Information processing systems, information processing methods, and programs

Abstract

This invention provides a system, method, and program to ensure the quality of services provided to service providers involved in video streaming. [Solution] In the information processing system, the data acquisition unit 121 of the control unit acquires distribution video-related data including first related data. The first related data includes at least one of statistical data, distributor characteristic data, and behavioral history data. The statistical data includes at least one of data relating to the number of viewers of the distribution video and data relating to the distribution time of the distribution video. The distributor characteristic data is data that shows the characteristics of the distributor of the distribution video, and the behavioral history data is data that shows the behavioral history of the distributor or viewers. The data estimation unit 122 generates distribution video estimation data including first estimation data based on the distribution video-related data. The first estimation data includes at least one of data relating to the performance of the distributor or the distribution video and viewer demographic data.

Description

Technical Field

[0001] The present invention relates to an information processing system, an information processing method, and a program.

Background Art

[0002] In recent years, services that realize communication by video distribution have been rapidly spreading, and various information processing systems related to video distribution platforms have been proposed (for example, Patent Document 1). Along with the spread of such video distribution platforms, various services are being developed for distributors, companies, etc. related to the distributed video (hereinafter also referred to as service recipients), such as displaying banner roll advertisements in live distributor media.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In an information society where the types and amounts of data used are increasing, it has become extremely complicated to determine what data to use and how to provide services using it. In such a situation, it is desired to further improve the quality of services provided to distributors, viewers, companies, etc. related to distributed videos.

[0005] The present invention aims to ensure the quality of services provided to those directly or indirectly related to distributed videos.

Means for Solving the Problems

[0006] According to the present invention, an information processing system is provided that uses distributed video-related data, comprising a data acquisition unit and a data estimation unit, wherein the data acquisition unit is configured to acquire the distributed video-related data, the distributed video-related data includes first related data, the first related data includes at least one of statistical data, distributor characteristic data and behavioral history data, the statistical data includes at least one of data relating to the number of viewers of the distributed video and data relating to the distribution time of the distributed video, the distributor characteristic data is data indicating the characteristics of the distributor of the distributed video, the behavioral history data is data indicating the behavioral history of the distributor or the viewers, and the data estimation unit is configured to generate distributed video estimation data based on the distributed video-related data, the distributed video estimation data includes first estimation data, the first estimation data includes at least one of data relating to the performance of the distributor or the distributed video and the demographic data of the viewers.

[0007] According to the present invention, the first related data included in the distributed video related data includes at least one of statistical data, distributor characteristic data, and behavioral history data, thereby enabling the data estimation unit to generate first estimated data that includes at least one of data related to the performance of the distributor or the distributed video, and demographic data of the viewers. Therefore, service providers can provide services related to distributed videos using the first estimated data, and can ensure the quality of services provided to those directly or indirectly involved with the distributed video. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 shows an example of the system configuration of the information processing system 100 according to the embodiment. [Figure 2] Figure 2 is a block diagram showing the hardware configuration of the information processing device 1. [Figure 3] Figure 3 is a functional block diagram of the control unit 12 shown in Figure 2. [Figure 4] Figure 4 is an explanatory diagram of the data related to the input and output of the data estimation unit 122. [Figure 5] Figure 5 shows a schematic diagram of a neural network 122N, which is an example of a pre-trained model. [Figure 6] Figure 6 is an explanatory diagram of an example of how the information processing device 1 can be used when providing useful information to any company. [Figure 7] Figure 7 is an explanatory diagram illustrating example 2 of the use of the information processing device 1 when providing useful information to any company. [Figure 8] Figure 8 is an explanatory diagram of an example of how the information processing device 1 can be used when providing useful information to each distributor. [Figure 9] Figure 9 is an explanatory diagram of an example of how the information processing device 1 can be used when providing useful information to each distributor. [Figure 10] Figures 10A and 10B are explanatory diagrams illustrating example 3 of the use of the information processing device 1 when providing useful information to each distributor. Figure 10C is a modified version of Figures 10A and 10B. [Figure 11] Figure 11 is an explanatory diagram of an example 4 of how the information processing device 1 can be used when providing useful information to each distributor. [Modes for carrying out the invention]

[0009] Embodiments of the present invention will be described below with reference to the drawings. The various features shown in the embodiments below can be combined with each other. Furthermore, each feature constitutes an independent invention.

[0010] 1. System configuration of information processing system 100 As shown in Figure 1, the information processing system 100 includes an information processing device 1 configured to connect to a communication network 2 (for example, the Internet). The information processing system 100 is configured to generate distribution video estimation data d2 using distribution video related data d1. The information processing system 100 (information processing device 1) has various functions such as data acquisition, generation, editing, storage, and output, and is configured to manage various types of data.

[0011] Here, the aforementioned streaming video-related data d1 is data relating to a specific streaming video or the distributor of a specific streaming video. Furthermore, the streaming video estimation data d2 is data based on the streaming video-related data d1 and is analysis result data to be provided to users of the information processing system 100 (information processing device 1). In other words, the information processing system 100 (information processing device 1) can provide data related to viewers' interests, concerns, and preferences through information processing (analysis) using the streaming video-related data d1. This makes it possible, for example, to recommend the selection of advertisements that are more effective for viewers when displaying advertisements on a video SNS (video streaming platform). Examples of video SNS (video distribution platforms) include live streaming services and various social networking services, but the theory is not limited to these and can be applied to video distribution platforms with similar functions.

[0012] It should be noted that the various functions of the information processing device 1 can be realized, for example, by exchanging data pre-stored in the storage unit 11 of the information processing device 1 or via an external storage medium. Therefore, it is not essential that the information processing device 1 has the function to communicate via the communication network 2. However, in this embodiment, it will be described as having this function.

[0013] Although not shown in FIG. 1, the information processing system 100 may include a user terminal (e.g., a desktop PC, a notebook PC, a smartphone, a tablet terminal, etc.) that can communicate with the information processing apparatus 1. In this case, it may be a method (native application method) of downloading a program of a desired application to the user terminal and communicating with the information processing apparatus 1 to receive service provision, or it may be a method (web application method) of communicating with the information processing apparatus 1 via the web browser of the user terminal to receive service provision.

