Advertising effectiveness prediction device

Abstract

This advertising effect prediction device predicts an advertising effect of advertising content, the device comprising: an acquisition unit that acquires advertisement information pertaining to the advertising content; and a calculation unit that calculates the advertising effect on the basis of the advertisement information. The advertisement information includes a plurality of images included in the advertising content. The calculation unit calculates the advertising effect on the basis of the arrangement order of the plurality of images in the advertising content.

Description

Advertising effectiveness prediction device

[0001] The present disclosure relates to an advertising effectiveness prediction device.

[0002] In recent years, advertisements including links for transitioning to the advertiser's web page have been distributed. The effectiveness of such advertisements is indicated by indicators such as click-through rate and conversion rate. Patent Literature 1 describes an advertising effectiveness estimation model that inputs one advertising image and outputs advertising effectiveness such as click-through rate and conversion rate.

[0003] Japanese Patent Application Laid-Open No. 2018-77615

[0004] An advertisement may contain multiple images. The advertising effectiveness estimation model described in Patent Document 1 allows only one image to be input, and therefore cannot take multiple images into consideration. It is therefore desirable to more accurately predict the advertising effectiveness of such advertisements.

[0005] The present disclosure describes an advertising effectiveness prediction device that can improve the accuracy of predicting advertising effectiveness.

[0006] An advertising effectiveness prediction device according to one aspect of the present disclosure is a device for predicting advertising effectiveness of advertising content. The advertising effectiveness prediction device includes an acquisition unit that acquires advertising information related to the advertising content, and a calculation unit that calculates advertising effectiveness based on the advertising information. The advertising information includes multiple images included in the advertising content. The calculation unit calculates the advertising effectiveness based on the arrangement order of the multiple images in the advertising content.

[0007] In this advertising effectiveness prediction device, the advertising effectiveness is calculated based on the arrangement order of multiple images in advertising content. It is considered that the target recipients who receive advertising content will view the advertising content in order from the beginning. Therefore, it is highly likely that the multiple images included in the advertising content will come into the target recipients' field of view in the order in which they are arranged in the advertising content. In such cases, it is considered that the order of the images will affect the target recipients' interest. In the above advertising effectiveness prediction device, the advertising effectiveness is predicted taking into account the arrangement order of multiple images in the advertising content, thereby enabling improved prediction accuracy of the advertising effectiveness of advertising content.

[0008] According to the present disclosure, it is possible to improve the accuracy of predicting advertising effectiveness.

[0009] FIG. 1 is a block diagram showing the functional configuration of an advertising effectiveness prediction device according to an embodiment. FIG. 2 is a diagram illustrating image acquisition processing and image processing. FIG. 3 is a diagram illustrating the configuration of a prediction model. FIG. 4 is a diagram illustrating the image model shown in FIG. 3. FIG. 5 is a diagram illustrating the configuration of a convolution LSTM (Long Short-Term Memory) shown in FIG. 4. FIG. 6 is a flowchart showing a series of processes of an advertising effectiveness prediction method performed by the advertising effectiveness prediction device shown in FIG. 1. FIG. 7(a) is a diagram illustrating the relationship between a CTR predicted by the advertising effectiveness prediction device shown in FIG. 1 and an actual CTR. FIG. 7(b) is a diagram illustrating the relationship between a CTR predicted by an advertising effectiveness prediction device of a comparative example and an actual CTR. FIG. 8 is a diagram illustrating the hardware configuration of the advertising effectiveness prediction device shown in FIG. 1.

[0010] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are designated by the same reference numerals, and duplicated explanations will be omitted.

[0011] The configuration of an advertising effectiveness prediction device according to one embodiment will be described with reference to Figures 1 to 5. Figure 1 is a block diagram showing the functional configuration of an advertising effectiveness prediction device according to one embodiment. Figure 2 is a diagram for explaining image acquisition processing and image processing. Figure 3 is a diagram for explaining the configuration of a prediction model. Figure 4 is a diagram for explaining the image model shown in Figure 3. Figure 5 is a diagram for explaining the configuration of the convolutional LSTM shown in Figure 4.

[0012] The advertising effectiveness prediction device 10 shown in FIG. 1 is a device that predicts the advertising effectiveness of advertising content (advertising copy). The advertising content includes, for example, images and text. An example of advertising content is an advertisement distributed by email. The advertising content is written in, for example, HTML (HyperText Markup Language). Examples of indices that indicate advertising effectiveness include a click-through rate (CTR), a conversion rate (CVR), the number of clicks, and a return on advertising spend (ROAS). An example of the advertising effectiveness prediction device 10 is an information processing device such as a server device.

