Estimation device
The estimation device uses dual learning models to accurately predict user behavior across services, addressing data limitations and bias, enabling precise user referrals.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- NTT DOCOMO INC
- Filing Date
- 2024-05-23
- Publication Date
- 2026-05-20
AI Technical Summary
Existing systems struggle to accurately identify users who are likely to transition from one service to another, as the data for common users between services is often limited, leading to biased estimations.
An estimation device that constructs a first estimation model based on a first database of a user group's history and a second estimation model based on a second database of common users, using deep learning techniques, to accurately predict user behavior and target referrals.
The device effectively reduces bias in estimations by leveraging a large first user group database, enabling precise identification of users likely to engage in specific actions, such as purchasing a popular appliance, for referral to a second service.
Smart Images
Figure 0007863262000001 
Figure 0007863262000002 
Figure 0007863262000003
Abstract
Description
Technical Field
[0001] This disclosure relates to an estimation device.
Background Art
[0002] Patent Document 1 discloses a system for escorting users between stores. This system reads the identification information of a user from an IC (Integrated Circuit) chip possessed by the user, and associates guidance information to another store with the read identification information.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] For example, a system for mutually escorting users who use respective services between different services can be considered. When escorting a user of a first service to a second service, it is required to escort a user who is highly likely to use the second service from among the users of the first service. However, among the users of the first service, it is unclear which users are highly likely to use the second service.
[0005] This disclosure has been made in view of the above, and an object thereof is to provide an estimation device that can accurately extract a target user.
Means for Solving the Problems
[0006] [[ID=四十五]] The estimation device relating to this disclosure comprises: a first learning unit that constructs a first estimation model for estimating the first actions of a first user group based on a first database having first data including a history of the first actions of each user constituting a first user group; a second learning unit that constructs a second estimation model for estimating the second actions of common users based on the first estimation model, using second data for common users in a second database having second data including a history of the second actions of each user constituting a second user group that includes at least a common user group common to a part of the first user group, and the first data for common users in the first database; and an estimation unit that estimates the second actions of the first user group excluding the common user group using the second estimation model.
[0007] In the estimation device described above, when a first user group uses the first service and a second user group uses the second service, the device can estimate which users from the first user group are most likely to perform the second action, thereby targeting those users for referral to the second service. In this device, actions are estimated by a second estimation model generated based on a second database containing the history of the second action, so the second action of the first user group can be estimated with high accuracy. Therefore, users who are eligible for referral can be extracted with high accuracy. However, if the estimation model is generated based only on data of common users, there is a possibility of bias in the estimation model, especially when the number of common users is small. In the above device, the impact of bias from common users is reduced because the second estimation model is based on the first estimation model generated based on the first database. [Effects of the Invention]
[0008] According to this disclosure, it is possible to accurately extract users who are the target of customer referral. [Brief explanation of the drawing]
[0009] [Figure 1] Figure 1 shows the configuration of an example estimation device according to an embodiment of this disclosure. [Figure 2]Figure 2 is a diagram illustrating the contents of a first example database. [Figure 3] Figure 3 is a diagram illustrating the contents of a second database as an example. [Figure 4] Figure 4 schematically shows one example of a first estimation model. [Figure 5] Figure 5 is a diagram illustrating the contents of an example of integrated data. [Figure 6] Figure 6 schematically shows one example of a second estimation model. [Figure 7] Figure 7 is a flowchart showing an example of the operation of the estimation device. [Figure 8] Figure 8 shows the hardware configuration of an example estimation device. [Modes for carrying out the invention]
[0010] The embodiments of the estimation device according to this disclosure will be described in detail below with reference to the drawings. In the description of the drawings, the same elements will be denoted by the same reference numerals, and redundant descriptions will be omitted.
