a data output unit that outputs performance data, the undelivered advertisement data, and the estimated delivery performance data.

According to one aspect of the present invention, an advertisement data acquisition function or an advertisement data acquisition unit provides first advertisement data indicating a first advertisement having a first advertisement format and second advertisement data having a second advertisement format different from the first advertisement format. Second advertisement data indicating an advertisement is acquired, and a machine learning function or a machine learning execution unit uses the first advertisement data and the second advertisement data to learn a machine learning device, and obtains advertisement data indicating an advertisement. The machine learning execution method causes the machine learning device to generate a delivery performance estimation program for outputting delivery performance estimation data indicating an estimation result of the delivery performance of the advertisement based on the data.

According to one aspect of the present invention, a condition-designated data obtaining function or a condition-designated data obtaining unit obtains condition-designated data that designates conditions for estimating the distribution performance of an undelivered advertisement, and the delivered advertisement data obtaining function or distribution A delivered advertisement data acquisition unit acquires delivered advertisement data indicating a delivered advertisement having a delivery record that satisfies the condition indicated by the condition specifying data and satisfies a predetermined standard, and a delivery performance data acquisition function or delivery performance data An acquisition unit acquires delivery performance data indicating delivery performance of the delivered advertisement indicated by the delivered advertisement data, and an undelivered advertisement data acquisition function or an undelivered advertisement data acquisition unit acquires the delivered advertisement and data format. is acquired, and the first advertisement data indicating the first advertisement having the first advertisement format, the first advertisement data indicating the first advertisement having the first advertisement format, and the second using a delivery performance estimation program generated using a second advertisement data indicating a second advertisement having a second advertisement format different from the one advertisement format, under the conditions indicated by the condition specifying data; estimating the delivery performance of the undelivered advertisement, generating delivery performance estimation data indicating the estimation result of the delivery performance of the undelivered advertisement, and outputting the delivered advertisement data, the delivery performance data, It is an advertisement production support method for outputting the undelivered advertisement data and the delivery performance estimation data.

According to the present invention, even if the advertisement format of the undelivered advertisement and the advertisement format of the delivered advertisement are different, it is possible to present the grounds for determining whether or not there is merit in delivering the undelivered advertisement.

It is a figure which shows an example of the function with which the machine-learning execution program which concerns on embodiment is provided. FIG. 10 is a diagram showing an example of teacher data that the machine learning execution program according to the embodiment inputs to the machine learning device in order to cause the machine learning device to generate a distribution performance estimation program; It is a figure for demonstrating an example of the delivery performance estimation program produced|generated by the machine-learning execution program which concerns on embodiment. It is a figure for demonstrating an example of the delivery performance estimation program produced|generated by the machine-learning execution program which concerns on embodiment. 4 is a flowchart showing an example of processing executed by a machine learning execution program according to the embodiment; It is a figure which shows an example of the function with which the advertisement production assistance program which concerns on embodiment is provided. FIG. 10 is a diagram showing an example of a user interface displayed on the display when condition specifying data according to the embodiment is input; FIG. 10 is a diagram showing an example of a user interface displayed on the display when condition specifying data according to the embodiment is input; It is a figure which shows an example of the display mode of the delivered advertisement shown by the delivered advertisement data which concerns on embodiment, and the delivery performance shown by delivery performance data. FIG. 7 is a diagram showing an example of a user interface displayed on the display when undelivered advertisement data is acquired by the advertisement production support program according to the embodiment; FIG. 10 is a diagram showing an example of a display mode of a delivered advertisement, a delivery performance of the delivered advertisement, an undelivered advertisement, and an estimation result of the delivery performance of the undelivered advertisement displayed on the display by the advertisement production support program according to the embodiment; . 6 is a flow chart showing an example of processing executed by an advertisement production support program according to the embodiment;

[Embodiment]
First, a machine learning execution program according to an embodiment will be described with reference to FIGS. 1 to 5. FIG. FIG. 1 is a diagram illustrating an example of functions provided by a machine learning execution program according to an embodiment. As shown in FIG. 1 , the machine learning execution program 10 has an advertisement data acquisition function 11 and a machine learning execution function 12 .