[0013] The advertising effect prediction device 10 functionally comprises an acquisition unit 11, a processing unit 12, a calculation unit 13, and an output unit 14.

[0014] The acquisition unit 11 is a functional unit that acquires advertising information related to advertising content. The advertising information includes one or more images included in the advertising content. The acquisition unit 11 acquires all images included in the advertising content, for example, by scraping the URL (Uniform Resource Locator) of the advertising content. As shown in FIG. 2 , the acquisition unit 11 acquires one or more images in the order (display order) of the images in the advertising content. The acquisition unit 11 may set the order of images in the source code of the advertising content as the order of images in the advertising content.

[0015] The advertisement information may further include text included in the advertisement content. Examples of the text include the title of the advertisement content and the main text of the advertisement content. The acquisition unit 11 may acquire the text included in the advertisement content by, for example, scraping the URL of the advertisement content. The acquisition unit 11 may acquire the title of the advertisement content from a memory (not shown).

[0016] The advertising information may further include distribution information. The distribution information is information related to the distribution of advertising content. The distribution information includes, for example, target information related to distribution targets, distribution number information related to the number of distributions, and budget information related to the budget. Examples of target information include a minimum age, a maximum age, and gender. Examples of distribution number information include the number of distributions per week, the number of distributions per day of the week, and the number of distributions per hour. Examples of budget information include a daily budget and a monthly budget. The acquisition unit 11 acquires the distribution information from, for example, a memory (not shown). Note that various information such as the URL, title, and distribution information of the advertising content is, for example, supplied from an external source and is pre-stored for each advertising content in a memory (not shown).

[0017] The processing unit 12 is a functional unit that processes advertising information. The processing unit 12 processes the advertising information acquired by the acquisition unit 11 into a format that can be input into a prediction model, which will be described later. The processing unit 12 performs the following processing on images, for example. Since the sizes (number of pixels) of images included in advertising content vary, the processing unit 12 changes (resizes) the size (number of pixels) of each image to a predetermined size (number of pixels). The image resizing is performed using a known method.

[0018] When the number of images obtained from one piece of advertising content is less than a specified number N (N is an integer equal to or greater than 2), the processing unit 12 adds dummy images to adjust the number of images to the specified number N. The processing unit 12 adds the missing dummy images after the images obtained from the advertising content. In other words, the dummy images are arranged after the images included in the advertising content that are arranged in order.

[0019] The processing unit 12 uses, for example, a blackened image (an image in which all pixel values ​​are 0) as the dummy image. The processing unit 12 may use the immediately preceding image as the dummy image, or may use an image obtained by averaging the pixel values ​​of multiple images included in the advertising content as the dummy image. Note that, in order to improve prediction accuracy, the type of dummy image may be matched to the type of dummy image used to train the prediction model M. For example, if the prediction model M is trained using a blackened image as the dummy image, the processing unit 12 may use the blackened image as the dummy image. Similarly, if the prediction model M is trained using the immediately preceding image as the dummy image, the processing unit 12 may use the immediately preceding image as the dummy image.

[0020] In the example shown in FIG. 2, advertising content AC includes five images G1 to G5. Images G1 to G5 are arranged from top to bottom of advertising content AC in the order of image G1, image G2, image G3, image G4, and image G5. In this example, the specified number N is set to 6. In this case, images G1 to G5 are acquired by the acquisition unit 11, and the size of each image is changed to a predetermined size by the processing unit 12. In this example, images G1 to G4 are reduced and changed to images G11 to G14, and image G5 is enlarged and changed to image G15. Furthermore, a dummy image is added by the processing unit 12 as image G16.

[0021] The processing unit 12 performs the following processing on the text, for example: The processing unit 12 divides the text into words by performing morphological analysis, assigns an index to each word, and further vectorizes each word; The processing unit 12 adjusts the number of characters in the text to a specified number; If the number of characters in the text is less than the specified number, the processing unit 12 adjusts the number of characters in the text to the specified number by performing padding (for example, adding zeros); The processing unit 12 performs the following processing on the distribution information, for example: The processing unit 12 normalizes the distribution information and vectorizes the normalized distribution information.

