JP2017168027A - Model generation device, model generation method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a model generation device for generating a novel model for a smartphone user based on an existing model for a PC user.SOLUTION: A model generation device for generating a model for a first device for estimating a user attribute from a behavior history on a first device of a user who uses the first device comprises: estimation means for estimating the same user between a user of the first device and a user of a second device different in kind from the first device; identification means for identifying the user attribute of a model for the second device corresponding to a behavior history of a user who is estimated as the same user, on the second device, using an existing model for the second device for estimating the user attribute from the behavior history of the user who uses the second device on the second device; and model generation means for, setting the identified user attribute to the user attribute of the model for the first device of the estimated user, for generating the model for the first device.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a model generation device, a model generation method, and a program.

  Conventionally, there is a model for estimating user attributes (for example, gender, age, annual income, family composition, preference, etc.) based on the past behavior history of the user on the Internet. In particular, there are already many models for estimating user attributes based on past behavior histories on the Internet for PC (personal computer) users. And by grasping | ascertaining the user attribute of a PC user using the said model, according to the attribute of a PC user, content, such as an effective advertisement, can be provided.

  In recent years, however, the number of young people who do not have a PC and access the Internet with only a mobile terminal such as a smartphone (hereinafter simply referred to as a smartphone) to collect information and perform purchasing behavior has increased. I am doing.

  For this reason, it is necessary to construct a new model for estimating user attributes based on past behavior histories on the Internet for smartphone users. However, in order to construct such a model, first, it is necessary to conduct a questionnaire survey on a wide range of smartphone users (for example, tens of thousands of people), which takes a long time.

  In addition, for example, as a questionnaire survey, items related to personal information such as “Is your annual income over XX yen?”, “Are you married or unmarried?”, “Are you interested in stock investment?” And a questionnaire about information about preferences. For this reason, in order for smartphone users to answer the questionnaire honestly, it is necessary to provide some kind of reward (points that can be used on the Internet, etc.), resulting in a large cost.

  On the other hand, a user behavior model is created probabilistically from user behavior history data without conducting an enormous questionnaire survey, and is used probabilistically according to predetermined conditions based on the behavior model A behavior prediction device is known that predicts a user's behavior and expands the condition to predict a user's behavior when the number of data of behavior history data matching the predetermined condition is less than a predetermined threshold. (For example, refer to Patent Document 1).

JP 2010-198243 A

  However, in the prior art of Patent Document 1, since the number of action history data is increased by performing processing such as expanding the value of a numerical input variable, the user attribute is estimated based on past action history. In such a case, there is a problem in that the difference between the estimated attribute and the actual attribute may increase.

  An object of the present invention is to provide a model generation apparatus, a model generation generation, and a program that can more suitably generate a model for a new device without taking time and cost.

In order to solve the above-mentioned problem, the invention according to claim 1
In a model generation device that generates a first device model for estimating a user attribute from an action history on a first device of a user who uses the first device,
Estimating means for estimating the same user between a user of the first device and a user of a second device of a different type from the first device;
Using the existing second device model for estimating the user attribute from the action history on the second device of the user who uses the second device, the user who is estimated to be the same user by the estimating means A specifying means for specifying a user attribute of the second device model corresponding to an action history on the second device;
Model generating means for generating the first device model by using the user attribute specified by the specifying means as the user attribute of the estimated user device model of the first device. To do.

The invention according to claim 2 is the model generation device according to claim 1,
The model generation unit may perform the estimation when the estimated action history of the first user on the first device matches or resembles the action history of the user other than the estimated user on the first device. The user attribute of the second device model corresponding to the user who has been made is the user attribute of the first device model of a user other than the estimated user.

The invention according to claim 3 is the model generation device according to claim 1 or 2,
The estimation means estimates that the user of the first device and the user of the second device are the same user when logging in with the same ID during a predetermined period.

The invention according to claim 4 is the model generation apparatus according to any one of claims 1 to 3,
The model generation unit repeats a process of generating a second device model based on the generated first device model after generating the first device model based on the second device model. It is characterized by.

The invention according to claim 5 is the model generation apparatus according to any one of claims 1 to 4,
The first device is a mobile terminal, and the second device is a personal computer.

The invention according to claim 6 is the model generation device according to any one of claims 1 to 4,
The first device is a personal computer, and the second device is a mobile terminal.