[0014] Further, a part or all of the communication network 2 may be a closed network separated from the Internet. Each component of the information processing system 100 such as the information processing apparatus 1 has one or more functions (functional units). Each component may be configured as a single device as shown in FIG. 1, or may be configured by a plurality of independent devices configured to be able to exchange information. The same applies to each functional unit such as the control unit 12 of the information processing apparatus 1 described later.

[0015] 1-1. Information Processing Apparatus 1 As shown in FIG. 2, the information processing apparatus 1 includes a communication unit 10, a storage unit 11, a control unit 12, an output unit 13, and an input unit 14, and these components are electrically connected via a communication bus 15 inside the information processing apparatus 1. As shown in FIG. 3, the control unit 12 includes a data acquisition unit 121, a data estimation unit 122, and a data presentation unit 123.

[0016] Each of the above components may be implemented by software or by hardware. When implemented by software, various functions can be realized by a CPU executing a computer program. The program may be stored in a non-transitory computer-readable recording medium, may be provided for download from an external server, or may be realized by so-called cloud computing that reads a program stored in an external storage unit to realize functions. When implemented by hardware, for example, it can be implemented by various circuits such as ASIC, FPGA, or DRP. In the embodiments, various information and concepts including these are handled, and these are represented by the high and low of signal values or quantum bits as a set of binary bits composed of 0 or 1, and communication and operations can be executed in the above software or hardware modes. Note that the software may be a general-purpose OS or a dedicated OS.

[0017] The communication unit 10 may adopt a wired communication means such as, for example, USB, IEEE1394, Thunderbolt (registered trademark), wired LAN network communication, etc. Note that the communication unit 10 may adopt a configuration connected to the communication network 2 via a wireless communication means such as, for example, wireless LAN network communication, mobile communication such as 3G / LTE / 5G, Bluetooth (registered trademark) communication, etc. Further, the communication unit 10 may have a configuration that combines the above-described wired communication means and wireless communication means.

[0018] The storage unit 11 stores various values, such as various programs, constants, variables, and settings of the information processing device 1 executed by the control unit 12. The storage unit 11 also stores data obtained by communicating with external devices or external servers of the information processing device 1. The storage unit 11 can employ storage devices such as solid-state drives (SSDs) or storage media such as random-access memory (RAM) that stores temporarily necessary information (arguments, arrays, etc.) related to program calculations. In addition to the storage unit 11, the information processing device 1 may also use external storage (for example, external storage media, cloud storage, etc.).

[0019] The control unit 12 is configured to execute information processing of the information processing device 1. The control unit 12 can be configured as, for example, a central processing unit (CPU), and in this embodiment, the control unit 12 is an example of a processor capable of executing programs related to the operation (steps) of the information processing device 1, which will be described later. The control unit 12 realizes various functions related to the information processing device 1 by, for example, reading programs stored in the storage unit 11. Furthermore, the information processing of software in the information processing device 1 is realized by, for example, the processing of various programs stored in the storage unit 11 by the control unit 12 as hardware.

[0020] The output unit 13 is, for example, a display unit of the information processing device 1. The output unit 13 may be, for example, included in the housing of the information processing device 1 or attached externally. The output unit 13 displays a graphical user interface (GUI) screen that can be operated by the user. The output unit 13 can employ display devices such as a CRT display, liquid crystal display, organic EL display, plasma display, or electronic paper display, as well as display devices such as a lit-up light or a projector. It is optional whether or not the information processing device 1 includes an output unit 13. For example, the output of the information processing device 1 may be displayed on a display unit located in a separate location independent of where the information processing device 1 is installed. The output unit 13 may also have a device that outputs sound.

[0021] The input unit 14 is configured to receive operation inputs, for example, from the administrator of the information processing device 1. The input unit 14 may be included in the housing of the information processing device 1 or it may be an external component. The input unit 14 can be, for example, a touch panel, switch buttons, a mouse, a keyboard camera, a scanner, etc. Whether or not the information processing device 1 includes an input unit 14 is optional. For example, operation inputs to the information processing device 1 may be received by the information processing device 1 via an information processing terminal located in a separate location independent of where the information processing device 1 is installed.

[0022] 2. Functional Configuration The functional configuration of the information processing device 1 according to this embodiment will be described with reference to Figures 3 and 4. Information processing by software stored in the storage unit 11 is specifically realized by the control unit 12, which is an example of hardware, and each functional unit included in the control unit 12 is executed.

[0023] 2-1. Data acquisition unit 121 The data acquisition unit 121 is configured to acquire data (for example, streaming video-related data d1) that will be processed by the data estimation unit 122, which will be described later. When the data acquisition unit 121 acquires such data, the streaming video-related data d1 and the like are stored in the storage unit 11. The data acquisition unit 121 is also configured to acquire various types of data received from the input unit 14. The data acquisition unit 121 may also be configured to collect (acquire) various types of data from any external server connected to the Internet via the communication unit 10. The data acquisition unit 121 may also have the function of receiving similar data and transmitting it to other functional units. Furthermore, the data acquisition unit 121 may have the function of performing predetermined processing (preprocessing) on ​​the received data to convert it into desired data before transmitting it to other functional units, or it may have both of these functions. In other words, acquisition may include not only simply acquiring data, but also processing to modify the data through preprocessing.

[0024] Next, we will explain the various types of data used by the information processing device 1 (acquired by the data acquisition unit 121). The data acquisition unit 121 has the function of acquiring input data for the data estimation unit 122. The input data here refers to the distributed video-related data d1.

[0025] The distributed video-related data d1 includes first related data d11 and second related data d12. In this embodiment, the distributed video-related data d1 is described as including both first related data d11 and second related data d12, but it may include only one of them (for example, only first related data d11 or only second related data d12). The distributed video-related data d1 can encompass various data related to the video distribution platform and is a data set that forms the basis for the analysis and estimation processing of the information processing system 100.