[0022] The calculation unit 13 is a functional unit that calculates the advertising effectiveness of advertising content based on advertising information. As shown in FIG. 3, the calculation unit 13 includes a prediction model M for predicting advertising effectiveness. The prediction model M is a machine learning model that uses each piece of advertising information as an explanatory variable and the advertising effectiveness as a target variable, and is configured, for example, by a neural network. The prediction model M is generated by performing machine learning. In the machine learning, for example, a pair of advertising information of advertising content for which an actual measured value of advertising effectiveness has been obtained and the actual measured value of advertising effectiveness is used as ground truth data. The prediction model M includes a delivery information model Mb, an image model Mi, a text model Mt, and a combined model Mc.

[0023] The distribution information model Mb receives as input (feature quantities) of the distribution information processed by the processing unit 12, and outputs feature quantities of the entire distribution information. The feature quantities of the entire distribution information are, for example, important parts of the entire distribution information. In the distribution information model Mb, processing is performed in the following order: linear transformation, calculation using a ReLU (Rectified Linear Unit) function, linear transformation, and calculation using a Tanh function.

[0024] The image model Mi receives as input (feature quantities) of N images processed by the processing unit 12, and outputs feature quantities of the entire image. The feature quantities of the entire image are, for example, important parts of the entire image. In this embodiment, the image model Mi is configured by a convolutional long short-term memory (ConvLSTM). The convolutional LSTM is a recurrent neural network (RNN) in which the linear operation of the LSTM is replaced with a convolution operation. The LSTM is a neural network that receives each element of time-series data in which multiple elements are arranged as input in order, and is configured so that the influence of previously input elements affects the output.

[0025] As shown in FIGS. 4 and 5, the convolutional LSTM 30 calculates an input X at time t. ｔ and the cell state C at the previous time (t-1) ｔ－１and output h ｔ－１ From this, the cell state C at time t ｔ and output h ｔ Specifically, the forget gate 31 outputs the input X ｔ and output h ｔ－１ and outputs f ｔ The forget gate 31 outputs the weight matrix W ｆ , weight matrix R ｆ , and bias b ｆ By performing the calculation shown in equation (1) using ｔ The weight matrix W ｆ , weight matrix R ｆ , and bias b ｆ is preset in the convolutional LSTM 30. The weight matrix W ｆ Input X ｔ A convolution operation is performed to convolve the weight matrix R ｆ Output to h ｔ－１ A convolution operation is performed to convolve the following. The function σ(·) represents the sigmoid function.

[0026] The input gate 32 receives the input X ｔ and output h ｔ－１ and outputs i ｔ The input gate 32 outputs the weight matrix W ｉ , weight matrix R ｉ , and bias b ｉ By performing the calculation shown in equation (2) using ｔ The weight matrix W ｉ , weight matrix R ｉ , and bias b ｉ is preset in the convolutional LSTM 30. The weight matrix W ｉ Input X ｔ A convolution operation is performed to convolve the weight matrix R ｉ Output to h ｔ－１ A convolution operation is performed to fold the

[0027] The tanh layer 33 receives the input X ｔ and output h ｔ－１ and receives a vector C' of state values. ｔThe tanh layer 33 outputs a weight matrix W ｃ , weight matrix R ｃ , and bias b ｃ By performing the operation shown in equation (3) using ｔ The weight matrix W ｃ , weight matrix R ｃ , and bias b ｃ is preset in the convolutional LSTM 30. The weight matrix W ｃ Input X ｔ A convolution operation is performed to convolve the weight matrix R ｃ Output to h ｔ－１ A convolution operation is performed to fold the

[0028] At the nodes 34 to 36, the operation shown in equation (4) is performed to obtain the cell state C ｔ Specifically, at node 34, the cell state C ｔ－１ and output f ｔ By multiplying by ｔ－１ The information in is selected and discarded. Output f ｔ is the cell state C ｔ－１ The output f has a value in the range of 0 to 1 for all values ​​of ｔ If the value of is 1, the cell state C ｔ－１ The value of is completely maintained, and the output f ｔ If the value of is 0, the cell state C ｔ－１ Similarly, at node 35, the value of vector C' ｔ and output i ｔ and are multiplied to obtain vector C' ｔ The state values ​​in the output i ｔ is the vector C' ｔ , the result of the operation at node 34 is added to the result of the operation at node 35, and the cell state C ｔ is obtained.