At least some of these functions are realized by reading and executing the advertisement production support program 20 stored in the storage device by a hardware processor such as a CPU (Central Processing Unit). The storage device referred to here is, for example, a hard disk drive (HDD: Hard Disk Drive), a solid state drive (SSD: Solid State Drive), a flash memory (Flash Memory), and a ROM (Read Only Memory).

The advertisement data acquisition function 11 acquires first advertisement data, second advertisement data, and other advertisement data. The first advertisement data is data indicative of a first advertisement having a first advertisement type. The second advertisement data is data indicating a second advertisement having a second advertisement format different from the first advertisement format. The advertisement format referred to here means a specific mode of advertisement. Specific examples of advertisements include still image advertisements, video advertisements, carousel advertisements, and text advertisements.

Also, the first advertisement data has the same data format as the second advertisement data, and the portion related to the second advertisement format is masked. For example, when the first advertisement data is data indicating a still image advertisement and the second advertisement data is data indicating a video advertisement with sound, the first advertisement data is masked in the portion indicating the moving image and the portion indicating the sound. It is the data that is

Similarly, the second advertising data has the same data format as the first advertising data, with the portion relating to the first advertising format being masked. For example, when the first advertisement data is data indicating a still image advertisement and the second advertisement data is data indicating a video advertisement, the second advertisement data is data in which the portion indicating the still image is masked. there is

As a masking method in these cases, it is preferable to replace all masked portions with numerals of zero. In addition, when masking the part showing the text, when masking the part showing the elements of the targeting design other than the creative elements, as a specific masking method, newly defined characters and symbols for masking A preferred method is to replace all masked portions with .

Also, the first advertisement data, the second advertisement data, and other advertisement data indicate elements other than moving images, still images, texts, voices, etc., which are creative elements. These advertisement data also indicate, for example, the age and gender of the target of advertisement, the budget of the target of advertisement, the delivery period of the advertisement, and the area where the advertisement is delivered, which are the elements of the targeting design of the advertisement.

The machine learning execution function 12 uses the first advertisement data and the second advertisement data to cause the machine learning device to learn, and based on the advertisement data, outputs the delivery performance estimation data indicating the estimation result of the advertisement delivery performance. Let the machine learning device generate an estimation program. The distribution record is an index that indicates the effectiveness of the distributed advertisement. For example, the machine learning execution function 12 inputs teacher data including first advertisement data, second advertisement data, and other advertisement data to a machine learning device to make the machine learning device learn.

FIG. 2 is a diagram illustrating an example of teacher data that the machine learning execution program according to the embodiment inputs to the machine learning device in order to cause the machine learning device to generate the distribution record estimation program. As shown in Figure 2, the training data includes images as creative elements, age and gender of the target audience as elements of targeting design, budget of the target audience, period of advertisement delivery, and area where the advertisement is delivered. Data that specifically designates each of them is the question, and the delivery record for each of these combinations is the data that is the answer. In other words, the training data is data in which the first advertisement data, the second advertisement data, and other advertisement data are questioned, and the results of advertisement distribution indicated by each piece of advertisement data are the answers.

Also, the score, which is an example of the answer of the training data, means the delivery record related to the cost spent on advertising, such as ROAS (Return on Advertising Spend), ROI (Return on Investment), CPM (Cost Per Mile), etc. there is Note that the score may mean, for example, CTR (Click Through Rate), CPA (Cost per Action), or the number of impressions.

The machine learning device performs learning by, for example, stochastic gradient descent. Since the machine learning device performs masking processing on the first advertisement data, the second advertisement data, and other advertisement data, while avoiding the gradient in the stochastic gradient descent method from affecting the encoder described later, learning can be carried out.

Moreover, since the machine learning device includes various data in the teacher data, it is preferable that the machine learning device is a multimodal model machine learning device. A multimodal model machine learning device is a machine learning device that learns based on multiple types of data.

3 and 4 are diagrams for explaining an example of a distribution record estimation program generated by the machine learning execution program according to the embodiment. The distribution performance estimation program includes, for example, multiple encoders and a neural network, as shown in FIGS. 3 and 4. FIG.