[0026] The data acquisition unit 121 may also receive output from the generating AI model (large-scale language model) and acquire it as streaming video-related data d1. In other words, any information processing device (not limited to information processing device 1) may input data such as streaming videos, along with prompts (prompts to generate the output described later), into the generating AI model, causing the generating AI model to output data such as the streamer's statements in the streaming video and viewer comments in the streaming video. The generating AI model may then be made to perform further statistical processing on this output data. The data acquisition unit 121 can then acquire this data generated by the generating AI model as streaming video-related data d1. The generating AI model may be provided by the information processing device 1, or it may be provided by an external server independent of the information processing device 1.

[0027] 2-1-1. First related data d11 of the distributed video related data d1 The first related data d11 includes at least one of the following: statistical data d111, distributor characteristic data d112, and behavioral history data d113. In this embodiment, the first related data d11 is described as including all of these, but is not limited to this. Specifically, for example, the first related data d11 may include statistical data d111 and distributor characteristic data d112, or statistical data d111 and behavioral history data d113, or distributor characteristic data d112 and behavioral history data d113. Alternatively, the first related data d11 may consist only of statistical data d111, or only of distributor characteristic data d112, or only of behavioral history data d113.

[0028] At least a portion of the first related data d11 may be data obtained using an API. For example, the data acquisition unit 121 may be configured to acquire statistical data d111 as API-linked data from an external information processing device of the information processing system 100.

[0029] <Statistical data d111> The statistical data d111 includes at least one of the data t1 relating to the number of viewers of the streamed video and the data t2 relating to the streaming time of the streamed video. In this embodiment, it is described as including both the data t1 relating to the number of viewers of the streamed video and the data t2 relating to the streaming time of the streamed video, but it is not limited to this, and it may include only the data t1 relating to the number of viewers of the streamed video, or only the data t2 relating to the streaming time of the streamed video.

[0030] For data related to the number of viewers of a streamed video, t1 can, for example, use the average number of concurrent viewers. Here, the average number of concurrent viewers can be calculated by dividing the total number of concurrent viewers recorded during the stream (e.g., viewers per minute) by the streaming time (the number of minutes of time measured). By using the average number of concurrent viewers, it is possible to understand how engaged viewers were throughout the entire stream, and by looking at the overall average rather than a temporary peak, it can contribute to stable analysis. The average number of concurrent viewers can be used to understand how consistently the streamer's fan base is watching, and can be used to review streaming schedules and content. From a platform perspective, it can be used to evaluate the popularity and performance of a stream and reflect it in recommendation algorithms. From an advertiser perspective, it can be used as a criterion for determining when to place ads to achieve the greatest effect.

[0031] Furthermore, as data t1 related to the number of viewers of the streamed video, in addition to the average number of concurrent viewers, peak viewership, return viewer rate, viewer retention rate, viewership by region, viewership by device, and cumulative viewership may also be used. Of course, these may be used individually or in combination of two or more. Peak viewership corresponds to the maximum number of simultaneous viewers at any given point in time during a broadcast. This data can be used to identify the most attention-grabbing moments of a broadcast and contribute to the analysis of the broadcaster's actual fan base, popularity, popular content, and optimal broadcast timing. The return viewer rate corresponds to the percentage of viewers who have watched past broadcasts and can contribute to analysis of viewer and fan base retention. The viewer retention rate corresponds to the percentage of viewers who watched for a predetermined threshold percentage or more of the total broadcast time (e.g., setting the threshold at 50% if the broadcast is 1 hour long and the average viewing time is 30 minutes), and can contribute to the analysis of content quality. Regional viewer numbers are data that classifies viewers by country and region, and can contribute to analysis to understand regional trends. The number of viewers by device corresponds to the number of viewers for each viewing device, such as smartphones, PCs, tablets, and connected TVs, and can contribute to analysis to understand trends for each device.

[0032] In the above, data is specified by a rate (percentage), such as the return visitor rate, but this is not the only way to do so; absolute or relative values ​​may also be used. For example, if an absolute value is used, the return visitor rate would be the number of returning viewers. Alternatively, the change in each value described above may also be used. For example, it could be the change in the return visitor rate or the number of returning viewers over a predetermined period. Other examples of change include growth rate and fluctuation rate. As an example of growth rate, the increase in the number of viewers at the start of distribution and the end of distribution divided by the number of viewers at the start can be used. As an example of fluctuation rate, the ratio of the maximum number of viewers to the minimum number of viewers during the distribution period can be used. The same applies to the data described in Section 2-1 and Section 2-2 below.

[0033] The data t2 related to the streaming time of the streamed video can include, for example, viewing time slots or streaming time ratio classifications (for example, in the case of game video streaming, streaming time ratio classifications by game title). Viewing time refers to data about the time when a streamed video was viewed, corresponding to, for example, the exact time the video was being watched (e.g., 10 AM to 2 PM). Viewing time could indicate the time period with the most viewers, or it could indicate the time period when the number of viewers exceeded a certain threshold. Viewing time can contribute to analyses such as optimizing the streaming schedule. The streaming time ratio classification corresponds to the proportion of streaming time dedicated to each game title within a video, for example, if the video is about games. For instance, if a streamer broadcasts 2 hours of game A and 1 hour of game B during a 3-hour stream, game A would account for 66.7% and game B for 33.3%. The streaming time ratio classification allows us to understand which games the streamer is focusing on. Furthermore, since games that the streamer focuses on are likely to be of higher interest to viewers, it can indirectly contribute to analysis related to viewer interests.

[0034] The data t2 related to the streaming time of the streamed video may specifically include, for example, viewing time slots and streaming time ratio classifications (streaming time ratio classifications for game titles), as well as average streaming time, peak times for viewer growth rate and viewer decline rate, viewer dwell time, streaming interval, and total streaming time. Of course, these may be used individually or in combination of two or more.