[0029] The output gate 37 receives the input X ｔ and output hｔ－１ and outputs o ｔ The output gate 37 outputs the weight matrix W ｏ , weight matrix R ｏ , and bias b ｏ By performing the calculation shown in equation (5) using ｔ The weight matrix W ｏ , weight matrix R ｏ , and bias b ｏ is preset in the convolutional LSTM 30. The weight matrix W ｏ Input X ｔ A convolution operation is performed to convolve the weight matrix R ｏ Output to h ｔ－１ A convolution operation is performed to fold the

[0030] The tanh layer 38 and the node 39 perform the operation shown in equation (6), and the output h ｔ Specifically, the tanh layer 38 is in cell state C ｔ and receives the cell state C ｔ By applying the tanh function to the output o, each value is set to fall within the range of -1 to 1. Then, at node 39, ｔ and the calculation result of the tanh layer 38 are multiplied to obtain the output h ｔ The convolutional LSTM 30 receives the input X ｔ 4, N convolutional LSTMs 30 are connected in series, but the image is simply input to one convolutional LSTM 30 in a recursive manner.

[0031] Here, the calculation unit 13 arranges the N images received from the processing unit 12 in the order of arrangement in the advertising content, and calculates the feature amounts of the first image, the second image, ..., and the Nth image as input X １ , input X ２ , ...input X Ｎ The image model Mi outputs h Ｎis output as the feature of the entire image.

[0032] The text model Mt receives as input the text processed by the processing unit 12 and outputs the feature quantities of the entire text. The feature quantities of the entire text are, for example, important parts of the entire text. In the text model Mt, processing is performed, for example, in the order of embedding, convolution, calculation using the ReLU function, MaxPooling, and calculation using the ReLU function.

[0033] The combined model Mc receives as input the output of the distribution information model Mb, the output of the image model Mi, and the output of the text model Mt, and outputs the advertising effect. In the combined model Mc, processing is performed in the following order, for example: batch normalization, dropout, linear transformation, calculation using the ReLU function, linear transformation, calculation using the Tanh function, linear transformation, and calculation using the sigmoid function.

[0034] As described above, the calculation unit 13 calculates the advertising effect by inputting the N images one by one into the convolution LSTM 30 in the order in which the N images are arranged in the advertising content.

[0035] The output unit 14 is a functional unit that outputs information indicating the advertising effectiveness. The output unit 14 may, for example, output the information indicating the advertising effectiveness to a display device (not shown) and cause the display device to display the advertising effectiveness.

[0036] Next, an advertising effectiveness prediction method performed by the advertising effectiveness prediction device 10 will be described with reference to Fig. 6. Fig. 6 is a flowchart showing a series of processes of the advertising effectiveness prediction method performed by the advertising effectiveness prediction device shown in Fig. 1. The series of processes shown in Fig. 6 is started, for example, when a user sets information capable of identifying advertising content for which advertising effectiveness is to be predicted in the advertising effectiveness prediction device 10.

[0037] 6 , first, the acquisition unit 11 acquires advertising information (step S11). In step S11, the acquisition unit 11 acquires, as advertising information, for example, all images included in the advertising content, the title of the advertising content, and distribution information of the advertising content. Then, the acquisition unit 11 outputs the advertising information to the processing unit 12.

[0038] Next, the processing unit 12 processes the advertising information (step S12). In step S12, when the processing unit 12 receives the advertising information from the acquisition unit 11, the processing unit 12 processes the advertising information into a format that can be input to the prediction model M. For example, the processing unit 12 changes the size of each image to a predetermined size, and if the number of images obtained from one advertising content is less than the specified number N, adds dummy images to adjust the number of images to the specified number N. The processing unit 12 performs morphological analysis to divide the text into words, assigns an index to each word, and further vectorizes each word. The processing unit 12 normalizes the distribution information and vectorizes the normalized distribution information. Then, the processing unit 12 outputs the processed advertising information to the calculation unit 13.

[0039] Next, the calculation unit 13 calculates the advertising effectiveness of the advertising content (step S13). In step S13, when the calculation unit 13 receives the processed advertising information from the processing unit 12, the calculation unit 13 inputs the processed advertising information into the prediction model M. Specifically, the calculation unit 13 inputs the processed distribution information into the distribution information model Mb, inputs the N processed images into the image model Mi, and inputs the processed text into the text model Mt. The calculation unit 13 inputs the N images one by one in the order of the advertising content into the convolution LSTM 30 of the image model Mi. The combined model Mc then receives as input the output of the distribution information model Mb, the output of the image model Mi, and the output of the text model Mt, and outputs the advertising effectiveness. The calculation unit 13 then outputs information indicating the advertising effectiveness to the output unit 14.