The encoder converts the portion of the advertisement data that indicates numerical values, categories, still images, moving images, or audio into data in a format that can be input to the neural network. Categories are information other than numerical values, still images, moving images, or audio. Both the encoders shown in FIGS. 3 and 4 are preferably input responsive. For example, the encoder is preferably a multi-layer perceptron when a numerical or categorical part of the advertisement data is input. Also, for example, the encoder is preferably a convolutional neural network when a portion of advertisement data indicating a moving image or a still image is input. Also, for example, when the portion of the advertisement data indicating the voice is input, the encoder is preferably a recursive neural network.

Note that the encoder shown in FIG. 3 to which the part showing the moving image is input, the encoder to which the part showing the audio shown in FIG. 3 is inputted, and the encoder to which the part showing the still image shown in FIG. 4 is inputted. The crosses shown on the right side of the indicate that these portions are masked.

Based on the data input from each of the five encoders, the neural network generates and outputs delivery performance estimation data indicating estimated delivery performance of the advertisement indicated by the advertisement data. The distribution performance data may indicate the estimation result of the score described above, or may indicate the estimation result of the distribution performance other than the score.

Next, an example of processing executed by the machine learning execution program 10 according to the embodiment will be described with reference to FIG. FIG. 5 is a flowchart illustrating an example of processing executed by a machine learning execution program according to the embodiment;

In step S11, the advertisement data acquisition function 11 acquires first advertisement data indicating a first advertisement having the first advertisement format and second advertisement data indicating a second advertisement having a second advertisement format different from the first advertisement format. to get

In step S12, the machine learning execution function 12 causes the machine learning device to learn using the first advertisement data and the second advertisement data, and based on the advertisement data indicating the advertisement, the delivery performance indicating the estimation result of the advertisement delivery performance. A machine learning device is caused to generate a delivery performance estimation program that outputs estimated data.

Next, an advertisement production support program according to the embodiment will be described with reference to FIGS. 6 to 12. FIG. FIG. 6 is a diagram illustrating an example of functions included in an advertisement production support program according to the embodiment; As shown in FIG. 6, the advertisement production support program 20 includes a condition specification data acquisition function 21, a delivered advertisement data acquisition function 22, a delivery performance data acquisition function 23, an undelivered advertisement data acquisition function 24, and a delivery performance It has an estimated data generation function 25 , a data output function 26 and a display control function 27 .

At least some of these functions are implemented by, for example, reading and executing the advertisement creation support program 20 stored in the storage device by a hardware processor such as a CPU. Note that the storage devices referred to here are, for example, hard disk drives, solid state drives, flash memories, and ROMs.

The condition specifying data acquisition function 21 acquires condition specifying data specifying conditions for estimating the delivery performance of undelivered advertisements. For example, the condition specifying data acquisition function 21 may acquire condition specifying data that specifies at least part of the targeting design of the undelivered advertisement as a condition. Alternatively, the condition specifying data acquisition function 21 acquires condition specifying data specifying at least a part of the creative of the undelivered advertisement as a condition.

The undelivered advertisement referred to here is an advertisement that has not been delivered, for example, an advertisement whose production has been completed and whether or not to deliver it is under consideration. Also, the advertisement referred to here includes at least one of online advertisement and offline advertisement. Online advertisements are, for example, retargeting advertisements and SNS (Social Networking Service) advertisements. Offline advertisements are, for example, advertisements broadcast on television or radio, advertisements published in magazines or newspapers.

The aforementioned conditions are imposed on at least part of the targeting design and creative of undelivered advertisements. Targeting design is design for narrowing down users to whom advertisements are distributed. For example, the targeting design includes at least one of the advertising medium that publishes the advertisement, the advertiser of the advertisement, the age, gender, preferences, and area of residence of the user targeted by the advertisement, the period during which the advertisement is distributed, and the user's browsing It narrows down the users to whom the advertisement is distributed from the viewpoint of the website etc. A creative is a production that includes at least one of moving images, still images, text, and audio.

For example, the condition specifying data acquisition function 21 acquires condition specifying data input using a user interface. 7 and 8 are diagrams showing an example of a user interface displayed on the display when condition specifying data according to the embodiment are input. The user interface shown in FIG. 7 includes an image M2, tab T1, tab T2, tab T3 and button B. FIG. The user interface shown in FIG. 8 additionally includes an image M31 and an image M32.