[0035] Average broadcast time refers to, for example, the average broadcast duration for a single broadcast. This can be useful for analysis, such as determining whether long broadcasts or short broadcasts are the norm. The peak time for viewer growth is, for example, the time during a live stream when the number of viewers increased most sharply. Since it is assumed that specific actions or events during the live stream had an impact on viewers, this can contribute to the analysis of such influences. Also, for example, if there is a correlation between the peak times of viewer growth in two live stream videos, it is possible to analyze that there is a correlation between the content of those streams. Similarly, if there is a correlation between the peak time of viewer growth in one live stream video and the peak time of viewer decline in a different live stream video, it is possible to analyze that there is a correlation between the content of those streams. Note that here, we have explained using one pair (two) of live stream videos as an example, but it is also possible to analyze whether there is a correlation with three or more videos. Viewer engagement time is, for example, the average amount of time viewers continued to watch the stream. This allows for analysis of how well the audience was retained. The distribution interval is, for example, the time between the previous distribution and the next distribution. Since a correlation is expected between distribution frequency and the number of viewers, such a correlation can contribute to the analysis. Total streaming time refers to, for example, the total streaming time within a predetermined period. Since a correlation is expected between total streaming time and the number of viewers, this correlation can contribute to the analysis.

[0036] <Streamer Characteristics Data d112> The streamer characteristics data d112 is data that indicates the characteristics of the streamer of a streamed video. The streamer characteristics data d112 may be, for example, analytics information held by various video streaming platforms, or information obtained through research. The streamer characteristics data d112 includes at least one of the following: platform type data s1, streamer analytics data s2, and streamer data s3.

[0037] In this embodiment, the distributor characteristics data d112 is described as including, but is not limited to, platform type data s1, distributor analytics data s2, and distributor data s3. It may include platform type data s1 and distributor analytics data s2, or platform type data s1 and distributor data s3, or distributor analytics data s2 and distributor data s3. Of course, the distributor characteristics data d112 may include any of these individually.

[0038] Platform type data s1 is data used to identify the type of video streaming platform used by the streamer. Platform type data s1 can contribute to analysis related to the platform used by the streamer.

[0039] The streamer analytics data s2 includes data such as the streamer's MAU (Monthly Active Users), revenue, and the gender, age, and location of the streamer's viewers. Streamer analytics data s2 can contribute to the analysis of factors particularly important to advertisers, such as the streamer's MAU. As with the above, streamer analytics data s2 can be used individually or in combination of two or more. Furthermore, streamer analytics data s2 may be obtained under the streamer's authentication process. For example, if a piece of content (streaming video, streaming channel, streamer) has 1 million monthly views, and each viewer plays it 10 times, the MAU would be 100,000. Normally, it is difficult to estimate MAU solely from the number of views. Therefore, if the MAU of a piece of content (streaming video, streaming channel, streamer) can be estimated, it is possible to understand the channel's reach (how many people are watching) and improve the quality of services provided to those directly or indirectly involved with the streaming video.

[0040] The streamer data s3 corresponds to data such as whether the streamer shows their face in video streaming (including avatar + voice if they don't), the streamer's gender, and the streamer's age. Streamer data s3 is basic data corresponding to the streamer's status and attributes, and can contribute to the analysis of such basic information. As with the above, streamer data s3 can be used individually or in combination of two or more.

[0041] <Behavioral history data d113> Behavioral history data d113 is data that shows the behavioral history of the broadcaster or viewer. Behavioral history data d113 corresponds to data such as viewing history, web browsing history, bookmarks, and the number of "likes". A "like" is a way for viewers to show positive feedback on a broadcasted video or content, and positive feedback is identified when a viewer clicks or taps the desired button when they feel it is good. Behavioral history data d113 can contribute to the analysis of the behavioral history of the streamer themselves and their viewers. It is especially beneficial if behavioral history data d113 contains time information. For example, if there is a positive or negative action towards a streamed video at a certain point in time, that timing represents the moment when the person's emotions were stirred, which can lead to useful analysis. In addition, behavioral history data d113 can reveal the websites that the streamer or viewers frequently visit, or the genres of websites they visit, which can contribute to estimating their hobbies and preferences. As with the above, behavioral history data d113 may be used individually or in combination of two or more.

[0042] 2-1-2. Related data d12 for streaming video related data d1 The second related data d12 includes the streamer video data r1 and the comment data r2 from the comment section of the streamed video. In this embodiment, it is described as including both the streamer video data r1 and the comment data r2 from the comment section of the streamed video, but it is not limited to this, and it may include only the streamer video data r1 or only the comment data r2 from the comment section of the streamed video. The streamer video data r1 is video data streamed by a specific streamer. Note that streamer video data r1 consists of a combination of audio, video, text within the video, and subtitles within the video, not just video alone. However, it is not necessary for all of these to be combined; video alone is also acceptable.

[0043] The comment data r2 in the comment section of a live stream video is data that shows, for example, text-based feedback, impressions, questions, opinions, and chat counts posted by viewers to the live stream video. Note that the comment data r2 in the comment section of a live stream video may be used after pre-processing to remove so-called spam or trolling comments. Furthermore, the comment data r2 in the comment section of a live stream video is not limited to Japanese; it may be in foreign languages ​​and may include symbols and emojis. In this embodiment, the language of the comment data r2 in the comment section of the streamed video can be analyzed, and the viewer's place of residence can be estimated by assuming that the viewer is watching from a region where that language is spoken. The place of residence data obtained in this way based on the comment data r2 may be organized as belonging to the statistical data d111 described above.

[0044] 2-2. Data Estimation Unit 122 The data estimation unit 122 is configured to process, analyze, and estimate the input data collected by the data acquisition unit 121. In this embodiment, the data estimation unit 122 is configured to generate estimated streaming video data d2 based on a learning model that takes streaming video-related data d1 as input and outputs estimated streaming video data d2.

[0045] The generation of estimated streaming video data d2 makes it possible to ensure the quality of services provided to those directly or indirectly involved with the streaming video. For example, if the estimated streaming video data d2 is the estimated monthly active users (MAU) of a certain streamer (described later), advertisers can use this estimated MAU to decide whether or not to display their own products or services when this streamer's videos are being streamed.

[0046] Here, the learning model in the data estimation unit 122 is pre-trained using the streaming video-related data d1 and the corresponding streaming video estimation data d2.