[0040] Next, the output unit 14 outputs the advertising effect (step S14). In step S14, when the output unit 14 receives the information indicating the advertising effect from the calculation unit 13, the output unit 14 outputs the information indicating the advertising effect to, for example, a display device (not shown), and causes the display device to display the advertising effect.

[0041] This completes the series of processes in the advertising effectiveness prediction method.

[0042] Next, the effects of the advertising effectiveness prediction device 10 will be described with reference to (a) and (b) of Figure 7. (a) of Figure 7 is a diagram showing the relationship between the CTR predicted by the advertising effectiveness prediction device shown in Figure 1 and the measured CTR. (b) of Figure 7 is a diagram showing the relationship between the CTR predicted by the advertising effectiveness prediction device of the comparative example and the measured CTR. The advertising effectiveness prediction device of the comparative example differs from the advertising effectiveness prediction device 10 mainly in that the image model Mi of the prediction model M is configured using a Residual Network (Resnet) instead of a convolutional LSTM, and in that the image model Mi receives only the first image included in the advertising content as input instead of all images included in the advertising content. The learning rate is 1e-6 in both cases.

[0043] As shown in (b) of Figure 7, the predicted CTR predicted by the advertising effect prediction device of the comparative example is greater than the actually measured CTR. The prediction accuracy (mean square error) by the advertising effect prediction device of the comparative example was 4.57e-4. On the other hand, as shown in (a) of Figure 7, the predicted CTR predicted by the advertising effect prediction device 10 is closer to the actually measured CTR than the advertising effect prediction device of the comparative example. The prediction accuracy (mean square error) by the advertising effect prediction device 10 was 3.12e-4. Therefore, compared to the advertising effect prediction device of the comparative example, the mean square error is reduced by approximately 32% in the advertising effect prediction device 10.

[0044] The advertising effectiveness prediction device 10 described above calculates the advertising effectiveness based on the arrangement order of multiple images in advertising content. It is considered that a target recipient who receives advertising content will view the advertising content in order from the beginning of the advertising content. Therefore, it is highly likely that the multiple images included in the advertising content will be in the target recipient's field of view in the order in which they are arranged in the advertising content. In such a case, it is considered that the order of the images will affect the target recipient's interest. The advertising effectiveness prediction device 10 predicts the advertising effectiveness taking into account the arrangement order of multiple images in the advertising content, thereby improving the accuracy of predicting the advertising effectiveness of the advertising content. In other words, the advertising effectiveness prediction device 10 can express both the number of images and the quality of the advertising content as feature quantities, thereby improving the accuracy of predicting the advertising effectiveness of the advertising content. Accurate prediction of advertising effectiveness allows an appropriate submission price to be set for the advertising content.

[0045] The calculation unit 13 includes a prediction model M, which is a machine learning model that uses advertising information as an explanatory variable and advertising effectiveness as a target variable. Therefore, advertising effectiveness can be obtained simply by inputting advertising information into the prediction model M.

[0046] The calculation unit 13 calculates the advertising effectiveness by inputting the multiple images included in the advertising content one by one in the order of their arrangement into the convolutional LSTM 30. This configuration makes it possible to capture the feature quantities of the images while allowing the influence of previously input images to be exerted on the output. Therefore, it is possible to take into account the order of the multiple images included in the advertising content, thereby improving the accuracy of predicting the advertising effectiveness.

[0047] A predetermined number N of images are input to the prediction model M. Therefore, if the number of images included in the advertising content is less than the predetermined number N, the processing unit 12 adjusts the number of images to the predetermined number N by adding dummy images. This configuration adjusts the number of images to the input of the prediction model M, allowing for appropriate predictions using the prediction model M. It is believed that advertising effectiveness, such as click-through rates, will change (increase) significantly when the number of images exceeds a certain number. This number may vary depending on the media in which the advertising content is used. Therefore, by measuring in advance the number of images that significantly changes advertising effectiveness and setting the predetermined number N to this number (or greater), the impact of the number of images on advertising effectiveness can be extracted as a feature. With this configuration, whether or not the advertising content contains more than the predetermined number N of images can be reflected in the advertising effectiveness. Furthermore, in this embodiment, the prediction model M performs a convolution operation on each input image, allowing important areas within the image to be extracted. Therefore, the impact of the quality of the images included in the advertising content on advertising effectiveness can be extracted as a feature. As a result, the accuracy of predicting advertising effectiveness can be further improved.