The image M2 displays the message "Please select an advertisement medium" by the user and clicks the tab T1, the tab T2 or the tab T3. displayed to prompt the user to select "Medium Y" or "Advertising Medium Z." Also, the advertising medium referred to here is, for example, a social networking service. Tab T1, tab T2, and tab T3 are clicked when advertising medium X, advertising medium Y, or advertising medium Z is selected as the conditions described above. Button B is used by the user when the undelivered advertisement data acquisition function 24 acquires undelivered advertisement data. The image M31 is clicked when the above-described advertiser "XX Co., Ltd." is selected. The image M32 is clicked when the above-described advertiser "XX Co., Ltd." is selected.

The condition-designated data acquisition function 21 conditionally designates data indicating "advertisement medium X", "advertisement medium Y" or "advertisement medium Z" selected by the user clicking on tab T1, tab T2 or tab T3. Get it as data. In addition, the condition-designated data acquisition function 21 acquires, as condition-designated data, data indicating "XX Co., Ltd." or "XX Co., Ltd." selected by the user clicking the image M31 or the image M32.

The delivered advertisement data acquisition function 22 acquires delivered advertisement data indicating a delivered advertisement that satisfies the conditions indicated by the condition specifying data and has a delivery record that satisfies a predetermined standard. Further, the delivery record may be calculated in advance, or may be calculated immediately before the delivered advertisement data acquisition function 22 acquires the delivery record data. Furthermore, the predetermined criteria referred to here are, for example, a threshold value set for the delivery record and a threshold value set for the ranking of the delivery record.

If the delivered advertisement is an online advertisement, the delivered advertisement data acquisition function 22 may acquire the delivered advertisement data directly from the Internet or from a database in which the delivered advertisement data is stored. good too. On the other hand, when the delivered advertisement is an online advertisement, the delivered advertisement data acquisition function 22 acquires the delivered advertisement data from a database in which the delivered advertisement data is stored, for example.

The delivery performance data acquisition function 23 acquires delivery performance data indicating the delivery performance of the delivered advertisement indicated by the delivered advertisement data.

FIG. 9 is a diagram illustrating an example of a display mode of a delivered advertisement indicated by delivered advertisement data and a delivery performance indicated by delivery performance data according to the embodiment. The user interface shown in FIG. 9 includes an image M4 and a display area D in addition to the tabs T1, T2, T3 and button B described above.

The image M4 is an image displaying the advertiser selected by clicking the image M31 or the image M32 described above. For example, the image M4 shown in FIG. 9 displays that "XX Co., Ltd." has been selected as the advertiser. The display area D is an area in which a character string indicating the target design of the delivered advertisement, the creative of the delivered advertisement, the delivery performance of the delivered advertisement, and the order of the delivery performance of the delivered advertisement are displayed. For example, in the display area D shown in FIG. "SCORE 0.8" indicating the distribution performance of the advertisement and "1st place" indicating the ranking of the distribution performance are displayed. Also, the creative S is an example of a video included in the video advertisement.

The undelivered advertisement data acquisition function 24 acquires undelivered advertisement data indicating an undelivered advertisement whose data format is different from that of the delivered advertisement. Further, when a plurality of pieces of delivered advertisement data are acquired, the undelivered advertisement data acquisition function 24 acquires undelivered advertisement data indicating an undelivered advertisement having a data format different from that of at least one of the delivered advertisements.

FIG. 10 is a diagram illustrating an example of a user interface displayed on the display when undelivered advertisement data is acquired by the advertisement production support program according to the embodiment; The user interface shown in FIG. 10 displays five undelivered advertisements, creative C1, creative C2, creative C3, creative C4, and creative C5.

Creative C1 is an example of a still image included in a still image advertisement. Creative C2 is an example of text included in a text advertisement. Creative C3 and creative C4 are examples of still images included in still image advertisements. Creative C5 is an example of a video included in a video advertisement. These five creatives are creatives included in undelivered advertisements that have the same targeting design as delivered advertisements including creative S.

For example, the undelivered advertisement data acquisition function 24 obtains the undelivered advertisement data and click data of the undelivered advertisement including the creative dragged to the upper center of the user interface shown in FIG. 10 among the five creatives shown in FIG. Get the undelivered ad data for the undelivered ad that contains the modified creative.