The invention described in claim 7
A model generation method for generating a first device model for estimating a user attribute from an action history on a first device of a user who uses the first device using a computer,
Estimating the same user between a user of the first device and a user of a second device of a different type from the first device;
On the second device of the user who is estimated to be the same user using the existing second device model for estimating the user attribute from the action history on the second device of the user who uses the second device Identifying a user attribute of the second device model corresponding to the behavior history of
Generating the first device model by using the identified user attribute as the user attribute of the estimated user device model of the first device.

The invention according to claim 8 provides:
Computer
Estimating means for estimating the same user between a user of a first device and a user of a second device of a different type from the first device;
Using the existing second device model for estimating the user attribute from the behavior history of the user using the second device on the second device, the user who is estimated to be the same user by the estimating means A specifying means for specifying a user attribute of the second device model corresponding to an action history on the second device;
A program that functions as a model generation unit that generates the first device model by using the user attribute specified by the specifying unit as the user attribute of the estimated first device model of the user.

  According to the present invention, it is possible to more suitably create a model for a new device without spending time and cost.

It is a block diagram which shows the structure of the model production | generation apparatus which concerns on embodiment. It is a flowchart which shows an example of operation | movement of a model production | generation apparatus. It is explanatory drawing which shows the relationship between the user attribute derived | led-out from the model for PC users, and action history. It is explanatory drawing which shows an example of a smart user's action log | history. It is explanatory drawing which shows the relationship between the user attribute derived | led-out from the produced | generated model for smartphone users, and action history. It is explanatory drawing which shows an example of the relationship between age sex and the score of a household annual income model. It is explanatory drawing which shows an example of the relationship between age sex and the score of a household asset model. It is explanatory drawing which shows an example of the relationship between the TV program and the score of a household annual income model which are often seen. It is explanatory drawing which shows an example of the relationship between the television program and the score of a household asset model which are often seen.

(Embodiment)
[1. Description of configuration]
Hereinafter, a model generation apparatus according to an embodiment of the present invention will be described in detail with reference to the drawings. However, the scope of the invention is not limited to the illustrated examples.

[1-1. Description of model generator configuration]
The configuration of the model generation apparatus according to the embodiment of the present invention will be described with reference to FIG. FIG. 1 is a schematic configuration diagram illustrating functions of a model generation device 100 (hereinafter simply referred to as the device 100) as a block diagram.

  As shown in FIG. 1, the apparatus 100 includes a control unit 1, a storage unit 2, a display unit 3, an operation unit 4, a communication unit 5, a PC login history storage unit 6, a smartphone login history storage unit 7, and a PC user attribute storage unit. 8 and a smartphone action history storage unit 9. In the device 100, the control unit 1, the storage unit 2, the display unit 3, the operation unit 4, the communication unit 5, the PC login history storage unit 6, the smartphone login history storage unit 7, the PC user attribute storage unit 8, and the smartphone action history. The storage units 9 are connected to each other by an internal bus or the like.

  The control unit 1 performs central control of the operation of the device 100. Specifically, the control unit 1 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. Each unit is controlled in cooperation with the program data stored in the CPU and the CPU. Further, as will be described later, the control unit 1 operates as an estimation unit, a specification unit, and a model generation unit.

  The storage unit 2 stores data such as program data and various setting data in a readable / writable manner from the control unit 1. For example, the storage unit 2 may be an HDD (Hard Disk Drive), a semiconductor memory, or the like.

  The display unit 3 displays an image based on the display control signal output from the control unit 1 on the display screen. For example, the display unit 3 may be an LCD (Liquid Crystal Display), an FPD (Flat Panel Display) using an organic EL (Electro Luminescence) element, or the like. The display unit 3 may be a touch panel formed integrally with the operation unit 4 such as a smartphone or a tablet terminal.

  The operation unit 4 receives an operation input from a user and outputs an operation signal corresponding to the operation input to the control unit 1. For example, the operation unit 4 may be a touch panel formed integrally with the display unit 3 such as a smartphone or a tablet terminal.

  The communication unit 5 communicates with an external device via a network (not shown) and transmits / receives data. For example, PC or smartphone login information stored in the external server is acquired and output to the control unit 1.