[0047] An example of how the data estimation unit 122 generates streaming video estimation data d2 will be described. Figure 5 is a schematic diagram of a neural network 122N, which is an example of a trained model. Input data (streaming video related data d1) defined by various parameters is input to the first layer N1. The input data is output from the computation node of the first layer N1 to the computation node of the second layer N2. At this time, the value output from the computation node of the first layer N1 is multiplied by the weight w set between each computation node and input to the computation node of the second layer N2. The computation node of the second layer N2 adds up the input values ​​from the computation node of the first layer N1 and inputs this value (or a value obtained by adding a predetermined bias value to this) to a predetermined activation function. The output value of the activation function is then propagated to the next node, the computation node of the third layer N3. At this time, the value obtained by multiplying the weight w set between the computation node of the second layer N2 and the computation node of the third layer N3 by the above output value is input to the computation node of the third layer N3. The computation node in the third layer N3 sums the input values ​​and outputs the sum as the output signal. Alternatively, the computation node in the third layer N3 may sum the input values, add a bias value to the sum, input the result into an activation function, and output that result as the output signal. This generates the estimated streaming video data d2 as output data. Note that the figures shown in Figure 5 are for illustrative purposes only and are not limited to this trained model.

[0048] The data output by the data estimation unit 122 (estimated data) is classified into first estimated data d21 and second estimated data d22. In other words, the streamed video estimated data d2 includes both the first estimated data d21 and the second estimated data d22. It should be noted that the first estimated data d21 is not necessarily estimated from the first related data d11, nor is the second estimated data d22 necessarily estimated from the second related data d12. In other words, the second estimated data d22 may be estimated from the first related data d11, or the first estimated data d21 may be estimated from the second related data d12. Furthermore, the first estimated data d21 may be estimated from the first related data d11 and the second related data d12. Conversely, the second estimated data d22 may be estimated from the first related data d11 and the second related data d12. In this way, the dataset consisting of input data and output data can be determined depending on how the learning model (neural network) of the data estimation unit 122 is trained.

[0049] Here, the data estimation unit 122 may have multiple independent learning models (neural networks). For example, it may have a first learning model (first neural network) that generates the first estimated data d21 from the first related data d11, and a second learning model (second neural network) that generates the second estimated data d22 from the second related data d12, with separate learning models depending on the estimated data to be generated.

[0050] 2-2-1. Estimated data d21 of the streamed video data d2 The first estimated data d21 includes at least one of the following: data u1 relating to the performance of the broadcaster or the broadcast video, demographic data u2 of the viewer, and broadcaster u3 different from the broadcaster of the broadcast video. In the embodiment, the first estimated data d21 is described as including all of these, but is not limited to this. Specifically, for example, the first estimated data d21 may include data u1 related to the performance of the broadcaster or the broadcast video, and demographic data u2 of the viewers. Alternatively, the first estimated data d21 may include data u1 related to the performance of the broadcaster or the broadcast video, and broadcaster u3 different from the broadcaster of the broadcast video. Alternatively, the first estimated data d21 may include demographic data u2 of the viewers and broadcaster u3 different from the broadcaster of the broadcast video. Alternatively, the first estimated data d21 may consist only of data u1 related to the performance of the broadcaster or the broadcast video, or only of demographic data u2 of the viewers, or only of broadcaster u3 different from the broadcaster of the broadcast video.

[0051] Data u1 related to the performance of a streamer or streamed video can include at least one of the following: the estimated MAU (Monthly Active Users) of the streamer or streamed video, and the streamer's revenue. Of course, these can be used individually or in combination of two or more.

[0052] The viewer's demographic data u2 can include at least one of the following: the gender of the streamer or viewer of the streamed video; the age of the streamer or viewer of the streamed video; the occupation of the streamer or viewer of the streamed video; the family structure of the streamer or viewer of the streamed video; the annual income of the streamer or viewer of the streamed video; and the region of the streamer or viewer of the streamed video. Of course, these can be used individually or in combination of two or more.

[0053] A different streamer (u3) from the streamer of the video being streamed can be used when providing services to streamers. In other words, any streamer may collaborate with other streamers who have similar interests to stream videos. Therefore, this data is useful when recommending other streamers as collaboration partners to any given streamer. Furthermore, even if the streamers' interests are not similar, collaboration may be beneficial if the viewer attributes are similar, so this data is also useful in cases where the viewer attributes are similar.

[0054] Note that the input data (streaming video-related data d1) and the output data (streaming video estimated data d2) may be identical in content. For example, the input data (streaming video-related data d1) may be the MAU of a certain streamer, and the output data (streaming video estimated data d2) may be the estimated MAU of the same streamer. In this case, the MAU used as input data (streaming video-related data d1) would be, for example, past (factual) data, while the estimated MAU used as output data (streaming video estimated data d2) would be an estimate of future MAU.

[0055] 2-2-2. Second estimated data d22 of the streamed video estimated data d2 The second estimated data set, d22, includes psychographic data of at least one of the broadcaster and / or viewers. This psychographic data may include hobbies, favorite movie / video / game genres, favorite music genres, preferred design elements related to streaming videos (e.g., color schemes), preferred type of person, favorite food, favorite brands, favorite video streaming platforms, favorite animals, skills to learn, and media consumption data. These can be used individually or in combination of two or more.

[0056] 2-3. Data presentation unit 123 The data presentation unit 123 is configured to present the data to the recipient. The data presentation unit 123 is configured to present the data, for example, through the output unit 13 (display, etc.). The recipient can be various people, but for example, it could be the administrator or user of the information processing device 1. In addition, if the functions of the information processing device 1 are used by, for example, a broadcaster, a person belonging to a company, a talent agency, etc., these people would also be included. The presented data is recommendation data based on the distribution video estimation data d2. The presented data may be the distribution video estimation data d2 itself, or it may be data that has been processed from the distribution video estimation data d2 in a way that is easy for the recipient to understand.

[0057] 3. Explanation of Information Processing Methods The information processing (information processing method) according to this embodiment can be divided into a learning model creation stage and a learning model operation stage.