The distribution performance estimation data generation function 25 uses the distribution performance estimation program generated by learning the above-described machine learning device to estimate the delivery performance of undelivered advertisements under the conditions indicated by the condition specifying data. , to generate distribution performance estimation data indicating the estimation result of the distribution performance of the undelivered advertisement.

The data output function 26 outputs delivered advertisement data, delivery performance data, undelivered advertisement data, and estimated delivery performance data. For example, data output function 26 outputs these data to a display. Note that the data output function 26 may output these data to a device other than the display, such as a speaker.

The display control function 27 displays the delivered advertisement indicated by the delivered advertisement data, the delivery performance indicated by the delivery performance data, the undelivered advertisement indicated by the undelivered advertisement, and the delivery performance estimation result indicated by the delivery performance estimation data. control the display to For example, the display control function 27 controls the display to display these in the manner shown in FIG.

FIG. 11 shows an example of a display mode of a delivered advertisement displayed on the display by the advertisement production support program according to the embodiment, the delivery performance of the delivered advertisement, the undelivered advertisement, and the estimation result of the delivery performance of the undelivered advertisement. FIG. 10 shows. As shown in FIG. 11, the display control function 27 controls the display to display Creative C1, Creative C2, Creative C3, Creative C4, and Creative C5 in addition to the content shown in FIG. At the same time, the display control function 27 controls the display to display the estimation results of the delivery results in the lower right of each of these creatives and to display the order of the estimated delivery results in the lower left of each of these creatives.

In addition, the display control function 27 controls at least one of the delivered advertisement indicated by the delivered advertisement data and the delivery performance indicated by the delivery performance data, and the undelivered advertisement indicated by the undelivered advertisement and the delivery indicated by the estimated delivery performance data. The display may be controlled so that at least one of the performance estimation results is displayed in a different manner. For example, the display control function 27 sets the character string "SCORE0.9" and The color, size, etc. of the character string “1st place” indicating the order of delivery results in the five creatives may be different from the other character strings.

Next, an example of processing executed by the advertisement production support program 20 according to the embodiment will be described with reference to FIG. 12 . FIG. 12 is a flowchart illustrating an example of processing executed by an advertisement production support program according to the embodiment;

In step S21, the condition specification data acquisition function 21 acquires condition specification data that specifies conditions for estimating the delivery performance of undelivered advertisements.

In step S22, the delivered advertisement data acquisition function 22 acquires delivered advertisement data indicating a delivered advertisement that satisfies the conditions indicated by the condition designation data and has a delivery record that satisfies a predetermined standard.

In step S23, the delivery performance data acquisition function 23 acquires delivery performance data indicating the delivery performance of the delivered advertisement indicated by the delivered advertisement data acquired in step S22.

In step S24, the undelivered advertisement data acquisition function 24 acquires undelivered advertisement data indicating an undelivered advertisement whose data format is different from that of the delivered advertisement.

In step S25, the delivery performance estimation data generation function 25 uses the delivery performance estimation program to estimate the delivery performance of the undelivered advertisement under the conditions indicated by the condition specifying data acquired in step S21. Generating delivery performance estimation data indicating an estimation result of advertisement delivery performance.

In step S26, the data output function 26 outputs delivered advertisement data, delivery performance data, undelivered advertisement data, and estimated delivery performance data.

In step S27, the display control function 27 controls the delivered advertisement indicated by the delivered advertisement data, the delivery performance indicated by the delivery performance data, the undelivered advertisement indicated by the undelivered advertisement data, and the delivery performance indicated by the estimated delivery performance data. Control the display to show the estimation result of

The machine learning execution program 10 and the advertisement production support program 20 according to the embodiment have been described above.

The machine learning execution program 10 has an advertisement data acquisition function 11 and a machine learning execution function 12 . The advertisement data acquisition function 11 acquires first advertisement data indicating a first advertisement having a first advertisement format and second advertisement data indicating a second advertisement having a second advertisement format different from the first advertisement format. The first advertising data has the same data format as the second advertising data, with the portion relating to the second advertising format being masked. The second advertisement data has the same data format as the first advertisement data, with the portion relating to the first advertisement format being masked. The machine learning execution function 12 uses the first advertisement data and the second advertisement data to train the machine learning device, and outputs distribution performance estimation data indicating the estimation result of the advertisement distribution performance based on the advertisement data indicating the advertisement. A machine learning device is caused to generate a distribution performance estimation program for

As a result, the machine learning execution program 10 can handle a plurality of pieces of advertisement data as data having a unified data format, regardless of the specific aspect of the advertisement indicated by the advertisement data. Therefore, the machine learning execution program 10 generates a delivery performance estimation program capable of outputting delivery performance estimation data indicating delivery performance estimation results regardless of the specific aspect of the advertisement indicated by the advertisement data. can do.

