Advertising media evaluation device, advertising media evaluation method, and advertising media evaluation program

Abstract

To enable proper evaluation of the effectiveness of advertising media. [Solution] The visibility probability estimation unit 223 estimates the visibility probability for the advertising medium for each of several routes that pass through an area where the advertising medium for estimation is visible. The number of viewers calculation unit 224 calculates the number of people who viewed the advertising medium for estimation along each of the several routes by multiplying the traffic volume for the route by the visibility probability estimation unit 223 for that route, and then calculates the number of people who viewed the advertising medium for estimation from the number of people calculated for each of the several routes.

Description

Technical Field

[0001] The present disclosure relates to a technique for evaluating the effect of an advertising medium.

Background Art

[0002] Attempts have been made to evaluate the effect of outdoor advertisements displayed on bulletin boards, displays, etc. posted on buildings. In recent years, there is a trend that it should be evaluated not by a rough index such as "how many people passed in front of the advertisement" but by an index of "how many people actually viewed the advertisement", and methods associated with this are being studied.

[0003] Patent Document 1 describes a recognition evaluation system for evaluating the recognition of outdoor advertisements. In Patent Document 1, a simulation is executed in which a station and a road are virtually represented on a computer, a person is virtually represented on a computer in a virtual space that discretely represents the passage of time, and various actions are performed within the virtual space. Then, recognition is evaluated based on whether or not a person recognizes an advertisement during the simulation.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In Patent Document 1, when an advertisement is included in the field of view during an action, it is evaluated whether or not the advertisement is recognized based on the visual recognition probability for each action. If this visual recognition probability is not appropriately set, it is impossible to appropriately evaluate whether or not the advertisement is recognized, and it is impossible to appropriately evaluate the effect of the advertising medium. An object of the present disclosure is to enable appropriate evaluation of the effect of an advertising medium.

Means for Solving the Problems

[0006] The advertising media evaluation device related to this disclosure is A visibility probability estimation unit estimates the visibility probability for each of several paths that pass through an area where the target advertising medium is visible, A viewer count calculation unit calculates the number of people who viewed the estimated target advertising medium on each of the multiple routes by multiplying the traffic volume on the target route by the visibility probability estimation unit for the target route, and then calculates the number of people who viewed the estimated target advertising medium from the number of people calculated for each of the multiple routes. It is equipped with. [Effects of the Invention]

[0007] This disclosure estimates the visibility probability separately for each of multiple paths that pass through the visible area of ​​the advertising medium. This makes it possible to estimate a realistic and appropriate visibility probability. As a result, it becomes possible to appropriately evaluate the effectiveness of the advertising medium. [Brief explanation of the drawing]

[0008] [Figure 1] Configuration diagram of the advertising media evaluation device 10 according to Embodiment 1. [Figure 2] An explanatory diagram of the survey information storage unit 31 according to Embodiment 1. [Figure 3] An explanatory diagram of the investigation position storage unit 32 according to Embodiment 1. [Figure 4] An explanatory diagram of the model storage unit 33 according to Embodiment 1. [Figure 5] An explanatory diagram of the visible area storage unit 34 according to Embodiment 1. [Figure 6] An explanatory diagram of the road information storage unit 35 according to Embodiment 1. [Figure 7] An explanatory diagram of the visible path storage unit 36 ​​according to Embodiment 1. [Figure 8] An explanatory diagram of the traffic volume storage unit 37 according to Embodiment 1. [Figure 9]Explanatory drawing of the advertisement medium information storage unit 38 according to Embodiment 1. [Figure 10] Explanatory drawing of the environment information storage unit 39 according to Embodiment 1. [Figure 11] Explanatory drawing of the in-path traffic volume storage unit 40 according to Embodiment 1. [Figure 12] Explanatory drawing of the visual recognition probability storage unit 41 according to Embodiment 1. [Figure 13] Explanatory drawing of the number of viewers storage unit 42 according to Embodiment 1. [Figure 14] Explanatory drawing of the overall processing of the advertisement medium evaluation device 10 according to Embodiment 1. [Figure 15] Explanatory drawing of the reliability determination method by Method 2 according to Embodiment 1. [Figure 16] Explanatory drawing of the display example according to Modification 1. [Figure 17] Configuration diagram of the advertisement medium evaluation device 10 according to Embodiment 2. [Figure 18] Explanatory drawing of the reference attribute storage unit 43 according to Embodiment 2. [Figure 19] Explanatory drawing of the overall processing of the advertisement medium evaluation device 10 according to Embodiment 2. [Figure 20] Explanatory drawing of the adjustment process according to Embodiment 2. [Figure 21] Explanatory drawing of the adjustment process according to Embodiment 2. [Figure 22] Explanatory drawing of the method for generating the estimation model according to Embodiment 2. [Figure 23] Explanatory drawing of the adjustment using the estimation model according to Embodiment 2. [Figure 24] Explanatory drawing of the reference attribute storage unit 43 according to Modification 3. [Figure 25] Explanatory drawing of the adjustment process according to Modification 3. [Figure 26] Configuration diagram of the advertisement medium evaluation device 10 according to Embodiment 3. [Figure 27] Explanatory drawing of the visual recognition probability storage unit 41 according to Embodiment 3. [Figure 28] Explanatory drawing of the overall processing of the advertisement medium evaluation device 10 according to Embodiment 3. [Figure 29]Flowchart of the model construction process according to Embodiment 3. [Figure 30] An explanatory diagram of the location identification process according to Embodiment 3. [Figure 31] An explanatory diagram of the location identification process according to Embodiment 3. [Figure 32] Diagram illustrating the learning process according to Embodiment 3. [Figure 33] An explanatory diagram of the visibility probability estimation process according to Embodiment 3. [Modes for carrying out the invention]

[0009] Embodiment 1. ***Explanation of the structure*** Referring to Figure 1, the configuration of the advertising media evaluation device 10 according to Embodiment 1 will be described. The advertising media evaluation device 10 is a computer. The advertising media evaluation device 10 comprises hardware including a processor 11, memory 12, storage 13, and a communication interface 14. The processor 11 is connected to the other hardware via signal lines and controls this other hardware.

[0010] Processor 11 is an IC that performs processing. IC stands for Integrated Circuit. Specific examples of processor 11 include CPU, DSP, and GPU. CPU stands for Central Processing Unit. DSP stands for Digital Signal Processor. GPU stands for Graphics Processing Unit.

[0011] Memory 12 is a storage device that temporarily stores data. Specific examples of memory 12 include SRAM and DRAM. SRAM stands for Static Random Access Memory. DRAM stands for Dynamic Random Access Memory.

[0012] Storage 13 is a storage device for storing data. A concrete example of storage 13 is an HDD. HDD stands for Hard Disk Drive. Alternatively, storage 13 may be a portable recording medium such as an SD® memory card, CompactFlash®, NAND flash, flexible disk, optical disk, compact disk, Blu-ray® disc, or DVD. SD stands for Secure Digital. DVD stands for Digital Versatile Disk.

[0013] Communication interface 14 is an interface for communicating with external devices. Specific examples of communication interface 14 include Ethernet®, USB, and HDMI® ports. USB stands for Universal Serial Bus. HDMI stands for High-Definition Multimedia Interface.

[0014] The advertising media evaluation device 10 comprises a control unit 20 and a storage unit 30 as functional components. The functions of the control unit 20 are implemented by software. The functions of the storage unit 30 are implemented by the storage 13.

[0015] The control unit 20 comprises functional components: a learning unit 21 and a calculation unit 22. The learning unit 21 comprises functional components: a reliability determination unit 211, an exclusion data specification unit 212, and a model construction unit 213. The calculation unit 22 comprises functional components: a visible route determination unit 221, a traffic volume calculation unit 222, a visibility probability estimation unit 223, and a visibility count calculation unit 224. The storage 13 stores programs that implement the functions of each functional component of the control unit 20. These programs are loaded into the memory 12 by the processor 11 and executed by the processor 11. This enables the implementation of the functions of each functional component of the control unit 20.

[0016] The memory unit 30 includes storage units for survey information 31, survey location 32, model 33, visible area 34, road information 35, visible route 36, traffic volume 37, advertising media information 38, environmental information 39, route traffic volume 40, visibility probability 41, and number of viewers 42.

[0017] In Figure 1, only one processor 11 was shown. However, there may be multiple processors 11, and multiple processors 11 may work together to execute programs that implement each function.

[0018] ***Explanation of operation*** Referring to Figures 2 to 15, the operation of the advertising media evaluation device 10 according to Embodiment 1 will be explained. The operating procedure of the advertising media evaluation device 10 according to Embodiment 1 corresponds to the advertising media evaluation method according to Embodiment 1. Furthermore, the program that implements the operation of the advertising media evaluation device 10 according to Embodiment 1 corresponds to the advertising media evaluation program according to Embodiment 1.