[0048] Of the multiple images included in the advertising content, the target audience's gaze may stop on the last image. In this way, the impression of the image immediately before the dummy image is likely to have a significant impact on the target audience's interest. Therefore, when the image immediately before the dummy image is used as the dummy image, it is possible to further improve the accuracy of predicting the advertising effectiveness for target audiences who perform the above-mentioned browsing behavior.

[0049] After a target user finishes viewing advertising content, the target user may look over the entire advertising content. This behavior of looking over the entire advertising content can be simulated by averaging the pixel values ​​of multiple images included in the advertising content. Therefore, when an image obtained by averaging the pixel values ​​of multiple images included in the advertising content is used as a dummy image, it is possible to further improve the accuracy of predicting advertising effectiveness for target users who perform the above-mentioned browsing behavior.

[0050] The processing unit 12 changes the number of pixels of the multiple images to a predetermined number of pixels. By making the number of pixels of the multiple images the same, each image can be treated under the same conditions. Therefore, predictions can be made appropriately using the prediction model M. As a result, it is possible to further improve the prediction accuracy of advertising effectiveness.

[0051] Not only images but also text contained in the advertising content may attract the interest of the target audience. When the advertising information further includes text contained in the advertising content, the text is further taken into consideration when predicting the advertising effectiveness. Therefore, it is possible to further improve the accuracy of predicting the advertising effectiveness.

[0052] It is believed that the title of the advertising content (the subject of the advertising email) determines to some extent whether the target recipient will be attracted to the advertising content. If the advertising information includes the title of the advertising content as text, the title of the advertising content is further taken into consideration when predicting the advertising effectiveness. Therefore, it is possible to further improve the accuracy of predicting the advertising effectiveness.

[0053] Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above embodiments.

[0054] The advertising effect prediction device 10 may be configured as a single device that is physically or logically coupled, or may be configured as multiple devices that are physically or logically separated from each other. For example, the advertising effect prediction device 10 may be realized by multiple computers that are distributed over a network, such as in cloud computing. As described above, the configuration of the advertising effect prediction device 10 may include any configuration that can realize the functions of the advertising effect prediction device 10.

[0055] The advertising effectiveness prediction device 10 may not include the processing unit 12. The acquisition unit 11 may acquire only images included in advertising content as advertising information, or may acquire at least one of delivery information and text in addition to images as advertising information. The prediction model M may not include at least one of the delivery information model Mb and the text model Mt, depending on the acquired advertising information.

[0056] If the number of images obtained from one advertising content is less than a specified number N, the processing unit 12 may add a dummy image at any position, not just after the images included in the advertising content arranged in order.

[0057] The image model Mi may be configured by a convolutional neural network (CNN) and an LSTM instead of the convolutional LSTM 30. For example, each image is input to the CNN, and the output of the CNN for each image is input to the LSTM in the order of the images in the advertising content, thereby obtaining features that take into account the entire image and the order.

[0058] The block diagrams used to explain the above embodiments show functional blocks. These functional blocks (components) are realized by any combination of at least one of hardware and software. The method for realizing each functional block is not particularly limited. That is, each functional block may be realized using a single device that is physically or logically coupled, or may be realized using two or more physically or logically separated devices that are directly or indirectly connected (e.g., wired, wireless, etc.) and these multiple devices. The functional block may be realized by combining the single device or multiple devices with software.

[0059] Functions include, but are not limited to, judgment, determination, judgment, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, regard, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating, mapping, and assignment. For example, a functional block (component) that performs transmission is called a transmitting unit or transmitter. As mentioned above, there are no particular limitations on how these functions are implemented.

[0060] For example, the advertising effect prediction device 10 according to an embodiment of the present disclosure may function as a computer that performs the processing of the present disclosure. Fig. 8 is a diagram illustrating an example of a hardware configuration of an advertising effect prediction device according to an embodiment of the present disclosure. The above-described advertising effect prediction device 10 may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, etc.