The advertisement production support program 20 includes a condition specification data acquisition function 21, a delivered advertisement data acquisition function 22, a delivery performance data acquisition function 23, an undelivered advertisement data acquisition function 24, and a delivery performance estimation data generation function 25. and a data output function 26 .

The condition specifying data acquisition function 21 acquires condition specifying data specifying conditions for estimating the delivery performance of undelivered advertisements. The delivered advertisement data acquisition function 22 acquires delivered advertisement data indicating a delivered advertisement that satisfies the conditions indicated by the condition specifying data and has a delivery record that satisfies a predetermined standard. The delivery performance data acquisition function 23 acquires delivery performance data indicating the delivery performance of the delivered advertisement indicated by the delivered advertisement data.

The undelivered advertisement data acquisition function 24 acquires undelivered advertisement data indicating an undelivered advertisement whose data format is different from that of the delivered advertisement. The distribution performance estimation data generation function 25 uses first advertisement data indicating a first advertisement having a first advertisement format and second advertisement data indicating a second advertisement having a second advertisement format different from the first advertisement format. Using the program for estimating delivery performance generated by Generate. The data output function 26 outputs delivered advertisement data, delivery performance data, undelivered advertisement data, and estimated delivery performance data.

As a result, the advertisement production support program 20 uses the above-described delivery performance estimation program to provide a delivery performance estimation result that is unified regardless of the specific aspect of the advertisement indicated by the advertisement data. data can be obtained. In addition, this allows the advertisement production support program 20 to subjectively determine whether or not there is merit in distributing an undelivered advertisement compared to a delivered advertisement, regardless of the specific aspect of the advertisement indicated by the advertisement data. It is possible to present the grounds for judging from an objective point of view rather than from a physical point of view.

In the above-described embodiment, the case where each function shown in FIG. 1 is implemented by a hardware processor that reads and executes the machine learning execution program 10 has been described as an example, but the present invention is not limited to this.

At least part of the functions shown in FIG. It may be implemented by hardware including: Alternatively, at least part of the functions shown in FIG. 1 may be implemented by cooperation of software and hardware. In addition, these pieces of hardware may be integrated into one, or may be divided into a plurality of pieces.

Further, in the above-described embodiment, the case where each function shown in FIG. 6 is implemented by a hardware processor that reads and executes the advertisement production support program 20 has been described as an example, but the present invention is not limited to this.

At least part of the functions shown in FIG. 6 may be implemented by hardware including circuit units such as LSI, ASIC, FPGA, and GPU. Alternatively, at least part of the functions shown in FIG. 6 may be implemented by cooperation of software and hardware. In addition, these pieces of hardware may be integrated into one, or may be divided into a plurality of pieces.

The embodiments of the present invention have been described in detail above with reference to the drawings. However, the specific configurations of the embodiments of the present invention are not limited to the above-described embodiments, and at least one of various combinations, modifications, replacements, and design changes can be made without departing from the scope of the present invention. It may be added to the embodiment.

Moreover, the effects of the present invention described in the above-described embodiments are the effects shown as examples. Therefore, the present invention can produce other effects that can be understood from the description of the above embodiments in addition to the effects described above.

10... machine learning execution program, 11... advertisement data acquisition function, 12... machine learning execution function, 20... advertisement production support program, 21... condition specification data acquisition function, 22... delivered advertisement data acquisition function, 23... delivery performance data Acquisition function 24 Undelivered advertisement data acquisition function 25 Delivery result estimation data generation function 26 Data output function 27 Display control function