[0019] Referring to Figure 2, the survey information storage unit 31 according to Embodiment 1 will be described. The survey information storage unit 31 stores multiple survey results data that investigate whether or not the user viewed the educational advertising media. Specifically, the survey information storage unit 31 stores the user ID, advertising media ID, and survey responses as survey result data. The user ID is the identification information of the user who is the subject of the survey. The advertising media ID is the identification information of the advertising media that is the subject of the survey. The survey responses are the answers to the questions in the survey. In this case, the survey is a questionnaire conducted with multiple individuals who were in an area where educational advertising materials were visible as subjects. The survey responses in the survey results data are the answers to the questionnaire obtained from the subjects. For example, the survey response is a response of "yes" or "no" to the question of whether they saw the advertising material indicated by the advertising material ID.

[0020] Referring to Figure 3, the investigation position storage unit 32 according to Embodiment 1 will be described. The survey location storage unit 32 stores the location information of the user who is the subject of the survey described above. Specifically, the survey location storage unit 32 stores the user ID, advertising media ID, location information list, and movement speed. The user ID and advertising media ID are the same as those stored in the survey information storage unit 31. The location information list is information indicating the location of the user indicated by the user ID at each point in time, and includes latitude and longitude for each date and time. The movement speed is the movement speed of the user indicated by the user ID in an area where the advertising media indicated by the advertising media ID is visible.

[0021] The model storage unit 33 according to Embodiment 1 will be described with reference to Figure 4. The model storage unit 33 stores an estimation model for estimating the probability of visibility. The estimation model will be described later.

[0022] Referring to Figure 5, the visible area storage unit 34 according to Embodiment 1 will be described. The visible area storage unit 34 stores information indicating the area where the advertising medium can be viewed. Specifically, the visible area storage unit 34 stores a list of visible area node coordinates for each advertising media ID. The advertising media ID is the identification information of the target advertising media. The visible area node coordinate list is a list of coordinates for identifying the area in which the advertising media indicated by the advertising media ID is visible. For example, the visibility area node coordinate list may have three coordinates set, and the triangular area indicated by these three coordinates will be the visible area. Alternatively, the visibility area node coordinate list may have multiple coordinates set, and the area obtained by connecting each coordinate in sequence will be the visible area. Furthermore, if there are areas that cannot be seen due to obstacles, etc., the visibility area node coordinate list may include information indicating the unseen areas. In this case, for example, the visible area will be the triangular area indicating the visible area, excluding the unseen areas.

[0023] Referring to Figure 6, the road information storage unit 35 according to Embodiment 1 will be described. The road information storage unit 35 stores information indicating roads in an area that includes the area where each advertising medium is visible. Specifically, the road information storage unit 35 stores a list of node coordinates for each road ID. The road ID is road identification information. The node coordinate list is a list of coordinates for identifying a road. The line or area obtained by connecting each coordinate in order represents the area of ​​the road.

[0024] Referring to Figure 7, the visible path storage unit 36 ​​according to Embodiment 1 will be described. The visible route storage unit 36 ​​stores information indicating the roads included in the visible area for each advertising medium. Specifically, the visible route storage unit 36 ​​stores a road ID list for each advertising media ID. The advertising media ID is the identification information of the target advertising media. The road ID list is a list of road IDs of roads that are included in the area where the advertising media indicated by the advertising media ID is visible.

[0025] Referring to Figure 8, the traffic volume storage unit 37 according to Embodiment 1 will be described. The traffic volume storage unit 37 stores information indicating the traffic volume for each road. Specifically, the traffic volume storage unit 37 stores the number of people for each road ID and inbound / outbound classification. The road ID is road identification information. The inbound / outbound classification is classification information indicating whether the direction is inbound or outbound. The number of people is the number of people passing through per unit of time.

[0026] Referring to Figure 9, the advertising media information storage unit 38 according to Embodiment 1 will be described. The advertising media information storage unit 38 stores attribute information of the advertising media. Specifically, the advertising media information storage unit 38 stores attribute information such as media size, height, and coordinates for each advertising media ID. The advertising media ID is the identification information of the target advertising media. The media size is the length and width of the advertising media indicated by the advertising media ID. The height is the height at which the advertising media indicated by the advertising media ID is installed. The coordinates are the latitude and longitude of the installation location of the advertising media indicated by the advertising media ID. Furthermore, attribute information for the advertising medium may include whether it is digitized or not, distance, presence or absence of lighting, presence or absence of motion, location, road, orientation, clarity, number of clusters, maximum viewing distance, obstacles, and video / still image classification. Whether it is digitized or not refers to whether the advertising medium is digital or analog. Distance is the distance from the main road to the advertising medium. Presence or absence of lighting refers to whether or not there is lighting illuminating the advertising medium. Presence or absence of motion refers to whether or not the advertising medium performs scrolling actions, etc. Location is the geographical location of the advertising medium. Road refers to information indicating the surrounding road environment. Orientation is the angle of the advertising medium relative to the main road. Clarity refers to the visual clarity of the advertising medium. Number of clusters is the number of advertising mediums with the same content that are displayed. Maximum viewing distance is the distance at which the advertising medium can be viewed. Obstacles refer to the presence or absence of obstacles that obstruct the view of the advertising medium. Video / still image classification refers to whether the advertising medium is video or still image.

[0027] Referring to Figure 10, the environmental information storage unit 39 according to Embodiment 1 will be described. The environmental information storage unit 39 stores information indicating the installation environment of the advertising medium. Specifically, the environmental information storage unit 39 stores the surrounding clutter level and situation for each advertising medium ID. The advertising medium ID is the identification information of the target advertising medium. The surrounding clutter level is information indicating the degree to which advertisements are densely placed in the location where the advertising medium indicated by the advertising medium ID is installed. The situation is information indicating the installation situation of the advertising medium indicated by the advertising medium ID. Specific examples of installation situations include whether it is installed in a building, in a retail store, or on public transportation such as buses, taxis, and trains.

[0028] Referring to Figure 11, the in-route traffic volume storage unit 40 according to Embodiment 1 will be described. The route traffic volume storage unit 40 stores the traffic volume of roads included in the visible area for each advertising medium. Specifically, the route traffic volume storage unit 40 stores a list of people for each advertising media ID. The advertising media ID is the identification information for the target advertising media. The list of people stores the number of people for each road ID and inbound / outbound classification. The road ID, inbound / outbound classification, and number of people are the same as the road ID, inbound / outbound classification, and number of people stored in the traffic volume storage unit 37.

[0029] Referring to Figure 12, the visibility probability storage unit 41 according to Embodiment 1 will be described. The visibility probability storage unit 41 stores the visibility probability for each advertising medium. Specifically, the visibility probability storage unit 41 stores the visibility probability for each advertising medium ID. The advertising medium ID is the identification information of the target advertising medium. The visibility probability is the probability that a person who is in the visible area of ​​the advertising medium indicated by the advertising medium ID will see the advertising medium indicated by the advertising medium ID.

[0030] Referring to Figure 13, the viewer count storage unit 42 according to Embodiment 1 will be described. The viewer count storage unit 42 stores the number of viewers for each advertising medium. Specifically, the viewer count storage unit 42 stores the viewer count for each advertising medium ID. The advertising medium ID is the identification information of the target advertising medium. The viewer count is the number of people who viewed the advertising medium indicated by the advertising medium ID.

[0031] Referring to Figure 14, the overall processing of the advertising media evaluation device 10 according to Embodiment 1 will be explained. (Step S11: Reliability determination process) The reliability determination unit 211 determines the reliability of each of the multiple survey result data stored in the survey information storage unit 31, which investigate whether or not the learning advertising medium was viewed. In this case, the reliability determination unit 211 may also determine the reliability by referring to the location information of the survey subject, the user, stored in the survey location storage unit 32.

[0032] (Step S12: Exclude data specification process) The exclusion data designation unit 212 designates exclusion data, which are survey result data from among multiple survey result data that will not be used to estimate the visibility probability for advertising media, based on the reliability determination result in step S11. Specifically, the exclusion data designation unit 212 designates as exclusion data survey result data that was determined in step S11 to have a reliability lower than the standard.

[0033] (Step S13: Model building process) The model building unit 213 uses the remaining survey data, which were not designated as excluded data in step S12, as training data to construct an estimation model that estimates the visibility probability of the target advertising medium. In other words, the model building unit 213 does not include survey data that is judged to have a reliability lower than the standard in the training data. At this time, the model building unit 213 also uses the attribute information of the training advertising medium stored in the advertising medium information storage unit 38 and the information indicating the installation environment of the training advertising medium stored in the environment information storage unit 39 as training data. The generated estimation model is a model that estimates the visibility probability for the target advertising medium by taking the attribute information of the target advertising medium and the information indicating the installation environment of the target advertising medium as input. Existing technologies can be used for the method of constructing the estimation model. The model building unit 213 stores the constructed estimation model in the model storage unit 33.