[0061] In the following description, the term "device" can be interpreted as a circuit, a device, a unit, etc. The hardware configuration of the advertising effect prediction device 10 may be configured to include one or more of the devices shown in the figure, or may be configured to exclude some of the devices.

[0062] Each function of the advertising effectiveness prediction device 10 is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, causing the processor 1001 to perform calculations, control communication via the communication device 1004, and control at least one of reading and writing of data in the memory 1002 and storage 1003.

[0063] The processor 1001 controls the entire computer by running, for example, an operating system. The processor 1001 may be configured by a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, and a register. For example, each function of the above-described advertising effect prediction device 10 may be realized by the processor 1001.

[0064] The processor 1001 reads programs (program codes), software modules, data, etc. from at least one of the storage 1003 and the communication device 1004 into the memory 1002, and executes various processes in accordance with these. The program used is a program that causes a computer to execute at least some of the operations described in the above-described embodiments. For example, each function of the advertising effect prediction device 10 may be realized by a control program stored in the memory 1002 and running on the processor 1001. While the above-described various processes have been described as being executed by one processor 1001, they may also be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may also be transmitted from a network via a telecommunications line.

[0065] The memory 1002 is a computer-readable recording medium and may be configured by, for example, at least one of a read-only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), and a random access memory (RAM). The memory 1002 may also be referred to as a register, a cache, a main memory, or the like. The memory 1002 may store executable programs (program codes), software modules, and the like for implementing an advertising effectiveness prediction method according to an embodiment of the present disclosure.

[0066] Storage 1003 is a computer-readable recording medium, and may be composed of at least one of, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy (registered trademark) disk, and a magnetic strip. Storage 1003 may also be referred to as an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, a server, or other appropriate medium including at least one of memory 1002 and storage 1003.

[0067] The communication device 1004 is hardware (transmission / reception device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as, for example, a network device, a network controller, a network card, or a communication module. The communication device 1004 may be configured to include a high-frequency switch, a duplexer, a filter, a frequency synthesizer, etc. to realize at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the above-mentioned acquisition unit 11 and output unit 14 may be realized by the communication device 1004.

[0068] The input device 1005 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that receives input from the outside. The output device 1006 is an output device (e.g., a display, a speaker, an LED lamp, etc.) that outputs to the outside. Note that the input device 1005 and the output device 1006 may be integrated into one device (e.g., a touch panel).

[0069] The processor 1001, memory 1002, and other devices are connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or may be configured using different buses for each device.

[0070] The advertising effect prediction device 10 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), and a field programmable gate array (FPGA), and some or all of the functional blocks may be realized by the hardware. For example, the processor 1001 may be implemented using at least one of these pieces of hardware.

[0071] Notification of information is not limited to the aspects / embodiments described in this disclosure, and may be performed using other methods.

[0072] In the processing procedures, sequences, flowcharts, etc. of each aspect / embodiment described in this disclosure, the order of processing may be changed unless there is a contradiction. For example, the methods described in this disclosure present elements of various steps using an example order, and are not limited to the specific order presented.

[0073] Information etc. may be output from a higher layer to a lower layer, or may be output from a lower layer to a higher layer. Information etc. may be input / output via multiple network nodes.

[0074] Input and output information may be stored in a specific location (for example, memory) or may be managed using a management table. Input and output information may be overwritten, updated, or added to. Output information may be deleted. Input information may be transmitted to another device.

[0075] The determination may be made based on a value represented by one bit (0 or 1), a Boolean value (true or false), or a numerical comparison (e.g., comparison with a predetermined value).

[0076] The aspects / embodiments described in the present disclosure may be used alone, in combination, or switched depending on the implementation. Notification of predetermined information (e.g., notification that "X is true") is not limited to explicit notification, but may be performed implicitly (e.g., by not notifying the predetermined information).

[0077] Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is intended to be illustrative and does not have any limiting meaning on the present disclosure.

[0078] Software shall be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.

[0079] Software, instructions, information, etc. may be transmitted or received over a transmission medium. For example, if software is transmitted from a website, server, or other remote source using wired technologies (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL)), and / or wireless technologies (such as infrared, microwave), these wired and / or wireless technologies are included within the definition of transmission media.

[0080] The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

[0081] In addition, terms explained in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings.

[0082] As used in this disclosure, the terms "system" and "network" are used interchangeably.

[0083] The information, parameters, etc. described in this disclosure may be expressed using absolute values, relative values ​​from a predetermined value, or corresponding other information.