Claims (3)

to the computer,
a condition specification data acquisition function for acquiring condition specification data specifying conditions for estimating delivery results of undelivered advertisements;
a delivered advertisement data acquisition function for acquiring delivered advertisement data indicating a delivered advertisement that satisfies the condition indicated by the condition specifying data and has a delivery record that satisfies a predetermined standard;
a delivery performance data acquisition function for acquiring delivery performance data indicating the delivery performance of the delivered advertisement indicated by the delivered advertisement data;
an undelivered advertisement data acquisition function for acquiring undelivered advertisement data indicating the undelivered advertisement having a data format different from that of the delivered advertisement;
A delivery performance estimate generated using first ad data indicative of a first ad having a first ad type and second ad data indicative of a second ad having a second ad type different than the first ad type. A delivery record that uses a program to estimate the delivery record of the undelivered advertisement under the conditions indicated by the condition specifying data, and to generate delivery record estimation data indicating the estimation result of the delivery record of the undelivered advertisement. an estimated data generation function;
a data output function for outputting the delivered advertisement data, the delivery performance data, the undelivered advertisement data, and the delivery performance estimation data;
to realize
The delivery performance estimation data generation function includes the first advertisement data having the same data format as the second advertisement data and masking the portion related to the second advertisement format, and the same data format as the first advertisement data. and generating the delivery performance estimation data using the delivery performance estimation program generated using the second advertisement data in which the part related to the first advertisement format is masked;
Advertising production support program. a condition specification data acquisition unit for acquiring condition specification data specifying conditions for estimating delivery performance of undelivered advertisements;
a delivered advertisement data acquisition unit that acquires delivered advertisement data indicating a delivered advertisement that satisfies the condition indicated by the condition specifying data and has a delivery record that satisfies a predetermined criterion;
a delivery performance data acquisition unit for acquiring delivery performance data indicating a delivery performance of the delivered advertisement indicated by the delivered advertisement data;
an undelivered advertisement data acquisition unit for acquiring undelivered advertisement data indicating the undelivered advertisement having a data format different from that of the delivered advertisement;
A delivery performance estimate generated using first ad data indicative of a first ad having a first ad type and second ad data indicative of a second ad having a second ad type different than the first ad type. A delivery record that uses a program to estimate the delivery record of the undelivered advertisement under the conditions indicated by the condition specifying data, and to generate delivery record estimation data indicating the estimation result of the delivery record of the undelivered advertisement. an estimated data generator;
a data output unit that outputs the delivered advertisement data, the delivery performance data, the undelivered advertisement data, and the delivery performance estimation data;
with
The delivery performance estimation data generation unit has the same data format as the second advertisement data, and the first advertisement data in which the portion related to the second advertisement format is masked, and the same data format as the first advertisement data. and generating said delivery performance estimation data using said delivery performance estimation program generated using said second advertisement data in which a portion related to said first advertisement format is masked. Device. Acquiring condition-designated data specifying conditions for estimating delivery results of undelivered advertisements by means of the condition-designated data acquisition function or the condition-designated data acquisition unit;
acquiring delivered advertisement data indicating a delivered advertisement that satisfies the condition indicated by the condition specifying data and has a delivery track record that satisfies a predetermined standard by a delivered advertisement data acquisition function or a delivered advertisement data acquisition unit;
Acquiring delivery performance data indicating delivery performance of the delivered advertisement indicated by the delivered advertisement data by a delivery performance data acquisition function or a delivery performance data acquisition unit;
acquiring undelivered advertisement data indicating the undelivered advertisement having a data format different from that of the delivered advertisement by an undelivered advertisement data acquisition function or an undelivered advertisement data acquisition unit;
The first advertisement data indicating the first advertisement having the first advertisement format and the second advertisement having the second advertisement format different from the first advertisement format are provided by the distribution performance estimation data generation function or the delivery performance estimation data generation unit. estimating the delivery performance of the undelivered advertisement under the conditions indicated by the condition specifying data using a delivery performance estimation program generated using the second advertisement data, and delivering the undelivered advertisement; generating delivery performance estimation data indicating performance estimation results;
outputting the delivered advertisement data, the delivery performance data, the undelivered advertisement data and the estimated delivery performance data by a data output function or a data output unit ;
When generating the delivery performance estimation data, the first advertisement data having the same data format as the second advertisement data and masking the portion related to the second advertisement format, and the same as the first advertisement data generating the delivery performance estimation data using the delivery performance estimation program generated using the second advertisement data having a data format and masking the portion related to the first advertisement format;
Advertising production support method.

Images