[0034] (Step S14: Visible path determination process) The visible route determination unit 221 refers to the information stored in the visible area storage unit 34 and the information stored in the road information storage unit 35 to identify the roads included in the visible area for the advertising medium to be estimated. Specifically, the visible route determination unit 221 reads information indicating the visible area for the advertising medium to be estimated from the visible area storage unit 34. The visible route determination unit 221 reads the road IDs of the roads included in the read visible area from the road information storage unit 35. Then, the visible route determination unit 221 stores the read road IDs as a road ID list, along with the advertising medium IDs of the advertising medium to be estimated and the road ID list, in the visible route storage unit 36. Here, the visible route determination unit 221 simply reads the road IDs of roads included in the visible area. However, the visible route determination unit 221 may also consider the location of the advertising media stored in the advertising media information storage unit 38 and read only the road IDs of roads included in the visible area whose advertising media are within the field of view. Furthermore, if the visible route determination unit 221 only has advertising media within the field of view in either the uphill or downhill direction, it may also read the uphill / downhill classification of roads whose advertising media are within the field of view and set it in the road ID list.

[0035] (Step S15: Traffic volume calculation process) The traffic volume calculation unit 222 identifies the traffic volume for each road included in the visible area of ​​the advertising medium being estimated. Specifically, the traffic volume calculation unit 222 reads the road ID list for the advertising media to be estimated, which is stored in the visible route storage unit 36 ​​in step S14. For each road ID included in the read road ID list, the traffic volume calculation unit 222 reads the number of people for each inbound / outbound classification from the traffic volume storage unit 37. The traffic volume calculation unit 222 stores the read number of people along with the advertising media ID for the advertising media to be estimated in the route traffic volume storage unit 40. More precisely, the traffic volume calculation unit 222 sets the number of people for each road ID set in the road ID list in the number of people for the advertising media ID for the advertising media to be estimated, based on the number of people for that road ID. Furthermore, if the road ID in step S14 has an uphill / downhill classification in the road ID list where the advertising medium is within the field of view, the traffic volume calculation unit 222 reads only the number of people for the uphill / downhill classification set in the road ID list. In this case, for that road ID, only the number of people for the uphill / downhill classification set in the road ID list is set in the route traffic volume storage unit 40.

[0036] (Step S16: Visual Probability Estimation Process) The visibility probability estimation unit 223 estimates the visibility probability for the target advertising medium based on the remaining data extracted in step S11. In this process, the visibility probability estimation unit 223 estimates the visibility probability for the target advertising medium without using survey result data that has been determined to have a reliability lower than the standard. Here, if the advertising medium to be estimated is a different advertising medium from the advertising medium used for learning, that is, if the advertising medium to be estimated is not an advertising medium that has been surveyed, the visibility probability estimation unit 223 estimates the visibility probability for the advertising medium to be estimated using the estimation model constructed in step S13 based on the remaining data. In other words, the visibility probability estimation unit 223 inputs the attribute information about the advertising medium to be estimated stored in the advertising medium information storage unit 38 and the information indicating the installation environment of the advertising medium to be estimated stored in the environment information storage unit 39 into the estimation model and obtains the visibility probability estimated by the estimation model. On the other hand, if the advertising medium to be estimated is a training advertising medium, that is, if the advertising medium to be estimated is an advertising medium that has been surveyed, the visibility probability estimation unit 223 may estimate the visibility probability for the advertising medium to be estimated using an estimation model, or it may estimate the visibility probability for the advertising medium to be estimated using residual data. In other words, in this case, the visibility probability estimation unit 223 may estimate the visibility probability for the advertising medium to be estimated using residual data without using an estimation model. For example, the visibility probability estimation unit 223 may use the percentage of respondents who answered "yes" to the question "Did you see the advertising medium?" in the survey responses in the residual data as the visibility probability. Also, if the survey responses include a degree of confidence in having seen the advertising medium, the visibility probability estimation unit 223 may use the degree of confidence in the percentage of respondents who answered "yes" to the question "Did you see the advertising medium?" in the survey responses as the visibility probability. The visibility probability estimation unit 223 stores the visibility probability along with the advertising medium ID for the advertising medium to be estimated in the visibility probability storage unit 41.

[0037] (Step S17: Calculation process for the number of viewers) The viewer count calculation unit 224 calculates the number of people who viewed the advertising medium by multiplying the traffic volume in the area where the advertising medium is visible by the visibility probability estimated in step S16. Specifically, the viewer count calculation unit 224 calculates the number of people who viewed the advertising medium by multiplying the total number of people set in the list of people for the advertising medium to be estimated, which was set in the route traffic volume storage unit 40 in step S15, by the viewability probability stored in the viewability probability storage unit 41. The total number of people set in the list of people is the total number of people for each road ID and each uphill / downhill classification set in the list of people. The viewer count calculation unit 224 stores the calculated number of viewers, which is the number of people who viewed the advertising medium, along with the advertising medium ID for the advertising medium to be estimated, in the viewer count storage unit 42.

[0038] The reliability determination process according to Embodiment 1 (step S11 in Figure 14) will be described below. The reliability determination unit 211 sets each of the multiple survey result data as the target survey result data and estimates the reliability of the target survey result data using at least one of the following methods 1 and 2. As described above, the survey is a questionnaire survey conducted on people who were in an area where the educational advertising media was visible. The questionnaire survey is conducted by distributing questionnaires to people who were in an area where the educational advertising media was visible.

[0039] (Method 1) The reliability determination unit 211 determines reliability from the responses to the questionnaire survey. There are five possible methods for determining reliability from responses to a questionnaire survey, from method A to method E. The reliability determination unit 211 may calculate reliability using one of methods A to E, or it may calculate reliability using two or more methods. When using two or more methods, the reliability determination unit 211 sets weights for each method so that the sum of the weights for each of the two or more methods used is 1, and then calculates the final reliability by weighted average.

[0040] (Method A) The reliability determination unit 211 determines the reliability of the survey results data based on the bias in the choices included in the responses. Determining reliability based on the bias in the answer choices means that the more the answers to multiple questions are biased towards a single choice, the lower the reliability is judged. For example, if each question has choices from 1 to 4, and only choice 4 is selected, the reliability will be judged as low. This is because if the same choice is selected repeatedly, it may indicate that the survey participants are answering carelessly. For example, the reliability determination unit 211 calculates the reliability (Score_Linear[%]) using method A with the formula: "Reliability (Score_Linear[%]) = Standard deviation of answer choices / Standard deviation when answers are given randomly". When calculated using this formula, if all answers are from a single choice, the standard deviation of the answer choices becomes 0, and the reliability becomes 0%. On the other hand, if the answers are scattered, the standard deviation of the answer choices becomes larger and approaches the standard deviation when answers are given randomly, so the reliability value becomes larger.

[0041] Alternatively, reliability could be reduced only when the bias exceeds the bias threshold, and not when the bias does not exceed the bias threshold. This is because some degree of bias is possible even when respondents answer honestly.

[0042] (Method B) The reliability determination unit 211 determines the reliability of the survey results data based on the time taken to respond. Determining reliability based on the time taken to answer means that the shorter the total time taken to answer multiple questions, the lower the reliability. This is because if respondents answer quickly, there is a possibility that they are answering carelessly. For example, the reliability determination unit 211 determines the reliability (Score_Time[%]) according to method B based on the range of the determination time, which is calculated as "determination time = reference time - time taken to respond". Specifically, if it is less than 0 seconds, the reliability (Score_Time[%]) is set to 100%. If it is 0 seconds or more but less than 1 minute, the reliability (Score_Time[%]) is set to 75%. If it is 1 minute or more but less than 2 minutes, the reliability (Score_Time[%]) is set to 50%. If it is 2 minutes or more but less than 3 minutes, the reliability (Score_Time[%]) is set to 25%. If it is 3 minutes or more, the reliability (Score_Time[%]) is set to 0%.

[0043] (Method C) The reliability determination unit 211 determines the reliability of the survey results data based on the response to the question of whether the respondent saw an advertisement that was not displayed. The question of whether the respondent saw an advertisement that was not displayed is a determination question used to determine reliability. Determining reliability based on responses to the question of whether an ad was seen that was not displayed means that the reliability is judged as low if the respondent answers that they saw an ad that was not displayed. This is because if a respondent answers that they saw an ad that was not displayed, their memory may be vague. For example, suppose the question asks whether or not the respondent saw an advertisement that was not displayed, and if so, also to indicate their level of confidence. In this case, the reliability determination unit 211 calculates the reliability (Score_memory[%]) using method C, with the formula: Reliability (Score_memory[%]) = 1 - (average level of confidence when the respondent answered "yes"). As a specific example, if there are four questions asking whether the respondent saw an advertisement that was not displayed, and the respondent answered "yes" to three of them, with confidence levels of 75%, 50%, and 25% respectively, then the reliability (Score_memory[%]) = 1 - the average of [75%, 50%, 25%] = 1 - 50% = 50%.