  The PC login history storage unit 6 stores a login history including information such as date and time and a user ID when logged in from the PC in a readable / writable manner from the control unit 1. For example, the PC login history storage unit 6 may be an HDD or a semiconductor memory. Further, the function of the PC login history storage unit 6 may be assigned to a partial area of the storage unit 2.

  The smartphone login history storage unit 7 stores a login history including information such as date and time and a user ID when logged in from the smartphone in a readable / writable manner from the control unit 1. For example, the smartphone login history storage unit 7 may be an HDD or a semiconductor memory. Further, the function of the smartphone login history storage unit 7 may be assigned to a partial area of the storage unit 2.

  The PC user attribute storage unit 8 stores the generated user attribute (estimated data) from the existing PC user model together with the past action history on the PC so as to be readable and writable from the control unit 1. For example, the PC user attribute storage unit 8 may be an HDD or a semiconductor memory. Further, the function of the PC user attribute storage unit 8 may be assigned to a partial area of the storage unit 2. Further, the model for the existing PC user is stored, and the action history on the past PC is input to the model for the existing PC user under the control of the control unit 1, and the user attribute (estimated data) is set. It may be generated.

  The smartphone action history storage unit 9 stores the action history for the smartphone user so as to be readable and writable from the control unit 1. Further, the control unit 1 extracts and writes the user attribute corresponding to the behavior history for the smartphone user from the existing PC user attribute storage unit 8. For example, the smartphone action history storage unit 9 may be an HDD or a semiconductor memory. Further, the function of the smartphone action history storage unit 9 may be assigned to a partial area of the storage unit 2.

[2. Explanation of operation]
A specific operation of the apparatus 100 according to the embodiment of the present invention will be described in detail with reference to FIGS. Hereinafter, for convenience of explanation, the processing that is mainly performed by the control unit 1 of the apparatus 100 will be described with the “apparatus 100” as the main body of the processing.
[2-1. Description of operation of apparatus 100]
Here, the operation of the apparatus 100 will be described with reference to the flowchart of FIG.
The apparatus 100 reads out the login histories from the PC login history storage unit 6 and the smartphone login history storage unit 7 (step S21), and compares the PC login history with the smartphone login history to estimate the same user (step S21). S22: Estimating means).

  Estimating the same user is, for example, by matching the login history of the PC with the login history of the smartphone and logging in from both the PC and the smartphone using the same user ID for a predetermined period. It is to estimate that the user and the smartphone user are the same person.

  When it is determined that the same user is the same user (step S23: Yes), the apparatus 100 reads the user attribute of the same user from the PC user attribute storage unit 8 (step S24: specifying means), and for the smartphone user As a user attribute of the same user in this model, it is written in the smartphone action history storage unit 9 (step S25: model generation means), and the process proceeds to step S30.

  Specifically, for example, the PC user attribute storage unit 8 stores an action history in the past PC as shown in FIG. 3 and user attributes corresponding thereto. On the other hand, the smartphone action history storage unit 9 stores only the action history in the past smartphone as shown in FIG. 4, the user attribute required for model creation is blank, and only such data is stored. It is impossible to construct a model for a smartphone user.

  For this reason, for example, when it is estimated that the user with the PC user identifier “P002” and the user with the smartphone user identifier “S002” are the same user, as indicated by WR51 in FIG. 5, the PC user identifier “P002” The user attribute of the user (for example, “I am interested in sports”) is written in the smartphone action history storage unit 9 as the user attribute of the user with the smartphone user identifier “S002”, and the action history on the smartphone user as shown in FIG. The user attribute combination is completed, and the model for the smartphone user is tuned and completed using such data.

  Similarly, for example, when it is estimated that the user with the PC user identifier “P004” and the user with the smartphone user identifier “S004” are the same user, as indicated by WR52 in FIG. 5, the PC user identifier “P004” The user attribute of the user (for example, “I am interested in a car…”) is written in the smartphone action history storage unit 9 as the user attribute of the user with the smartphone user identifier “S004”, and the behavior on the smartphone user is shown in FIG. A combination of history and user attributes is sequentially completed, and a model for a smartphone user is tuned and completed using such data.

  On the other hand, when the device 100 determines that they are not the same user (step S23: No), the behavior history for the smartphone user of the user who is estimated to be the same user in the previous process is sequentially read from the smartphone behavior history storage unit 9. Collation is performed (step S26), and it is determined whether or not there is a user having an action history for the same or similar smartphone user (step S27).