[0058] In the model creation phase, training data is prepared, including input data (corresponding to streaming video-related data d1) and corresponding output data (corresponding to streaming video estimation data d2). The necessary data can be collected via APIs, databases, or external servers. Depending on the situation, preprocessing (data cleansing) can be performed to remove missing data. The collected data is then assigned corresponding labels as output data. Furthermore, data normalization and standardization can be performed to unify data of different scales, thereby improving the learning efficiency of the model. The prepared training data can then be split, for example, into training data and test data, which can be used for training and evaluating the model.

[0059] Next, a learning model is created using an algorithm suitable for the data characteristics. The architecture of the learning model is not limited, and an example is explained in Section 2-2, so details are omitted here. The training data is supplied to the learning model as input, and the error between the output result and the correct label is calculated using a loss function. The value of this loss function is then minimized. As a result, the weight coefficients of the learning model are updated sequentially, and learning is completed when, for example, the value of the loss function falls below a predetermined threshold. After that, the performance of the learning model is evaluated using test data.

[0060] In the operational phase, the information processing (information processing method) comprises a data acquisition step, a data estimation step, and a data presentation step. In the data acquisition step, the data acquisition unit 121 acquires the streaming video-related data d1. In the data estimation step, the data estimation unit 122 generates the estimated streaming video data d2. In the data presentation step, the data presentation unit 123 presents the data to be presented. Regarding the streaming video-related data d1, it is as explained in sections 2-1-1 and 2-1-2, and regarding the streaming video estimation data d2, it is as explained in sections 2-2-1 and 2-2-2. The presented data is as explained in section 2-3, so a detailed explanation will be omitted. The generated streaming video estimation data d2 and presentation data can be used, for example, as suggestions to the recipient (e.g., improvement suggestions to the streamer). Furthermore, the estimation accuracy of the learning model can be continuously monitored, and if the estimation results fall below a certain level of accuracy, the model can be retrained using newly collected data to maintain its performance. This redeploys the learning model, and the streaming video estimation data d2 is generated using the latest weight coefficients. In the data estimation step, the data estimation unit 122 may generate a confidence level, which is an indicator of how reliable the estimation results generated by the learning model are. The confidence level may be, for example, a score, a probability value, or something other than a numerical value such as high, medium, or low confidence level. Here, the confidence level can be estimated, for example, using a distance function. Specifically, the confidence level of the output streaming video estimation data d2 can be generated by applying a distance function between the streaming video estimation data d2 and the training data (group) used for training the learning model. In other words, the data estimation unit 122 can further generate the confidence level of the generated streaming video estimation data d2 using the learning model. The distance function is a function used to calculate distance function values, such as similarity, between data points. The calculation method is not particularly limited, but for example, cosine similarity can be used. The distance function value calculated by the distance function is not limited to similarity; for example, it may be a quantity corresponding to the difference between data points (the value of the Euclidean distance), and various values ​​can be used. The data estimation unit 122 can then generate a confidence score based on the distance function value. Here, various evaluation methods can be used to generate the confidence score based on the distance function value. For example, the confidence score may be calculated according to whether the distance function value falls within a specific threshold range, or according to the magnitude of the distance function value. The confidence score generated by the data estimation unit 122 can then be presented as presented data by the data presentation unit 123. For example, it can be presented in a format that is easy for the user to understand, such as "The confidence score of this streaming video estimation data d2 is high (confidence score: 90%)."

[0061] 4. Examples of the use of Information Processing System 100 Next, we will describe an example of how the embodiment can be used (an example of an embodiment). Figure 6 is an explanatory diagram of an example of how the information processing device 1 can be used when providing useful information to any company. In Figure 6, the input data (distributed video-related data d1) is the distributed video data of each distributor as second related data d12, and the output data (distributed video estimation data d2) is the psychographic data of each distributor's viewers (second estimation data d22). The input data (distributed video-related data d1) includes, for example, the audio of the distributed video, comments on the distributed video, the date and time the comments were made, etc. As described in "2-1. Data acquisition unit 121", this data can be acquired using a generation AI model. In this application example, psychographic data of each broadcaster's viewers can be generated. The user of the information processing device 1 can use this psychographic data to understand the interests of each broadcaster's fans, etc. Furthermore, the user of the information processing device 1 can use this psychographic data to infer the attributes of products and services that the broadcaster's fans, etc., are interested in. For example, if a broadcaster talks about sports during a broadcast, psychographic data indicating that viewers are interested in sports will be generated, and the user of the information processing device 1 can infer that the viewers of this broadcast video are interested in such things. Also, if the content of the viewers' comments is a discussion about movies, psychographic data indicating that they are interested in movies will be generated, and the user of the information processing device 1 can infer that the viewers of this broadcast video are interested in such things. In addition, since the time of the topics the broadcaster talks about can be estimated, time-related elements can be added to the psychographic data, which can lead to inferring the interests of viewers, etc. As a result, the user of the information processing device 1 can recommend influential broadcasters to companies such as advertisers and advertising agencies. Furthermore, the user of the information processing device 1 may be a company such as an advertiser or an advertising agency.

[0062] Figure 7 is an explanatory diagram of an example 2 of the use of the information processing device 1 when providing useful information to any company. In Figure 7, the input data (distributed video related data d1) is the distributed video data of each distributor as the second related data d12, and the output data (distributed video estimation data d2) is the psychographic data of each distributor (second estimation data d22). In this application example, psychographic data can be generated for each broadcaster. The user of the information processing device 1 can use this psychographic data to understand each broadcaster's interests, etc., and can also use this psychographic data to infer the attributes of the broadcaster. As a result, the user of the information processing device 1 can recommend promising broadcasters to companies such as talent agencies. In this application example, the user of the information processing device 1 may be a company such as a talent agency.

[0063] Figure 8 is an explanatory diagram of an example of how the information processing device 1 can be used when providing useful information to each distributor. In Figure 8, the input data (streaming video related data d1) is the behavioral history data d113 of each streamer as the first related data d11, and the output data (streaming video estimation data d2) is the psychographic data (second estimation data d22) of each streamer (themselves). The user of the information processing device 1 can use each broadcaster's psychographic data to make the broadcasters re-recognize their interests, etc., and the user of the information processing device 1 can also use this psychographic data to infer and suggest the attributes of products and services that the broadcasters are interested in. The user of the information processing device 1 may also be one of the broadcasters.