[0044] (Method D) The reliability determination unit 211 determines the reliability of the survey results data based on the response to the question, "Will you answer this question seriously?". The question, "Will you answer this question seriously?", is a determination question used to determine reliability. Determining reliability based on responses to the question of whether the answer will be taken seriously means that reliability will be lowered if the answer is "not taken seriously." This is because if the answer is "not taken seriously," there is a possibility that the answer was not taken seriously. For example, the reliability determination unit 211 sets the reliability (Score_serious) according to method D to 0% if the answer to the question "Will you answer seriously?" is "Will you answer seriously?", and sets the reliability (Score_serious) to 100% if the answer is "Will you answer seriously?".

[0045] (Method E) The reliability determination unit 211 determines the reliability of the survey results data based on the answers to questions that ask about impossible situations. The questions that ask about impossible situations are determination questions, which are questions used to determine reliability. Determining reliability based on responses to questions about impossible situations means that reliability is judged as low if the response indicates that an impossible situation is possible. This is because if the response indicates that an impossible situation is possible, it may indicate that the survey participant is giving an inaccurate answer. For example, suppose a question asks about an unlikely situation and the respondent is asked to indicate how likely they are. In this case, the reliability determination unit 211 lowers the reliability (Score_mockdrill[%]) according to method E the more strongly the respondent believes the situation is likely. Specifically, if the respondent does not believe the situation is likely, the reliability (Score_mockdrill[%]) will be 100%. If they don't think it's likely very much, the reliability (Score_mockdrill[%]) will be 75%. If neither agrees nor disagrees, the reliability (Score_mockdrill[%]) will be 50%. If they think it's somewhat likely, the reliability (Score_mockdrill[%]) will be 25%. If they think it's likely, the reliability (Score_mockdrill[%]) will be 0%.

[0046] (Method 2) The reliability determination unit 211 determines the reliability of the survey results data based on the relationship between the area where the learning advertisement medium is visible and the location of the survey participants. The location of the survey participants is the location indicated by the location information of the user who is a survey participant, which is stored in the survey location storage unit 32. The survey is conducted with participants who are in an area where educational advertising materials are visible. However, there is an error in the location information of the participants, and some people who are not actually in an area where educational advertising materials are visible may be selected as participants and sent the questionnaire. Therefore, the reliability is determined in a way that takes into account the error in the location information of the participants, and the reliability decreases as the likelihood of a participant not actually being in an area where educational advertising materials are visible increases.

[0047] Please refer to Figure 15 for a more detailed explanation. The reliability determination unit 211 calculates reliability based on the degree of overlap between the error range of the survey subject's location and the area where the learning advertisement medium is visible. The error range of the survey subject's location is, for example, the area enclosed by a circle centered on the location indicated by the location information and specified by a radius determined by the accuracy of the location information. For example, the reliability determination unit 211 calculates the reliability (Score_location[%]) using method 2, based on the formula: "Reliability (Score_location[%]) = overlapping area of ​​the error range and the visible area / area of ​​the error range". Therefore, as shown in Figure 15(A), the reliability (Score_location[%]) is high when most of the error range overlaps with the visible area. On the other hand, as shown in Figure 15(B), the reliability (Score_location[%]) is low when most of the error range does not overlap with the visible area.

[0048] In this scenario, the survey subjects are moving, and the degree of overlap between the error range of the survey subjects' positions and the area where the learning advertisements are visible varies depending on the survey subjects' positions. Therefore, the reliability determination unit 211 calculates reliability using the maximum value of the degree of overlap between the error range of the survey subjects' positions and the area where the learning advertisements are visible. Alternatively, the reliability determination unit 211 may calculate reliability using the average value of the degree of overlap while the survey subjects are within the area where the learning advertisements are visible.

[0049] When the reliability determination unit 211 determines reliability using both method 1 and method 2, it sets weights for each method such that the sum of the weights for the reliability calculated by method 1 and the reliability calculated by method 2 equals 1, and then calculates the final reliability by weighted average.

[0050] ***Effects of Embodiment 1*** As described above, the advertising media evaluation device 10 according to Embodiment 1 uses multiple survey results data to investigate whether or not the advertising media was viewed, and further estimates the viewing probability using the remaining data after removing the less reliable survey results data. This makes it possible to estimate an appropriate viewing probability that is in line with reality. As a result, it becomes possible to appropriately evaluate the effectiveness of the advertising media.

[0051] ***Other configurations*** <Example 1> The visibility probability estimation unit 223 may estimate the visibility probability using the remaining data, and may also estimate the visibility probability using pre-exclusion data, which is the survey result data before the exclusion of survey result data that has been determined to have a reliability lower than the standard. The advertising media evaluation device 10 is equipped with a display unit, which may display a comparison between the visibility probability estimated using the remaining data and the visibility probability estimated using the pre-exclusion data.

[0052] An example of a display using the display unit will be explained with reference to Figure 16. Figure 16 shows the viewability probability estimated using residual data (After in Figure 16) and the viewability probability estimated using pre-exclusion data (Before in Figure 16) for each advertising medium. Furthermore, Figure 16 shows the difference between the viewability probability estimated using residual data and the viewability probability estimated using pre-exclusion data (Before-After in Figure 16) for each advertising medium. In addition, Figure 16 shows the average value of the viewability probability estimated using residual data, the average value of the viewability probability estimated using pre-exclusion data, and the average value of the difference. By checking the display on the display unit, it becomes possible to understand the effectiveness of excluding unreliable survey results data.

[0053] Embodiment 2. Embodiment 2 differs from Embodiment 1 in that it adjusts the distribution of the attributes of the survey participants in the survey results data. Embodiment 2 explains this difference, while omitting explanations of the same points.

[0054] ***Explanation of the structure*** Referring to Figure 17, the configuration of the advertising media evaluation device 10 according to Embodiment 2 will be described. The advertising media evaluation device 10 differs from the advertising media evaluation device 10 shown in Figure 1 in the configuration of the functional components of the learning unit 21 and the storage unit of the memory unit 30. The learning unit 21 includes an adjustment unit 214 instead of a reliability determination unit 211. The memory unit 30 further includes a reference attribute storage unit 43.

[0055] ***Explanation of operation*** Referring to Figures 18 to 22, the operation of the advertising media evaluation device 10 according to Embodiment 2 will be explained. The operating procedure of the advertising media evaluation device 10 according to Embodiment 2 corresponds to the advertising media evaluation method according to Embodiment 2. Furthermore, the program that implements the operation of the advertising media evaluation device 10 according to Embodiment 2 corresponds to the advertising media evaluation program according to Embodiment 2.

[0056] Referring to Figure 18, the reference attribute storage unit 43 according to Embodiment 2 will be described. The reference attribute storage unit 43 stores a reference distribution that shows the distribution of reference attributes. Specifically, the reference attribute storage unit 43 stores a reference distribution for each advertising medium ID. The advertising medium ID is the identification information of the advertising medium for learning. The reference distribution shows the distribution of human attributes in the visible area of ​​the advertising medium for learning. In Embodiment 2, the reference distribution shows the attributes of each person in the visible area of ​​the advertising medium for learning. The attributes are of a type that contributes to the visibility probability. Here, it is assumed that the reference distribution has distributions of multiple types of attributes set. Possible attributes include gender, age group, household income, personal income, highest level of education, marital status, family structure, presence of children, occupation, prefecture of residence, residential area, home ownership status, and hobbies. Age group refers to age categories such as 10s, 20s, 30s, etc. Age groups can also be divided into 5-year increments, 20-year increments, etc. Household income refers to the annual income of the household. Personal income refers to the annual income of the individual. Highest level of education refers to whether the highest level of education attained was junior high school, high school, junior college, four-year university, graduate school, etc. Marital status refers to whether the person is married (no children), married (with children), cohabiting, single, divorced or separated, widowed, etc. Family structure refers to the number of people in the household. Presence of children refers to whether there are none, whether there are children, and if there are children, the age range of the children, etc. Occupation includes company employee (regular employee), company employee (contract employee), civil servant, self-employed or freelancer, company executive or manager, part-time or temporary worker, student, full-time housewife or househusband, unemployed, etc. Prefecture of residence is the prefecture where the person resides. Region of residence is the region where the person resides, such as Kanto or Tokai. Home ownership status includes owned detached house, owned apartment, rented, company housing, shared accommodation, etc. Hobbies include reading, cooking, etc. Furthermore, attributes may include the duration of stay in an area where the advertising medium is visible, the mode of transportation, the speed of transportation, and the trajectory of the movement. Modes of transportation include walking, cycling, and driving.

[0057] Referring to Figure 19, the overall processing of the advertising media evaluation device 10 according to Embodiment 2 will be described. The processes in steps S23, S24, and S26 are the same as the processes in steps S14, S15, and S17 in Figure 14.