[0011] Figure 1 shows the configuration of an example estimation device 10. This example estimation device 10 can be used in a system to refer users to each other's services, which are different services. Below, we will describe a form in which users are referred to the second service from a group of users of the first service. The first service and the second service may be services operated by different companies. In this embodiment, we will describe an example in which the estimation device 10 is used by the company operating the first service, with the company operating the second service being the company operating the second service as the partner company. In this example, the first service and the second service may be services in which points are awarded to users in response to purchases of goods, etc., by the user, but the content of the first service and the second service is not limited to this. The group of users who use the first service (first user group) and the group of users who use the second service (second user group) have some common users. That is, the first user group includes users who also use the second service (common users). The estimation device 10 is a device for estimating users from the first user group who are likely to use the second service. In the following embodiment, we will describe an example of referring users from the first user group who are highly likely to purchase popular home appliance A to the second service.
[0012] As shown in Figure 1, an example estimation device 10 includes an input unit 11, a preprocessing unit 12, a first learning unit 13, a second learning unit 15, an estimation unit 16, and an output unit 17. The input unit 11 includes a first data input unit 11A and a second data input unit 11B. The contents of a first database containing data (first data) of a first group of users who use the first service are input to the first data input unit 11A. The first data includes a history of predetermined actions (first actions) of each user constituting the first group of users. The first database may be managed by the company itself.
[0013] The second data input unit 11B receives data (second data) that includes the history of the actions of the common user group. The second data may be data extracted from a second database that includes the history of predetermined actions (second actions) of each user constituting the second user group, which includes at least the common users. The second database may be managed by a partner company. For example, the second data may be data of only the common users provided to the company by a partner company, and may be data extracted from a second database managed by the partner company.
[0014] Figure 2 shows the contents of the first database. Figure 3 shows the common user data in the second database. One example of the first database consists of a user ID (identifier) to identify the user and attribute data for each user. The user attribute data may include basic information and behavioral information. The basic information may include static characteristics such as the user's gender, age, and place of residence. The basic information may be based on contract information obtained from the user when they start using the first service.
[0015] Behavioral information includes dynamic characteristics resulting from user behavior, such as the amount of payments made by the user in relation to the first service. Behavioral information may include multiple behavioral histories with different characteristics. For example, behavioral information may include the usage history of an app that can access the first service, the usage history (visit history) of merchants affiliated with the first service, and the user's location information history. In the example in Figure 2, the behavioral information of the first data includes the visit history to a partner (mass retailer) that collaborates with the first service.
[0016] Furthermore, user behavior information may be obtained through an application running on the terminal device. The terminal device is, for example, a communication device operated by the user. The terminal device may be, for example, a mobile device such as a high-function mobile phone (smartphone), a mobile phone, or a personal digital assistant (PDA). Furthermore, the devices constituting the terminal device are not interpreted restrictively.
[0017] The second data of one example is composed of a common user ID for identifying a common user and data about the attributes of each user. The common user ID may have the same identification code as the user ID of the first data. Note that the user ID (common user ID) may be information that can identify an individual, such as an email address or a mobile phone number. In this case, users whose user IDs included in each of the first database and the second database are common can be extracted as common users. In the example shown in FIG. 3, users with user IDs 2, 3, 4, and 7 among the first user group constituting the first database are exemplified as common users.
[0018] In the second data, the attributes of the user may include basic information and behavioral information. The basic information includes static features such as the user's residential area. Note that in addition to the residential area, the basic information may include gender and the like similar to the first data. In this example, the basic information that overlaps with the first data included in the second database may be deleted before being input to the second data input unit 11B. The behavioral information includes dynamic features resulting from the user's behavior, such as the user's membership rank, usage points, and favorite stores in the second service. In this example, the purchase history of popular home appliance A in the franchise store of the second service is included as the behavioral attribute of the second data.
[0019] The preprocessing unit 12 performs a process of converting the first data and the second data input to the input unit 11 into a format that can be processed by the first learning unit 13 and the second learning unit 15. In one example, the preprocessing unit 12 may convert the string data included in the first data and the second data into numerical data.