[0084] The names used for the parameters described above are not limiting in any way, and the formulas using these parameters may differ from those explicitly disclosed in this disclosure.

[0085] As used in this disclosure, the terms "determining" and "determining" may encompass a wide variety of actions. "Determining" and "determining" may each include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, search, inquiry (e.g., looking up in a table, database, or another data structure), and ascertaining. "Determining" and "determining" may each include receiving (e.g., receiving information), transmitting (e.g., sending information), input, output, and accessing (e.g., accessing data in memory). "Determining" and "determining" may each include resolving, selecting, choosing, establishing, and comparing. That is, "determining" may include any action that can be considered a "judgment." A "decision" may include any action that can be considered a "decision." "Judgment (decision)" may be read as "assuming," "expecting," or "considering," etc.

[0086] The terms "connected," "coupled," or any variation thereof, refer to any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are "connected" or "coupled" to each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, "connected" may be read as "access." When "connected" or "coupled" is used in this disclosure, two elements may be considered to be "connected" or "coupled" to each other using at least one of an electrical wire, cable, and printed electrical connection, or may be considered to be "connected" or "coupled" to each other using electromagnetic energy having wavelengths in the radio frequency range, microwave range, and optical (both visible and invisible) range, as some non-limiting and non-exhaustive examples.

[0087] As used in this disclosure, the phrase "based on" does not mean "based only on," unless expressly stated otherwise. In other words, the phrase "based on" means both "based only on" and "based at least on."

[0088] As used in this disclosure, any reference to an element using designations such as "first" and "second" does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient method of distinguishing between two or more elements. Thus, a reference to a first and a second element does not imply either that only two elements may be employed or that the first element must in some way precede the second element.

[0089] The "unit" in the configuration of each of the above devices may be replaced with a "circuit" or a "device", etc.

[0090] When the terms "include," "including," and variations thereof are used in this disclosure, these terms are intended to be inclusive, similar to the term "comprising." Furthermore, when the term "or" is used in this disclosure, it is not intended to be an exclusive or.

[0091] In this disclosure, where articles are added by translation, such as "a," "an," and "the" in English, the disclosure may include that the nouns following these articles are plural.

[0092] In the present disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "coupled" may also be interpreted in the same way as "different."

[0093] 10...advertising effectiveness prediction device, 11...acquisition unit, 12...processing unit, 13...calculation unit, 14...output unit, 30...convolution LSTM, 1001...processor, 1002...memory, 1003...storage, 1004...communication device, 1005...input device, 1006...output device, 1007...bus, M...prediction model, Mb...model for distributed information, Mc...combined model, Mi...model for image, Mt...model for text.

Claims

1. An advertising effectiveness prediction device that predicts the advertising effectiveness of advertising content, comprising: an acquisition unit that acquires advertising information related to the advertising content; and a calculation unit that calculates the advertising effectiveness based on the advertising information, wherein the advertising information includes a plurality of images included in the advertising content, and the calculation unit calculates the advertising effectiveness based on the arrangement order of the plurality of images in the advertising content.

2. The advertising effectiveness prediction device according to claim 1, wherein the calculation unit includes a prediction model that is a machine learning model in which the advertising information is an explanatory variable and the advertising effectiveness is an objective variable.

3. The advertising effectiveness prediction device of claim 2, wherein the prediction model includes a convolutional LSTM, and the calculation unit calculates the advertising effectiveness by inputting the multiple images into the convolutional LSTM one image at a time in the arranged order.

4. An advertising effectiveness prediction device as described in claim 2 or claim 3, further comprising a processing unit that processes the advertising information, wherein if the number of the plurality of images is less than a specified number, the processing unit adjusts the number of the plurality of images to the specified number by adding a dummy image to the plurality of images.

5. The advertising effect prediction device according to claim 4, wherein the dummy image is an image immediately preceding the dummy image.

6. The advertising effect prediction device according to claim 4, wherein the dummy image is an image obtained by averaging pixel values ​​of the plurality of images.

7. The advertising effect prediction device according to any one of claims 4 to 6, wherein the processing unit changes the number of pixels of the plurality of images to a predetermined number of pixels.

8. An advertising effectiveness prediction device according to any one of claims 1 to 7, wherein the advertising information further includes text contained in the advertising content.

9. The advertising effect prediction device according to claim 8, wherein the text is a title of the advertising content.

Images