[0058] (Step S21: Adjustment process) The adjustment unit 214 adjusts the survey results data stored in the survey information storage unit 31, which shows whether or not the survey subjects viewed the educational advertising media, so that the distribution of the survey subjects' attributes approaches the standard distribution stored in the standard attribute storage unit 43, and generates adjusted data. Here, the survey results data stored in the survey information storage unit 31 is data obtained from an online questionnaire survey conducted with multiple people who were in an area where the learning advertisement medium was visible as the survey subjects. For example, the survey results data is data obtained from a questionnaire survey conducted by sending questionnaires via the network to devices such as smartphones owned by people who entered an area where the learning advertisement medium was visible. The survey subjects are people who were the target of the questionnaire survey and answered the questionnaire. In other words, each set of survey results data is composed of the answers obtained from each survey subject.

[0059] (Step S22: Model building process) The model building unit 213 uses the adjusted data generated in step S21 as training data to build an estimation model that estimates the visibility probability of the target advertising medium. In this process, the model building unit 213 also uses the attribute information of the training advertising medium stored in the advertising medium information storage unit 38 and the information indicating the installation environment of the training advertising medium stored in the environment information storage unit 39 as training data. The generated estimation model is a model that estimates the visibility probability for the target advertising medium by taking the attribute information of the target advertising medium and the information indicating the installation environment of the target advertising medium as input. Existing technologies can be used for the method of constructing the estimation model. The model building unit 213 stores the constructed estimation model in the model storage unit 33.

[0060] (Step S25: Visual Probability Estimation Process) The visibility probability estimation unit 223 estimates the visibility probability for the target advertising medium based on the adjusted data generated in step S21. Here, if the advertising medium to be estimated is a different advertising medium from the advertising medium used for learning, that is, if the advertising medium to be estimated is not an advertising medium that has been surveyed, the visibility probability estimation unit 223 estimates the visibility probability for the advertising medium to be estimated using the estimation model constructed in step S22 based on the remaining data. In other words, the visibility probability estimation unit 223 inputs the attribute information about the advertising medium to be estimated stored in the advertising medium information storage unit 38 and the information indicating the installation environment of the advertising medium to be estimated stored in the environment information storage unit 39 into the estimation model and obtains the visibility probability estimated by the estimation model. On the other hand, if the advertising medium to be estimated is a training advertising medium, that is, if the advertising medium to be estimated is an advertising medium that has been surveyed, the visibility probability estimation unit 223 may estimate the visibility probability for the advertising medium to be estimated using an estimation model, or it may estimate the visibility probability for the advertising medium to be estimated using adjusted data. In other words, in this case, the visibility probability estimation unit 223 may estimate the visibility probability for the advertising medium to be estimated using adjusted data without using an estimation model. The visibility probability estimation unit 223 stores the visibility probability along with the advertising medium ID for the advertising medium to be estimated in the visibility probability storage unit 41.

[0061] Referring to Figures 20 to 23, the adjustment process according to Embodiment 2 (step S21 in Figure 19) will be described. Refer to Figure 20 to explain the difference between an ideal questionnaire survey and an actual questionnaire survey. As shown in Figure 20(A), suppose that of the 15 people in the area where the advertisement could be seen, 12 were male and 3 were female. In other words, the ratio of men to women among those in the area where the advertisement could be seen was 4:1. In this case, suppose that 5 responses were obtained from an online questionnaire survey. In this case, as shown in Figure 20(B), suppose that of the 5 responses, 4 were from men and 1 was from a woman. In other words, the ratio of men to women among the responses was also 4:1. Thus, an ideal questionnaire survey is one in which the distribution of attributes of the people who actually participate in the survey matches the distribution of attributes of the people who provide the responses. In contrast, in actual surveys, as shown in Figure 20(C), it is possible that out of 5 responses, 2 are from men and 3 are from women. Furthermore, online surveys tend to attract a large number of women in their 30s and 40s. When such responses are obtained, as shown in Figure 20(D), if 15 people are in an area where the advertisement is visible, it would be estimated that 6 are men and 9 are women. In other words, an incorrect estimation is made. Therefore, the adjustment unit 214 adjusts the survey results data so that the distribution of the attributes of the survey subjects approaches the reference distribution, thereby generating adjusted data. By using the adjusted data for estimation, accurate estimation becomes possible.

[0062] As shown in Figure 21, the adjustment unit 214 generates adjusted data by adjusting the distribution of the respondents' attributes so that it approaches the reference distribution. Here, the reference distribution has multiple types of attributes set, such as age and gender. Therefore, the adjustment unit 214 needs to adjust the distribution of multiple types of attributes of the respondents so that it approaches the reference distribution. The adjustment unit 214 uses propensity scores to make the multiple types of attributes one-dimensional, and generates adjusted data by adjusting the distribution of propensity scores of the respondents' attributes so that it approaches the distribution of propensity scores in the reference distribution.

[0063] Specifically, as shown in Figure 22, the adjustment unit 214 takes the survey results data and the reference distribution data as input and generates a model 51 that calculates the propensity score. Model 51 is composed of existing techniques such as linear regression and logistic regression. More precisely, the adjustment unit 214 uses multiple types of attributes in each survey result data (age, gender, occupation, annual income, etc. in Figure 22) as explanatory variables and a binary classification indicating whether or not it is a standard distribution (NO=0 in Figure 22) as the dependent variable. Similarly, the adjustment unit 214 uses multiple types of attributes for each person shown by the standard distribution (age, gender, occupation, annual income, etc. in Figure 22) as explanatory variables and a binary classification indicating whether or not it is a standard distribution (YES=1 in Figure 22) as the dependent variable. Using these explanatory and dependent variables as training data, the adjustment unit 214 generates a model 51 that calculates a propensity score using the explanatory variables as input. The propensity score here is close to 1 for sets of attributes that appear mostly in sets of attributes in the standard distribution, and close to 0 for sets of attributes that appear mostly in sets of attributes in the survey result data. Furthermore, the propensity score here is 0.5 if the attribute sets appear equally in the reference distribution and in the survey results data.

[0064] The adjustment unit 214 inputs multiple types of attributes (age, gender, occupation, annual income, etc. in Figure 23) from each survey result data to the model 51 and calculates a propensity score for each survey result data. The adjustment unit 214 then calculates a weight W for each survey result data from the propensity score for that survey result data. Specifically, the adjustment unit 214 calculates the weight W using the formula W = 1 / (0.5 / propensity score). For example, if the propensity score is 0.1, the weight W = 1 / (0.5 / 0.1) = 0.2. If the propensity score is 0.8, the weight W = 1 / (0.5 / 0.8) = 1.6. For example, if the propensity score is 0.5, the weight W = 1 / (0.5 / 0.5) = 1.

[0065] The adjustment unit 214 then weights the target survey results data with the calculated weight W. As a result, the adjustment unit 214 weights each survey result data such that the further the attributes of the target survey result data are from the attributes that constitute the standard distribution, the less the target survey result data is weighted, and the closer the attributes of the target survey result data are to the attributes that constitute the standard distribution, the more weight the target data is weighted. Weighting can be interpreted as increasing or decreasing the number of items according to the weight W. For example, if the weight W of the target survey results data is 0.2, the number of target survey results data will be reduced to 0.2 times. In the example in Figure 23, there are too many survey results data for women in their 40s, so they are reduced to 0.2 times. If the weight W of the target survey results data is 1.6, the number of target survey results data will be increased to 1.6 times. In the example in Figure 23, there are too few survey results data for men in their 70s, so they are increased to 1.6 times. If the weight W of the target survey results data is 1, the number of target survey results data will remain unchanged. In the example in Figure 23, the survey results data for teenage students is accurate, so it is maintained as is.

[0066] ***Effects of Embodiment 2*** As described above, the advertising media evaluation device 10 according to Embodiment 2 adjusts the survey results data so that the distribution of the attributes of the survey subjects approaches the standard distribution. The visibility probability estimation unit 223 then estimates the visibility probability using the adjusted data, which is the adjusted survey results data. This suppresses bias in the attributes of the data used for estimation, improves the quality of the data, and makes it possible to estimate an appropriate visibility probability that is in line with reality. As a result, it becomes possible to appropriately evaluate the effectiveness of advertising media.

[0067] ***Other configurations*** <Variation 3> In Embodiment 2, the baseline distribution represents the attributes of each person in an area where the learning advertisement is visible. However, as shown in Figure 24, the baseline distribution may also represent information showing the distribution of various attributes in an area where the learning advertisement is visible. In this case, it is possible to generate baseline distribution information using statistical information of various attributes in a range that includes the area where the learning advertisement is visible. In other words, it is possible to generate baseline distribution information without conducting on-site surveys in the area where the learning advertisement is visible. As shown in Figure 24, if the reference distribution is information showing the distribution of attributes for each type, then, as shown in Figure 25, the adjustment unit 214 calculates the attribute weight W1 for each attribute of each type. Specifically, the adjustment unit 214 calculates the attribute weight W1 for attribute X in type A as follows: W1 = proportion of attribute X in type A in the reference distribution / proportion of attribute X in type A in the survey results data. Then, for each survey results data, the adjustment unit 214 calculates the weight W for that survey results data by multiplying it by the attribute weight W1 for each attribute in the survey results data. For example, suppose a survey result data set is for people in their 40s and female. In this case, the weight W of the survey result data would be 2, which is the product of the attribute weight W1=4 for the attribute "40s" in the category "age group" and the attribute weight W1=0.5 for the attribute "female" in the category "gender group".