[0020] The first learning unit 13 constructs a first estimation model using the first data processed by the preprocessing unit 12 as training data. For example, the first learning unit 13 constructs the first estimation model using deep learning techniques with a neural network. Figure 4 is a schematic diagram showing an example of a first estimation model. The first estimation model 20 may be composed of a neural network consisting of an input layer 21, a plurality of hidden layers 22, and an output layer 23. As shown in Figure 4, the input layer 21 of an example of the first estimation model 20 is input with data indicating basic information, usage history of the company's applications, usage history of partner services, and movement history of terminal devices (e.g., user location analysis information, user area information, etc.). The first estimation model 20 constructed by the first learning unit 13 is stored in a predetermined storage unit.
[0021] The second learning unit 15 constructs a second estimation model 30 (see Figure 6) that estimates the second behavior of common users based on the first data for common user groups in the first database and the second data for common user groups in the second database. The second estimation model 30 is a learning model based on the first estimation model 20 and inherits the features of the hidden layer 32 of the first estimation model 20.
[0022] The learning data used in the second learning unit 15 may be integrated data (integrated data) that combines the first data and the second data for a common user group. Figure 5 is a diagram illustrating the contents of the integrated data. One example of integrated data consists of user IDs and data about the attributes of each user. The user attributes include basic information and behavioral information from the first data and basic information and behavioral information from the second data. Therefore, the data columns of the integrated data are greater than those of the first data. In Figure 5, for ease of understanding, the data about common users from the first data shown in Figure 2 and the second data shown in Figure 3 are integrated as is, but for each data, data processed by the preprocessing unit 12 may be used.
[0023] In this embodiment, a first estimation model 20 is constructed by the first learning unit 13 to estimate a first behavior based on data from a large-scale first user group owned by the company. Furthermore, a second estimation model is constructed by the second learning unit 15 to estimate a second behavior based on data from common users owned by the company and data from common users provided by partner companies. The user behavior ultimately estimated by the estimation device 10 is the second behavior estimated by the second estimation model.
[0024] The second behavior may be a behavior related to the first behavior estimated by the first estimation model 20. For example, a behavior related to the first behavior may be a behavior corresponding to a sub-concept of the first behavior, a behavior that has a common superordinate concept with the first behavior, or a behavior corresponding to a superordinate concept of the first behavior. In this case, there is a certain correlation between the first behavior and the second behavior, and it is considered that users who take the first behavior are likely to take the second behavior. In this embodiment, the first learning unit 13 constructs a first estimation model 20 that estimates the generalized behavior (first behavior), and the second learning unit 15 constructs a second estimation model 30 that estimates the specialized behavior (second behavior). The specialized behavior corresponds to a sub-concept of the generalized behavior. Note that the sub-concept includes not only behaviors that are more concrete than the generalized behavior which is the superordinate concept, but also behaviors that can be taken following the generalized behavior in relation to the generalized behavior which is the superordinate concept.
[0025] In this embodiment, the users targeted for referral to the second service are assumed to be users who are highly likely to purchase popular home appliance A. Therefore, the second estimation model 30 constructed by the second learning unit 15 is a learning model that estimates users who are highly likely to purchase popular home appliance A. Furthermore, the specialized behavior estimated by the second estimation model 30 is the purchase of popular home appliance A. In this case, the generalized behavior estimated by the first estimation model 20 may be a higher-level concept than "purchase of popular home appliance A." In this embodiment, the generalized behavior estimated by the first estimation model 20 may be "visiting a mass retailer." That is, the first estimation model 20 constructed by the first learning unit 13 may be a learning model that estimates users who are likely to visit a mass retailer. Note that the data for the specialized behavior estimated by the second estimation model 30 is stored only in the second database and not in the first database. Similarly, the data for the generalized behavior estimated by the first estimation model 20 is stored only in the first database and not in the second database.
[0026] The first learning unit 13 uses the first data from the first database as training data to train the first estimation model 20. Specifically, when the first learning unit 13 inputs basic information constituting the first data, usage history of the company's app, usage history of the company's partner services, and movement history of terminal devices into the input layer 21, it learns the weight data of the hidden layer 22 so that users with a history of visiting mass retailers are output from the output layer 23.