[0064] Figure 9 is an explanatory diagram of an example of how the information processing device 1 can be used when providing useful information to each distributor. In Figure 9, the input data (distributed video-related data d1) is the viewer behavior history data d113 of each distributor, which is the first related data d11, and the output data (distributed video estimation data d2) is a different distributor from the one in question (first estimation data d21). In this example, it is possible to generate other streamers who can be collaboration partners for each streamer, and the user of the information processing device 1 can use this generated data to recommend suitable collaborators to each streamer. In this example, the user of the information processing device 1 may also be the streamer themselves who is looking for a collaboration partner.

[0065] Figures 10A and 10B are explanatory diagrams for example 3 of the use of the information processing device 1 when providing useful information to each distributor. Figures 10A and 10B show an example of use in which the necessary data is obtained through multiple stages (two stages in this case) of estimation. In Figure 10A, the input data (streaming video related data d1) is the behavioral history data d113 of each streamer's viewers, which is the first related data d11, and the output data (streaming video estimation data d2) is a different streamer (first estimation data d21). This means that other streamers that a viewer is interested in are estimated from the behavioral history (e.g., viewing history) of each streamer's viewers. In Figure 10B, the input data (streaming video-related data d1) is the estimated viewer behavior history data d113 of each other streamer, and the output data (streaming video estimation data d2) is the psychographic data of this viewer. This allows the streamer to estimate the interests of viewers of other streamers whose viewers are interested in their own video streams. This enables the streamer to understand the audience demographics of other streamers they consider rivals, and they can use this information to decide how to proceed with the production of their own video streams in the future. In this example, the user of the information processing device 1 may be the streamer themselves or not.

[0066] Figure 10C is a modified version of Figures 10A and 10B. Figures 10A and 10B illustrate an example in which the necessary data (psychographic data of each other's viewers) is generated step by step using separate learning models, but the method is not limited to this. As shown in Figure 10C, the learning model may be trained to output the necessary data all at once.

[0067] Figure 11 is an explanatory diagram of an example 4 of how the information processing device 1 can be used when providing useful information to each distributor. In Figure 11, the input data (streaming video related data d1) is the behavioral history data d113 of other streamers that each streamer is interested in, which is the first related data d11, and the output data (streaming video estimation data d2) is the psychographic data of the other streamer (second estimation data d22). The user of the information processing device 1 can use the psychographic data of other broadcasters to provide each broadcaster with data on the interests and concerns of other broadcasters. In other words, since it can be inferred that a broadcaster's own interests and concerns are related to the interests and concerns of other broadcasters that the broadcaster is interested in, providing the broadcaster with psychographic data of other broadcasters makes it possible to provide the broadcaster with, for example, effective video broadcasting plan proposals. In this example of use, the user of the information processing device 1 may be the broadcaster themselves or not.

[0068] Various embodiments are illustrated below. The embodiments shown below can be combined with each other. [Note 1] An information processing system that uses data related to streamed videos, It comprises a data acquisition unit and a data estimation unit. The data acquisition unit is configured to acquire the data related to the distributed video, The aforementioned data related to the distributed video includes the first related data, The aforementioned first related data includes at least one of statistical data, distributor characteristics data, and behavioral history data. The aforementioned statistical data includes at least one of the following: data relating to the number of viewers of the streamed video, and data relating to the streaming time of the streamed video. The aforementioned distributor characteristics data is data that shows the characteristics of the distributor of the distributed video, The aforementioned behavioral history data is data that shows the behavioral history of the broadcaster or viewer, The data estimation unit is configured to generate streaming video estimation data based on the streaming video-related data. The aforementioned streaming video estimation data includes the first estimation data, The first estimated data is an information processing system that includes data relating to the performance of the distributor or the distributed video, demographic data of the viewers, and at least one of the distributors other than the distributor of the distributed video. [Note 2] The information processing system described in Appendix 1, The aforementioned streaming video-related data further includes the second related data, The second related data includes an information processing system comprising at least one of the following: the distributed video data and the comment data of the distributed video. [Note 3] An information processing system as described in Appendix 1 or Appendix 2, The aforementioned estimated data for the streamed video further includes second estimated data, The second estimated data includes psychographic data of at least one of the broadcaster and the viewer in an information processing system. [Note 4] An information processing system described in any one of the appendices 1 to 3, It also has a memory unit, An information processing system in which the aforementioned storage unit stores the data related to the distributed video. [Note 5] An information processing system described in any one of the appendices 1 to 4, It further includes a data presentation unit, The data presentation unit is configured to present the data to the recipient. The aforementioned presented data is recommendation data based on the aforementioned streaming video estimation data, according to the information processing system. [Note 6] An information processing system described in any one of the appendices 1 to 5, The data estimation unit is an information processing system that generates the estimated streaming video data based on a learning model that takes the streamed video-related data as input and outputs the streamed video estimation data. [Note 7] The information processing system described in Appendix 6, The data estimation unit is an information processing system that uses the learning model to further generate a confidence level for the generated streaming video estimation data. [Note 8] An information processing system described in any one of the appendices 1 to 7, The data acquisition unit is configured to acquire the statistical data as API-linked data from an external information processing device of the information processing system, in the information processing system. [Note 9] An information processing system that uses data related to streamed videos, It comprises a data acquisition unit and a data estimation unit. The data acquisition unit is configured to acquire the data related to the distributed video, The aforementioned data related to the distributed video includes the second related data, The aforementioned second related data includes at least one of the streamed video data and the streamed video comment data. The data estimation unit is configured to generate streaming video estimation data based on the streaming video-related data. The aforementioned streaming video estimation data includes the second estimation data, The second estimated data includes psychographic data of at least one of the broadcaster and the viewer, in an information processing system. [Note 10] A method for processing information using data related to streamed videos, The system comprises a data acquisition step and a data estimation step. In the data acquisition step described above, the data related to the distributed video is acquired. The aforementioned data related to the distributed video includes the first related data, The aforementioned first related data includes at least one of statistical data, distributor characteristics data, and behavioral history data. The aforementioned statistical data includes at least one of the following: data relating to the number of viewers of the streamed video, and data relating to the streaming time of the streamed video. The aforementioned distributor characteristics data is data that shows the characteristics of the distributor of the distributed video, The aforementioned behavioral history data is data that shows the behavioral history of the broadcaster or viewer, In the data estimation step described above, estimated streaming video data is generated based on the streaming video-related data, The aforementioned streaming video estimation data includes the first estimation data, An information processing method wherein the first estimated data includes at least one of the following: data relating to the performance of the distributor or the distributed video; demographic data of the viewer; and a distributor different from the distributor of the distributed video. [Note 11] A method for processing information using data related to streamed videos, The system comprises a data acquisition step and a data estimation step. In the data acquisition step described above, the data related to the distributed video is acquired. The aforementioned data related to the distributed video includes the second related data, The aforementioned second related data includes at least one of the streamed video data and the streamed video comment data. In the data estimation step described above, estimated streaming video data is generated based on the streaming video-related data, The aforementioned streaming video estimation data includes the second estimation data, The second estimated data includes psychographic data of at least one of the broadcaster and the viewer, as part of an information processing method. [Note 12] A program that causes a computer to execute the information processing method described in Appendix 10 or Appendix 11. [Explanation of symbols]