[0068] <Modification 4> In Embodiment 2, the learning unit 21 is configured to include an adjustment unit 214 instead of a reliability determination unit 211. However, the learning unit 21 may also be configured to include an adjustment unit 214 in addition to a reliability determination unit 211. In this case, the adjustment unit 214 adjusts the distribution of the attributes of the survey subjects in the remaining data extracted by the reliability determination unit 211 to generate adjusted data. Then, the visibility probability estimation unit 223 estimates the visibility probability using the adjusted data. Alternatively, the reliability determination unit 211 determines the reliability of the adjusted data generated by the adjustment unit 214, and excludes data whose reliability is determined to be lower than the standard to generate remaining data. Then, the visibility probability estimation unit 223 estimates the visibility probability using the remaining data.

[0069] Embodiment 3. Embodiment 3 differs from Embodiments 1 and 2 in that it estimates the visibility probability for each of the multiple paths that pass through the area where the advertising medium is visible. Embodiment 3 explains this difference, while omitting explanations of the same points. Embodiment 3 describes a case in which a modification has been made to Embodiment 1. However, it is also possible to modify Embodiment 2.

[0070] ***Explanation of the structure*** Referring to Figure 26, the configuration of the advertising media evaluation device 10 according to Embodiment 3 will be described. The advertising media evaluation device 10 differs from the advertising media evaluation device 10 shown in Figure 1 in the configuration of the functional components of the learning unit 21. The learning unit 21 does not include a reliability determination unit 211.

[0071] ***Explanation of operation*** Referring to Figures 27 to 33, the operation of the advertising media evaluation device 10 according to Embodiment 3 will be explained. The operating procedure of the advertising media evaluation device 10 according to Embodiment 3 corresponds to the advertising media evaluation method according to Embodiment 3. Furthermore, the program that implements the operation of the advertising media evaluation device 10 according to Embodiment 3 corresponds to the advertising media evaluation program according to Embodiment 3.

[0072] Referring to Figure 27, the visibility probability storage unit 41 according to Embodiment 3 will be described. The visibility probability storage unit 41 stores the visibility probability for each advertising medium and route. Specifically, the visibility probability storage unit 41 stores the visibility probability for each advertising medium ID, road ID, and uphill / downhill classification.

[0073] Referring to Figure 28, the overall processing of the advertising media evaluation device 10 according to Embodiment 3 will be described. The processes in steps S32 and S33 are the same as the processes in steps S14 and S15 in Figure 14. (Step S31: Model building process) The model building unit 213 uses multiple survey results data, each of which investigates whether or not a survey participant who passed through one of several routes that pass through an area where the learning advertising medium is visible viewed the learning advertising medium, to build an estimation model that estimates the probability of a person passing through an area where the target advertising medium is visible viewing the target advertising medium. Specifically, the model building unit 213 sets each of the multiple survey results data as the target survey results data. The model building unit 213 uses the visibility probability for the training advertising media indicated by the target survey results data, as well as information on the paths taken by the survey subjects regarding the target survey results data, as training data to build an estimation model. In addition, the model building unit 213 also uses the attribute information of the training advertising media stored in the advertising media information storage unit 38 and the information indicating the installation environment of the training advertising media stored in the environment information storage unit 39 as training data.

[0074] The visibility probability for learning advertising media, as indicated by the survey results data, is determined from the survey responses in the survey results data. For example, suppose the survey responses regarding advertising media include "saw," "thought I saw," "don't know," "don't think I saw," and "didn't see." In this case, if the response is "saw," the visibility probability is set to 100%. If the response is "thought I saw," the visibility probability is set to 75%. If the response is "don't know," the visibility probability is set to 50%. If the response is "don't think I saw," the visibility probability is set to 25%. If the response is "didn't see," the visibility probability is set to 0%.

[0075] Information about the route taken by the survey participants includes at least one of the following: the distance within the area where the learning-related advertising media was visible along the route taken by the survey participants, and the positional relationship between the learning-related advertising media and the route taken by the survey participants. Here, information about the route taken by the survey participants is assumed to include both distance and positional relationship.

[0076] The generated estimation model takes path information, attribute information of the target advertising medium, and information indicating the installation environment of the target advertising medium as input to estimate the visibility probability for the target advertising medium. Existing technologies can be used to construct the estimation model. The model building unit 213 stores the constructed estimation model in the model storage unit 33.

[0077] (Step S34: Visual Probability Estimation Process) The visibility probability estimation unit 223 estimates the visibility probability for the target advertising medium for each of several paths that pass through the area where the target advertising medium is visible. Here, if the advertising medium to be estimated is a different advertising medium from the advertising medium used for learning, that is, if the advertising medium to be estimated is not an advertising medium that has been surveyed, the visibility probability estimation unit 223 uses the estimation model constructed in step S32 to estimate the visibility probability for the advertising medium to be estimated. Specifically, the visibility probability estimation unit 223 sets each of the multiple paths that pass through the visible area of ​​the advertising medium to be estimated as a target path. Then, the visibility probability estimation unit 223 inputs the length of the target path, the positional relationship between the target path and the advertising medium to be estimated, the attribute information about the advertising medium to be estimated stored in the advertising medium information storage unit 38, and the information indicating the installation environment of the advertising medium to be estimated stored in the environment information storage unit 39 into the estimation model and obtains the visibility probability estimated by the estimation model. On the other hand, if the advertising medium to be estimated is a learning advertising medium, that is, if the advertising medium to be estimated is an advertising medium that has been surveyed, the visibility probability estimation unit 223 may use an estimation model to estimate the visibility probability for the advertising medium to be estimated, or it may use the visibility probabilities shown by multiple survey results to estimate the visibility probability for the advertising medium to be estimated. In other words, in this case, the visibility probability estimation unit 223 may estimate the visibility probability for the advertising medium to be estimated using the visibility probabilities shown by multiple survey results without using an estimation model. For example, the visibility probability estimation unit 223 may estimate the average value of the visibility probabilities shown by the survey results when a survey subject travels along a target route for each of the multiple routes, and use that as the visibility probability for the target route. The visibility probability estimation unit 223 stores the visibility probability for each of the multiple routes in the visibility probability storage unit 41, along with the advertising media ID for the advertising media to be estimated, the road ID and uphill / downhill classification indicating the route to be estimated, and the visibility probability for the route to be estimated.

[0078] (Step S35: Calculation process for the number of observers) The viewer count calculation unit 224 sets each of the multiple routes that pass through the area where the target advertising medium can be seen as a target. The viewer count calculation unit 224 calculates the number of people who saw the target advertising medium on the target route by multiplying the traffic volume for the target route by the visibility probability for the target route estimated in step S34. Here, the traffic volume for the target route is the number of people for the target route set in the route traffic volume storage unit 40. The target route is identified by the advertising medium ID, road ID, and inbound / outbound classification of the target advertising medium. The viewer count calculation unit 224 calculates the number of people who saw the target advertising medium by summing the number of people calculated for each of the multiple routes that pass through the area where the target advertising medium can be seen. The viewer count calculation unit 224 stores the calculated number of viewers, which is the number of people who viewed the advertisement, along with the advertising medium ID for the advertising medium being estimated, in the viewer count storage unit 42.

[0079] The model building process (step S31 in Figure 28) will be explained with reference to Figures 29 to 32. The model building unit 213 sets each of the multiple survey result data as the target survey result data. In Embodiment 3, the model building unit 213 distributes the visibility probability for the learning advertising medium indicated by the target survey result data to each of the multiple locations along the path taken by the survey subject for the target survey result data. Then, for each of the multiple locations along the path, the model building unit 213 constructs an estimation model using the visibility probability at the target location, as well as the positional relationship between the target location and the learning advertising medium.

[0080] Please refer to Figure 29 for a more detailed explanation. (Step S311: Data acquisition process) The model building unit 213 acquires one survey result data from multiple survey result data as the target survey result data. The model building unit 213 uses the user ID included in the target survey result data as a key to search the data in the survey location storage unit 32, thereby acquiring the location history and movement speed of the survey subjects for the target survey result data from the survey location storage unit 32.

[0081] (Step S312: Path estimation process) The model building unit 213 estimates the route taken by the subject from the location history of the subject obtained in step S311. For example, in Figure 30, it is estimated that user A, the subject of the study, took route X.