[0027] The second learning unit 15 uses the integrated data as training data to train the second estimation model 30. Figure 6 is a schematic diagram showing an example of the second estimation model 30. The second estimation model 30 is composed of a neural network consisting of an input layer 31, multiple hidden layers 32, and an output layer 33, and is based on the first estimation model 20. Specifically, the input layer 31 has nodes 31a common to the first estimation model 20, into which basic information, usage history of the company's applications, usage history of the company's partner services, and movement history of terminal devices are input, and nodes 31b specific to the first estimation model 20, into which behavioral information derived from the second database is input. Similarly, the hidden layer 32 has nodes inherited from the first estimation model 20, including nodes 32a corresponding to each node 31a, and nodes 32b corresponding to nodes 31b into which the second data is input.
[0028] The second learning unit 15 learns the weight data of the hidden layer 32 so that when it receives basic information constituting the integrated data, usage history of the company's app, usage history of the company's partner services, movement history of terminal devices, and behavior history of the second service as input to the input layer, users with a purchase history of popular home appliance A are output from the output layer 33. More specifically, the second learning unit 15 fine-tunes the weight data of the hidden layer 22 inherited from the first estimation model 20. For example, the second learning unit 15 may adjust the weight data using a so-called fine-tuning method. Furthermore, the second learning unit 15 adds a node 32b corresponding to the second data to node 32a, whose weight data has been fine-tuned and inherited from the first estimation model 20, and learns the weight data related to the added node 32b. For example, the second learning unit 15 may learn the weight data related to the added node 32b using a so-called transfer learning method.
[0029] The estimation unit 16 uses the second estimation model 30 to estimate the specialized behavior of the first user group, excluding the common user group. That is, the estimation unit 16 uses the second estimation model 30 to estimate users who are likely to purchase popular home appliance A, using the first data of the first user group, excluding the common user group, as input data. The input layer 31 of the second estimation model 30 may contain the behavioral history of the second service based on the second data. The estimation unit 16 receives basic information based on the first data, usage history of the company's own app, usage history of the company's partner services, and terminal device movement history as substantial input data. Input of the behavioral history of the second service based on the second data may be omitted.
[0030] The output unit 17 outputs multiple users who are estimated by the estimation unit 16 to be highly likely to purchase popular home appliance A, in descending order of likelihood. The output user data can be used to refer customers from the company to partner companies.
[0031] Figure 7 is a flowchart showing an example of the operation of the estimation device 10. First, in the estimation device 10, the first data constituting the first data table and the second data (common data) constituting the second data table are input to the input unit 11 (step S1). The second data table may include users other than common users, but here it is assumed to consist only of common users. The first data and second data input to the input unit 11 are preprocessed by the preprocessing unit 12 (step S2). Next, the first learning unit 13 constructs a first estimation model 20 that estimates the generalized behavior of the first user group based on the preprocessed first data (step S3). Subsequently, the second learning unit 15 constructs a second estimation model that estimates the specialized behavior of the common user group based on the first estimation model 20 and integrated data obtained by integrating the first data and second data for the common user group (step S4). Next, using the second estimation model, the estimation unit 16 estimates which users are likely to take specialized actions from the first user group excluding the common user group, and the first users with a high probability of taking specialized actions are output as target users for customer referral (step S5).
[0032] As described above, the estimation device 10 in one example comprises: a first learning unit 13 that constructs a first estimation model 20 for estimating the first actions of the first user group based on a first database having first data including the history of the first actions of each user constituting the first user group; a second learning unit 15 that constructs a second estimation model 30 for estimating the second actions of common users based on the first estimation model 20, using second data for the common user group in a second database having second data including the history of the second actions of each user constituting the second user group which includes at least a common user group that is common to a part of the first user group, and first data for the common user group in the first database; and an estimation unit 16 that estimates the second actions of the first user group excluding the common user group using the second estimation model 30.
[0033] In the estimation device 10 described above, when the first user group uses the first service and the second user group uses the second service, the device estimates which users from the first user group are most likely to perform the second action. This allows the estimated users to be targeted for referral to the second service. In this device, the behavior is estimated by a second estimation model generated based on a second database containing the history of the second action, so the second action of the first user group can be estimated with high accuracy. Therefore, users who are targeted for referral can be extracted with high accuracy.