[0069] 100: Information Processing Systems 1: Information Processing Device 10: Communications Department 11: Storage section 12: Control Unit 121: Data acquisition unit 122: Data Estimation Unit 122N: Neural Network 123: Data Presentation Section 13: Output section 14: Input section 15: Communications bus 2: Communication Network d1: Data related to streaming videos d11: First related data d111: Statistical data t1: Data t2: Data d112: Streamer characteristics data s1: Platform type data s2: Streamer analytics data s3: Streamer data d113: Behavioral history data d12: Second related data r1: Streamer video data r2: Comment data d2: Estimated data for streaming video d21: First estimated data u1: Data u2: Demographic data u3: Streamer d22: Second estimated data

Claims

1. An information processing system that uses data related to streamed videos, It comprises a data acquisition unit and a data estimation unit. The data acquisition unit is configured to acquire the data related to the distributed video, The aforementioned streaming video-related data includes the first related data, The first related data includes at least one of statistical data, distributor characteristic data, and behavioral history data. The aforementioned statistical data includes at least one of the following: data relating to the number of viewers of the streamed video, and data relating to the streaming time of the streamed video. The aforementioned distributor characteristics data is data that shows the characteristics of the distributor of the distributed video, The aforementioned behavioral history data is data that shows the behavioral history of the broadcaster or viewer, The data estimation unit is configured to generate streaming video estimation data based on the streaming video-related data. The aforementioned streaming video estimation data includes the first estimation data, The first estimated data is an information processing system that includes data relating to the performance of the distributor or the distributed video, demographic data of the viewers, and at least one of the distributors other than the distributor of the distributed video.

2. The information processing system according to claim 1, The aforementioned streaming video-related data further includes the second related data, The second related data includes an information processing system comprising at least one of the following: the distributed video data and the comment data of the distributed video.

3. The information processing system according to claim 2, The aforementioned estimated data for the streamed video further includes second estimated data, The second estimated data includes psychographic data of at least one of the broadcaster and the viewer in an information processing system.

4. An information processing system according to any one of claims 1 to 3, It also has a memory unit, An information processing system in which the aforementioned storage unit stores the data related to the distributed video.

5. An information processing system according to any one of claims 1 to 3, It further includes a data presentation unit, The data presentation unit is configured to present the data to the recipient. The aforementioned presented data is recommendation data based on the aforementioned streaming video estimation data, according to the information processing system.

6. An information processing system according to any one of claims 1 to 3, The data estimation unit is an information processing system that generates the estimated streaming video data based on a learning model that takes the streamed video-related data as input and outputs the streamed video estimation data.

7. The information processing system according to claim 6, The data estimation unit is an information processing system that uses the learning model to further generate a confidence level for the generated streaming video estimation data.

8. An information processing system according to any one of claims 1 to 3, The data acquisition unit is configured to acquire the statistical data as API-linked data from an external information processing device of the information processing system, in the information processing system.

9. An information processing system that uses data related to streamed videos, It comprises a data acquisition unit and a data estimation unit. The data acquisition unit is configured to acquire the data related to the distributed video, The aforementioned data related to the distributed video includes the second related data, The second related data includes at least one of the streamed video data and the streamed video comment data. The data estimation unit is configured to generate streaming video estimation data based on the streaming video-related data. The aforementioned estimated data for the streamed video includes the second estimated data, The second estimated data includes an information processing system that includes psychographic data of at least one of the broadcaster and the viewer.

10. A method for processing information using data related to streamed videos, The system comprises a data acquisition step and a data estimation step. In the data acquisition step described above, the data related to the distributed video is acquired. The aforementioned streaming video-related data includes the first related data, The first related data includes at least one of statistical data, distributor characteristic data, and behavioral history data. The aforementioned statistical data includes at least one of the following: data relating to the number of viewers of the streamed video, and data relating to the streaming time of the streamed video. The aforementioned distributor characteristics data is data that shows the characteristics of the distributor of the distributed video, The aforementioned behavioral history data is data that shows the behavioral history of the broadcaster or viewer, In the data estimation step described above, estimated streaming video data is generated based on the streaming video-related data, The aforementioned streaming video estimation data includes the first estimation data, An information processing method wherein the first estimated data includes at least one of the following: data relating to the performance of the distributor or the distributed video; demographic data of the viewer; and a distributor different from the distributor of the distributed video.

11. A method for processing information using data related to streamed videos, The system comprises a data acquisition step and a data estimation step. In the data acquisition step described above, the data related to the distributed video is acquired. The aforementioned data related to the distributed video includes the second related data, The second related data includes at least one of the streamed video data and the streamed video comment data. In the data estimation step described above, estimated streaming video data is generated based on the streaming video-related data, The aforementioned estimated data for the streamed video includes the second estimated data, The second estimated data includes psychographic data of at least one of the broadcaster and the viewer, as part of an information processing method.

12. A program that causes a computer to execute the information processing method described in claim 10 or claim 11.

Images