[0082] (Step S313: Location identification process) The model building unit 213 estimates the time spent by the research subject in the area where the learning advertising medium is visible, based on the distance in the area where the learning advertising medium is visible along the estimated path in step S312 and the movement speed obtained in step S311. The model building unit 213 calculates the distribution number by dividing the time spent by the unit time. The model building unit 213 identifies the position of the distribution number along the path by identifying the position of the research subject at each unit time along the path estimated in step S312 within the area where the learning advertising medium is visible. The model building unit 213 calculates the positional relationship between the target position and the learning advertising medium for each of the identified distribution number locations. The positional relationship with the learning advertising medium includes the distance from the target position to the learning advertising medium, the angle between the direction of movement of the research subject at the target position and the learning advertising medium, and the displacement of the target position from the front of the learning advertising medium. The displacement of the target position from the front of the learning advertising medium is the shortest distance in the direction parallel to the surface of the learning advertising medium between a plane that passes through the center of the surface of the learning advertising medium and is perpendicular to the surface of the learning advertising medium, and the target position.

[0083] For example, in Figure 30, user A stayed within the area where the learning advertisement medium was visible for 10 seconds. Assuming a unit time of 0.1 seconds, the model building unit 213 calculates the distribution number as 100 (10 seconds / 0.1 seconds). The model building unit 213 then identifies the position of path X at 0.1-second intervals within the area where the learning advertisement medium is visible. As shown in Figure 31(A), user A's movement speed is 3.45 m / s (meters / second), so it travels 0.345 m in 0.1 seconds. Therefore, as shown in Figure 31(B), the model building unit 213 uses this distance to identify the positions 0.1 seconds, ..., and 10.0 seconds after entering the area where the learning advertisement medium is visible. The model building unit 213 then calculates the positional relationship with the learning advertisement medium at each position. Since each position and the direction of movement at each position are specified, and the advertising media information storage unit 38 stores coordinates indicating the position of the advertising media for learning, the positional relationship can be calculated.

[0084] (Step S314: Position probability estimation process) The model building unit 213 distributes the visibility probability for the training advertising medium, as shown in the target survey results data, to each of the distribution number positions identified in step S313. Specifically, the model building unit 213 divides the visibility probability for the training advertising medium shown in the target survey results data by the number of distributions to obtain the value that constitutes the distribution probability. The model building unit 213 then sets the distribution probability as the visibility probability at each position of the number of distributions.

[0085] (Step S315: Unprocessed determination process) The model building unit 213 determines whether or not there is any unprocessed survey result data remaining. If there is, the model building unit 213 returns to step S311 to acquire new survey result data as the target survey result data. On the other hand, if there is no data remaining, the model building unit 213 proceeds to step S316.

[0086] (Step S316: Attribute acquisition process) The model building unit 213 obtains attribute information of advertising media that is unrelated to location, such as the media size of the advertising media used for training, from the advertising media information storage unit 38. Then, as shown in Figure 32, the model building unit 213 adds the obtained attribute information to data such as the visibility probability for each location identified from each survey result data.

[0087] (Step S317: Learning process) The model building unit 213 constructs an estimation model using the data for each location identified from each survey result as training data. Specifically, as shown in Figure 32, the model building unit 213 uses information indicating the positional relationship with the advertising medium used for training and attribute information of the advertising medium as explanatory variables for each location identified from each survey result data, and uses the visibility probability (distribution probability) as the objective variable to perform training and build an estimation model.

[0088] Refer to Figure 33 to explain the visibility probability estimation process (step S34 in Figure 28). This section explains how to estimate the probability of visibility using an estimation model in the visibility probability estimation process. The visibility probability estimation unit 223 sets each of the multiple paths within the area where the advertising medium to be estimated is visible as a target path, and each of the multiple locations on the target path as a target location. The visibility probability estimation unit 223 inputs the positional relationship between the advertising medium to be estimated and the target location on the target path into the estimation model and estimates the visibility probability for the advertising medium to be estimated at the target location. The visibility probability estimation unit 223 estimates the visibility probability for the target advertising medium along the target route from the visibility probabilities estimated for each position along the target route. Here, the visibility probability estimation unit 223 estimates the visibility probability for the target advertising medium along the target route by summing the visibility probabilities estimated for each position along the target route.

[0089] ***Effects of Embodiment 3*** As described above, the advertising media evaluation device 10 according to Embodiment 3 estimates the visibility probability for each of several paths that pass through an area where the advertising media is visible. Even within an area where the advertising media is visible, the probability of seeing the advertising media changes depending on the path. Since the advertising media evaluation device 10 according to Embodiment 3 estimates the visibility probability for each path, it becomes possible to estimate an appropriate visibility probability for each path. As a result, it becomes possible to appropriately evaluate the effectiveness of the advertising media.

[0090] ***Other configurations*** <Modification 5> In Embodiment 3, the visibility probability estimation unit 223 estimated the visibility probability of the advertising medium in the area where the advertising medium is visible by considering multiple paths that pass through the area where the advertising medium is visible. In addition, the visibility probability estimation unit 223 may estimate the visibility probability of the advertising medium in the area where the advertising medium is visible without considering multiple paths that pass through the area where the advertising medium is visible. Furthermore, the advertising media evaluation device 10 may include a display unit, which may display a comparison between the visibility probability estimated by considering multiple routes that pass through the area where the advertising media to be estimated is visible, and the visibility probability estimated without considering multiple routes that pass through the area where the advertising media to be estimated is visible.

[0091] <Variation 6> In the above embodiment, attribute information of the advertising medium and information indicating the installation environment of the advertising medium were used as training data for input to the estimation model. The model building unit 213 may also use time of day and seasonality as training data. The time of day, when building the estimation model, is the time of day when the survey subjects were in an area where the training advertising medium could be seen, and when estimating the probability of visibility, it is the time of day at the time of estimation. Seasonality indicates the season and latitude.

[0092] <Example 7> In Embodiment 3, the learning unit 21 did not include a reliability determination unit 211 and an adjustment unit 214. However, the learning unit 21 may be configured to include at least one of the reliability determination unit 211 and the adjustment unit 214. In this case, the visibility probability estimation unit 223 estimates the visibility probability for each of the multiple routes that pass through the area where the advertising medium is visible, using the residual data generated by the reliability determination unit 211. Alternatively, the visibility probability estimation unit 223 estimates the visibility probability for each of the multiple routes that pass through the area where the advertising medium is visible, using the adjusted data generated by the adjustment unit 214.

[0093] <Differentiation Example 8> In Embodiment 1, each functional component was implemented in software. However, in Modification 8, each functional component may be implemented in hardware. The differences between this Modification 8 and Embodiment 1 will be explained below.

[0094] When each functional component is implemented in hardware, the advertising media evaluation device 10 includes an electronic circuit 15 instead of a processor 11, memory 12, and storage 13. The electronic circuit 15 is a dedicated circuit that implements the functions of each functional component, as well as the functions of the memory 12 and storage 13.

[0095] Electronic circuits 15 can include single circuits, complex circuits, programmed processors, parallel programmed processors, logic ICs, GAs, ASICs, and FPGAs. GA stands for Gate Array. ASIC stands for Application Specific Integrated Circuit. FPGA stands for Field-Programmable Gate Array. Each functional component may be implemented in a single electronic circuit 15, or each functional component may be implemented by distributing them across multiple electronic circuits 15.

[0096] <Modification 9> As a variation 9, some of the functional components may be implemented in hardware, while other functional components may be implemented in software.

[0097] The processor 11, memory 12, storage 13, and electronic circuit 15 are collectively referred to as the processing circuit. In other words, the function of each functional component is realized by the processing circuit.

[0098] Furthermore, the term "part" in the above explanation may be replaced with "circuit," "process," "procedure," "processing," or "processing circuit."