[0034] Furthermore, as in this embodiment, if there are overlapping common users between the first and second databases, it is conceivable to generate an estimation model that estimates users likely to take the second action based only on data from these common users. In other words, it is conceivable to estimate users who are likely to take the second action by constructing only the second estimation model, without relying on the first estimation model 20. However, the number of common users who utilize both the first and second services is often relatively small, making it difficult to prepare sufficient data to construct a highly accurate estimation model. When the number of common users is small, the estimation model may be biased. For example, if users who utilize both the first and second services, as in this embodiment, are more active (proactive, assertive) than other users, an estimation model constructed based on such common users is likely to extract active users as target users. In this case, regardless of what action is set as the second action, users with similar attributes (i.e., active users) will be extracted from the first user group. In the estimation device 10 according to this embodiment, a second estimation model 30 is constructed based on a second data set of a small common user group, using a first estimation model 20 generated based on a first database having a large first user group as the foundation. This reduces the impact of bias in common users in the second estimation model 30.
[0035] The second behavior in the example may be a behavior related to the first behavior. In particular, the second behavior may be a behavior that corresponds to a sub-concept of the first behavior. In this case, it is easier to construct a second estimation model 30 based on the first estimation model 20.
[0036] One example of a first estimation model 20 may be a model that outputs users who are estimated to perform a first action when first data contained in the first database is input. A second estimation model 30 may be a model that outputs users who are estimated to perform a second action when at least first data is input. In this case, the second action can be estimated by inputting the first data into the second estimation model 30 without making any special modifications to the contents of the first database. Therefore, the second actions of the first user group other than common users can be easily estimated.
[0037] Although one embodiment has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment. For example, an example of constructing a learning model using deep learning techniques with neural networks has been described, but other machine learning techniques can be used as long as they can construct a second estimation model based on the first estimation model.
[0038] The block diagram used in the description of the above embodiment shows functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or it may be realized using two or more physically or logically separated devices that are directly or indirectly connected (for example, using wired or wireless connections). A functional block may be realized by combining the above one device or the above multiple devices with software.
[0039] Functions include, but are not limited to, judgment, decision, judgment, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, assumption, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), and assigning. For example, a functional block (configuration part) that enables transmission is called a transmitting unit or transmitter. As mentioned above, the method of implementation is not particularly limited.
[0040] For example, the estimation device 10 in one embodiment of the present disclosure may function as a computer that performs the estimation processing of the present disclosure. Figure 8 is a diagram showing an example of the hardware configuration of the estimation device 10 according to one embodiment of the present disclosure. The estimation device 10 described above may be physically configured as a computer device including a processor 1001, memory 1002, storage 1003, communication device 1004, input device 1005, output device 1006, bus 1007, etc.
[0041] In the following explanation, the term "device" can be replaced with "circuit," "device," "unit," etc. The hardware configuration of the input unit 11, preprocessing unit 12, first learning unit 13, second learning unit 15, estimation unit 16, and output unit 17 may include one or more of the devices shown in the figure, or it may be configured to omit some of the devices.
[0042] Each function in the estimation device is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, which allows the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of data reading and writing in the memory 1002 and storage 1003.
[0043] The processor 1001 controls the entire computer, for example, by running an operating system. The processor 1001 may be composed of a central processing unit (CPU) that includes interfaces with peripheral devices, control units, arithmetic units, registers, etc. For example, the preprocessing unit 12, first learning unit 13, second learning unit 15, estimation unit 16, etc., described above may be implemented by the processor 1001.
[0044] Furthermore, the processor 1001 reads programs (program code), software modules, data, etc., from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes accordingly. The program used is one that causes the computer to execute at least a part of the operations described in the above embodiment. For example, the pre-processing unit 12 may be implemented by a control program stored in the memory 1002 and running on the processor 1001, and other functional blocks may be implemented similarly. The above-described various processes have been explained as being executed by one processor 1001, but they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may also be transmitted from a network via a telecommunications line.