[0099] The various aspects of this disclosure are summarized below as an appendix. (Note 1) A visibility probability estimation unit estimates the visibility probability for each of several paths that pass through an area where the target advertising medium is visible, A viewer count calculation unit calculates the number of people who viewed the estimated target advertising medium on each of the multiple routes by multiplying the traffic volume on the target route by the visibility probability estimation unit for the target route, and then calculates the number of people who viewed the estimated target advertising medium from the number of people calculated for each of the multiple routes. An advertising media evaluation device equipped with the following features. (Note 2) The visibility probability estimation unit estimates the visibility probability of the estimated advertising medium for each of the plurality of routes, using at least one of the distance to the area in which the estimated advertising medium is visible along the route and the positional relationship between the route and the estimated advertising medium. The advertising media evaluation device described in Appendix 1. (Note 3) The aforementioned advertising media evaluation device further, A model building unit constructs an estimation model for estimating the probability of viewing the target advertising medium for a person who has passed through an area where the target advertising medium is visible, using, for each of the multiple survey result data obtained by surveying the probability of viewing the learning advertising medium for a survey subject who has passed through one of several routes that pass through an area where the learning advertising medium is visible, the model building unit constructs an estimation model for estimating the probability of viewing the target advertising medium for a person who has passed through an area where the target advertising medium is visible, using, in addition to the probability of viewing the learning advertising medium shown in the target survey result data, the distance of the route taken by the survey subject for the target survey result data to the area where the learning advertising medium is visible, and the positional relationship between the route taken by the survey subject for the target survey result data and the learning advertising medium. Equipped with, The visibility probability estimation unit estimates the visibility probability for each of the multiple paths using the estimation model constructed by the model construction unit. The advertising media evaluation device described in Appendix 1 or 2. (Note 4) The model building unit distributes the visibility probability for the learning advertising medium indicated by the target survey results data to each of the multiple locations along the path taken by the survey subjects for the target survey results data, and for each of the multiple locations, constructs the estimation model using the visibility probability at the target location, as well as the positional relationship between the target location and the learning advertising medium. The advertising media evaluation device described in Appendix 3. (Note 5) The model building unit uses the distribution number, which is the value obtained by dividing the viewing probability for the learning advertising medium shown in the survey results data of the subject by the distribution number, which is the value obtained by dividing the time spent by the subject in an area where the learning advertising medium is visible by the unit time, as the distribution probability. The positions of the subject on the subject's path for each unit time are defined as the multiple positions, and the distribution probability is used as the viewing probability at each of the multiple positions. The advertising media evaluation device described in Appendix 4. (Note 6) The visibility probability estimation unit estimates the visibility probability of the advertising medium along the target route by inputting the positional relationship between the advertising medium and each of the multiple locations on the target route, and the advertising medium at each of the multiple locations, for each of the multiple routes in the area where the advertising medium to be estimated is visible, into the estimation model, and estimating the visibility probability of the advertising medium at each of the multiple locations. An advertising media evaluation device as described in any one of the items 3 to 5 in the appendix. (Note 7) The positional relationship between each of the aforementioned locations and the learning advertising medium includes the angle between the direction of movement of the survey subject at each of the aforementioned locations and the learning advertising medium. An advertising media evaluation device as described in any one of the items 4 to 6 of the appendix. (Note 8) The visibility probability estimation unit estimates the visibility probability of the advertising medium in the area where the advertising medium is visible, without considering multiple paths that pass through the area where the advertising medium is visible. The aforementioned advertising media evaluation device further, A display unit that displays the visibility probability estimated for each of the multiple routes and the visibility probability estimated without considering the multiple routes that pass through the area where the advertising medium being estimated is visible. An advertising media evaluation device as described in any one of the appendices 1 to 7, comprising: (Note 9) The computer estimates the probability of viewing the target advertising medium for each of several paths that pass through an area where the target advertising medium is visible. An advertising media evaluation method comprising a computer that, for each of the multiple routes, multiplies the traffic volume for the target route by the visibility probability for the target route to calculate the number of people who viewed the estimated target advertising media on the target route, and then calculates the number of people who viewed the estimated target advertising media from the number of people calculated for each of the multiple routes. (Note 10) A visibility probability estimation process that estimates the visibility probability of the advertising medium for each of several paths that pass through an area where the advertising medium to be estimated is visible, For each of the aforementioned multiple routes, the number of people who viewed the estimated target advertising medium on the target route is calculated by multiplying the traffic volume on the target route by the visibility probability estimation process, and the number of people who viewed the estimated target advertising medium is calculated from the number of people calculated for each of the multiple routes. An advertising media evaluation program that makes a computer function as an advertising media evaluation device.

[0100] The embodiments and variations of this disclosure have been described above. Some of these embodiments and variations may be implemented in combination. Alternatively, some or all of them may be implemented in part. However, this disclosure is not limited to the embodiments and variations described above, and various modifications are possible as needed. [Explanation of Symbols]

[0101] 10 Advertising media evaluation device, 11 Processor, 12 Memory, 13 Storage, 14 Communication interface, 20 Control unit, 21 Learning unit, 211 Reliability determination unit, 212 Model construction unit, 213 Adjustment unit, 22 Calculation unit, 221 Visible route determination unit, 222 Traffic volume calculation unit, 223 Visibility probability estimation unit, 224 Number of viewers calculation unit, 30 Memory unit, 31 Survey information storage unit, 32 Survey location storage unit, 33 Model storage unit, 34 Visible area storage unit, 35 Road information storage unit, 36 Visible route storage unit, 37 Traffic volume storage unit, 38 Advertising media information storage unit, 39 Environmental information storage unit, 40 Route traffic volume storage unit, 41 Visibility probability storage unit, 42 Number of viewers storage unit, 43 Reference attribute storage unit.

Claims

1. A visibility probability estimation unit estimates the visibility probability of the advertising medium for each of several paths that pass through an area where the advertising medium to be estimated is visible, A viewer count calculation unit calculates the number of people who viewed the estimated target advertising medium on each of the multiple routes by multiplying the traffic volume on the target route by the visibility probability estimation unit for the target route, and then calculates the number of people who viewed the estimated target advertising medium from the number of people calculated for each of the multiple routes. An advertising media evaluation device equipped with the following features.

2. The visibility probability estimation unit estimates the visibility probability of the estimated advertising medium for each of the plurality of routes, using at least one of the distance to the area in which the estimated advertising medium is visible along the route and the positional relationship between the route and the estimated advertising medium. The advertising media evaluation apparatus according to claim 1.

3. The aforementioned advertising media evaluation device further, A model building unit constructs an estimation model for estimating the probability of viewing the target advertising medium for a person who has passed through an area where the target advertising medium is visible, using, for each of the multiple survey result data obtained by surveying the probability of viewing the learning advertising medium for a survey subject who has passed through one of several routes that pass through an area where the learning advertising medium is visible, the model building unit constructs an estimation model for estimating the probability of viewing the target advertising medium for a person who has passed through an area where the target advertising medium is visible, using, in addition to the probability of viewing the learning advertising medium shown in the target survey result data, the distance of the route taken by the survey subject for the target survey result data to the area where the learning advertising medium is visible, and the positional relationship between the route taken by the survey subject for the target survey result data and the learning advertising medium. Equipped with, The visibility probability estimation unit estimates the visibility probability for each of the multiple paths using the estimation model constructed by the model construction unit. The advertising media evaluation apparatus according to claim 12.

4. The model building unit distributes the visibility probability for the learning advertising medium indicated by the target survey results data to each of the multiple locations along the path taken by the survey subjects for the target survey results data, and for each of the multiple locations, constructs the estimation model using the visibility probability at the target location, as well as the positional relationship between the target location and the learning advertising medium. The advertising media evaluation apparatus according to claim 3.

5. The model building unit uses the distribution number, which is the value obtained by dividing the viewing probability for the learning advertising medium shown in the survey results data of the subject by the distribution number, which is the value obtained by dividing the time spent by the subject in an area where the learning advertising medium is visible by the unit time, as the distribution probability. The positions of the subject on the subject's path for each unit time are defined as the multiple positions, and the distribution probability is used as the viewing probability at each of the multiple positions. The advertising media evaluation apparatus according to claim 4.

6. The visibility probability estimation unit estimates the visibility probability of the advertising medium along the target route by inputting the positional relationship between the advertising medium and each of the multiple locations on the target route, and the advertising medium at each of the multiple locations, for each of the multiple routes in the area where the advertising medium to be estimated is visible, into the estimation model, and estimating the visibility probability of the advertising medium at each of the multiple locations. The advertising media evaluation apparatus according to claim 3.

7. The positional relationship between each of the aforementioned locations and the learning advertising medium includes the angle between the direction of movement of the survey subject at each of the aforementioned locations and the learning advertising medium. The advertising media evaluation apparatus according to claim 4.

8. The visibility probability estimation unit estimates the visibility probability of the advertising medium in the area where the advertising medium is visible, without considering multiple paths that pass through the area where the advertising medium is visible. The aforementioned advertising media evaluation device further, A display unit that displays the visibility probability estimated for each of the multiple routes and the visibility probability estimated without considering the multiple routes that pass through the area where the advertising medium being estimated is visible. The advertising media evaluation device according to claim 1, comprising:

9. The computer estimates the probability of viewing the target advertising medium for each of several paths that pass through an area where the target advertising medium is visible. An advertising media evaluation method comprising a computer that, for each of the multiple routes, multiplies the traffic volume for the target route by the visibility probability for the target route to calculate the number of people who viewed the estimated target advertising media on the target route, and then calculates the number of people who viewed the estimated target advertising media from the number of people calculated for each of the multiple routes.

10. A visibility probability estimation process that estimates the visibility probability of the advertising medium for each of several paths that pass through an area where the advertising medium to be estimated is visible, For each of the aforementioned multiple routes, the number of people who viewed the estimated target advertising medium on the target route is calculated by multiplying the traffic volume on the target route by the visibility probability estimation process, and the number of people who viewed the estimated target advertising medium is calculated from the number of people calculated for each of the multiple routes. An advertising media evaluation program that makes a computer function as an advertising media evaluation device.

Images