[0045] Memory 1002 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. Memory 1002 may also be called a register, cache, main memory, etc. Memory 1002 can store executable programs (program code), software modules, etc., for carrying out the estimation method according to one embodiment of the present disclosure.
[0046] Storage 1003 is a computer-readable recording medium and may consist of at least one of the following: an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disc, a digital multipurpose disc, a Blu-ray® disc), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. Storage 1003 may also be called an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, server, or other suitable medium including at least one of memory 1002 and storage 1003.
[0047] The communication device 1004 is hardware (transceiver / receiver device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc. The communication device 1004 may be configured to include high-frequency switches, duplexers, filters, frequency synthesizers, etc., in order to implement at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the input section 11 and output section 17 described above may be implemented by the communication device 1004.
[0048] The input device 1005 is an input device that accepts input from an external source (e.g., a keyboard, mouse, microphone, switch, button, sensor, etc.). The output device 1006 is an output device that outputs to an external source (e.g., a display, speaker, LED lamp, etc.). The input device 1005 and the output device 1006 may be configured as an integrated unit (e.g., a touch panel).
[0049] Furthermore, each device, such as the processor 1001 and memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or different buses may be configured for each device.
[0050] Furthermore, the estimation device may include hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), or an FPGA (Field Programmable Gate Array), and some or all of the functional blocks may be implemented by such hardware. For example, processor 1001 may be implemented using at least one of these hardware components.
[0051] Information notification is not limited to the embodiments described herein and may be carried out by other means. For example, information notification may be carried out by physical layer signaling (e.g., DCI (Downlink Control Information), UCI (Uplink Control Information)), upper layer signaling (e.g., RRC (Radio Resource Control) signaling, MAC (Medium Access Control) signaling, broadcast information (MIB (Master Information Block), SIB (System Information Block))), other signals, or combinations thereof. RRC signaling may also be called RRC messages, and may be, for example, RRC Connection Setup messages, RRC Connection Reconfiguration messages, etc.
[0052] The processing procedures, sequences, flowcharts, etc., of each aspect / embodiment described herein may be reordered, provided they are consistent with each other. For example, the methods described herein present various step elements in an exemplary order and are not limited to that specific order.
[0053] Input and output information may be stored in a specific location (e.g., memory) or managed using a management table. Input and output information may be overwritten, updated, or appended to. Output information may be deleted. Input information may be transmitted to other devices.
[0054] The determination may be made by a value represented by 1 bit (0 or 1), by a boolean value (true or false), or by a numerical comparison (for example, a comparison with a predetermined value).
[0055] Each aspect / embodiment described herein may be used individually, in combination, or switched between as needed during implementation. Furthermore, notification of specific information (e.g., notification that "X is") is not limited to explicit notification, but may also be implicit (e.g., by not providing such notification).
[0056] Although the present disclosure has been described in detail above, it will be clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the intent and scope of the present disclosure as defined by the claims. Accordingly, the descriptions in the present disclosure are illustrative and not intended to be restrictive in any way.
[0057] Software should be broadly interpreted to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, execution threads, procedures, functions, and so on, whether they are called software, firmware, middleware, microcode, hardware description languages, or by any other name.
[0058] Furthermore, software, instructions, information, etc., may be transmitted and received via a transmission medium. For example, if software is transmitted from a website, server, or other remote source using at least one of wired technologies (such as coaxial cable, fiber optic cable, twisted pair, or digital subscriber line (DSL)) and wireless technologies (such as infrared or microwave), then at least one of these wired and wireless technologies is included in the definition of a transmission medium.
[0059] The information, signals, etc. described in this disclosure may be represented using any of the various different techniques. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.
[0060] In addition, terms used in this disclosure and terms necessary for understanding this disclosure may be replaced with terms that have the same or similar meaning.
[0061] Furthermore, the information, parameters, etc., described in this disclosure may be expressed using absolute values, relative values from a given value, or corresponding other information. For example, wireless resources may be indicated by an index.
[0062] As used in this disclosure, the terms “determining” and “determining” may encompass a wide variety of actions. “Determining” may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, or inquiring (e.g., searching in a table, database, or other data structure), or ascertaining. “Determining” may also include, for example, receiving (e.g., receiving information), transmitting (e.g., sending information), inputting, outputting, or accessing (e.g., accessing data in memory). Furthermore, "judgment" and "decision" can include considering something as having been "judged" or "decided" after resolving, selecting, choosing, establishing, comparing, etc. In other words, "judgment" and "decision" can include considering something as having been "judged" or "decided" after some action. Also, "judgment (decision)" can be reinterpreted as "assuming," "expecting," or "considering."
[0063] The terms “connected,” “coupled,” or any variation thereof, mean any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are “connected” or “coupled” with each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, “connection” may be reinterpreted as “access.” As used in this disclosure, two elements may be considered to be “connected” or “coupled” with each other using at least one of one or more wires, cables, and printed electrical connections, and, in some non-limiting and non-exclusive examples, electromagnetic energy having wavelengths in the radio frequency domain, microwave domain, and optical (both visible and invisible) domain.
[0064] In this disclosure, the phrase "based on" does not mean "based solely on" unless otherwise specified. In other words, the phrase "based on" means both "based solely on" and "based at least on."
[0065] Any reference to elements using designations such as “first,” “second,” etc., as used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Accordingly, references to first and second elements do not imply that only two elements may be adopted, or that the first element must precede the second element in any way.
[0066] Where the terms “include,” “including,” and their variations are used in this disclosure, these terms are intended to be inclusive, as is the term “comprising.” Furthermore, the term “or” as used in this disclosure is not intended to mean exclusive OR.
[0067] In this disclosure, if articles are added by translation, such as a, an, and the in English, this disclosure may include the fact that the noun following these articles is plural.
[0068] In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combine" may be interpreted similarly to "different."
[0069] The estimation device of this disclosure has the following configuration. [1] A first learning unit constructs a first estimation model for estimating the first actions of a first user group based on a first database having first data including the history of the first actions of each user constituting the first user group, A second learning unit constructs a second estimation model for estimating the second actions of the common users based on the first estimation model, using the second data for the common users in a second database which has second data including the history of the second actions of each user constituting a second user group which includes at least a common user group that is common to a part of the first user group, and the first data for the common users in the first database. An estimation device comprising: an estimation unit that estimates the second behavior of the first user group excluding the common user group using the second estimation model. [2] The estimation apparatus described in [1], wherein the second action is an action related to the first action. [3] The estimation device described in [1] or [2], wherein the second action corresponds to a sub-concept of the first action. [4] The first estimation model is a model that outputs users who are estimated to perform the first action when the first data contained in the first database is input, The estimation device according to any one of [1] to [3], wherein the second estimation model is a model in which a user is estimated to perform the second action when at least the first data is input. [Explanation of Symbols]
[0070] 10... Estimation device, 11... Input unit, 11A... First data input unit, 11B... Second data input unit, 12... Preprocessing unit, 13... First learning unit, 15... Second learning unit, 16... Estimation unit, 17... Output unit, 1001... Processor, 1002... Memory, 1003... Storage, 1004... Communication device, 1005... Input device, 1006... Output device, 1007... Bus.
Claims
1. A first learning unit constructs a first estimation model for estimating the first actions of a first user group based on a first database having first data including the history of the first actions of each user constituting the first user group, A second learning unit constructs a second estimation model for estimating the second actions of the common users based on the first estimation model, using the second data for the common users in a second database which has second data including the history of the second actions of each user constituting a second user group which includes at least a common user group that is common to a part of the first user group, and the first data for the common users in the first database. An estimation device comprising: an estimation unit that estimates the second behavior of the first user group excluding the common user group using the second estimation model.
2. The estimation device according to claim 1, wherein the second action is an action related to the first action.
3. The estimation device according to claim 2, wherein the second action is an action corresponding to a sub-concept of the first action.
4. The first estimation model is a model that outputs a user who is estimated to perform the first action when the first data contained in the first database is input, The estimation device according to claim 1, wherein the second estimation model is a model in which a user is estimated to perform the second action when at least the first data is input.