  If the device 100 determines that there is no identical or similar behavior history in the behavior history for the smartphone user (step S27: No), the device 100 proceeds to step S30.

  On the other hand, if the device 100 determines that there is an identical or similar behavior history in the behavior history for smartphone users (step S27: Yes), the user of the user who is estimated to be the same user from the PC user attribute storage unit 8 The attribute is read (step S28), written in the smartphone action history storage unit 9 as the user attribute of the other user having the same or similar action history in the model for smartphone users (step S29), and the process proceeds to step S30.

  That is, the behavior history in the past PC and the behavior history in the past smartphone are quite different even if they are the same person, so the behavior history in the past PC and the behavior history in the past smartphone are It cannot be matched as it is. For this reason, when there is an action history for the same or similar smartphone user by matching the action history of the user estimated to be the same user with the action history of the past smartphone of the other user and the action history of the other user with the past smartphone, The user attribute stored in the PC user attribute storage unit 8 of the estimated user is linked to the behavior history of other users in the past smartphone.

  Specifically, for example, the behavior histories of the user with the smartphone user identifier “S004” and the user with the smartphone user identifier “S001” match except for the number of times of “page view of car”. In other words, the action history is similar. Therefore, as indicated by WR53 in FIG. 5, the user attribute stored in the PC user attribute storage unit 8 of the user who is estimated to be the same user (the user with the PC user identifier “P004”) is replaced with the smartphone user identifier “S001”. ”Is written in the smartphone action history storage unit 9 to generate a model for the smartphone user.

  As shown in the flowchart of FIG. 2, the user attribute stored in the PC user attribute storage unit 8 is directly read, and the smartphone action history storage unit 9 is used as the user attribute of the other user having the same or similar action history. Alternatively, the user attribute stored in the smartphone action history storage unit 9 may be read out and written to the smartphone action history storage unit 9 as the user attribute of the other user having the same or similar action history.

  That is, the user attribute is written in the smartphone behavior history storage unit 9 corresponding to the user estimated to be the same user in the previous process, and the written user attribute is stored in the PC user attribute storage unit 8. Therefore, the user attributes stored in the smartphone behavior history storage unit 9 can be used.

  By the way, when reading the user attribute stored in the smartphone action history storage unit 9 and writing it in the smartphone action history storage unit 9 as the user attribute of the other user having the same or similar action history, the smartphone action history storage unit 9, it is necessary to determine whether there is a user history of a user having the same or similar behavior history.

  In addition, even if the behavior history of the smartphone user having the user attribute is “car page view” and the behavior history of the other smartphone user is “game page view”, it is similar by existing similarity determination logic When the determination is made, the user attribute of the smartphone user “interested in the car” can be replaced with “interested in the game” to be the user attribute of the other smartphone user.

  Finally, apparatus 100 determines whether or not processing has been completed for all users (step S30), and determines that processing has not been completed for all users (step S30: No). Returning to step S22, if it is determined that the processing has been completed for all users (step S30: Yes), the operation is terminated.

[3. inspection result]
The verification result of the model for smartphone users generated using the user attributes of the existing model for PC users will be described.

  FIG. 6 is an explanatory diagram showing an example of the relationship between age sex and household annual income score (score of household annual income model). In FIG. 6, the smartphone user's household annual income score (the higher the household annual income score is, the higher the household annual income is estimated) is plotted. The male has a higher household annual income score than the female, and the age is higher. In general, in the range from 20 to 70 years old, the household annual income score increases as the age increases.

  FIG. 7 is an explanatory diagram showing an example of the relationship between the age and gender and the household asset score (score of the household asset model). FIG. 7 is a plot of the smartphone user's household asset score (the higher the household asset score, the more household asset is estimated), and the male has a higher household asset score and the age is higher than the female. In the range of about 20 to 70 years old, the household asset score increases as the age increases.

  FIG. 8 is an explanatory diagram showing an example of the relationship between frequently viewed television programs and household annual income scores. As a program, “business program” and “late-night animation” are compared. In FIG. 8, at the same age, the user who often watches “business programs” has a higher household income score than the user who often watches “late-night animation”.

  Further, FIG. 9 is an explanatory diagram showing an example of the relationship between frequently watched television programs and household asset scores. As a program, “business program” and “late-night animation” are compared. In FIG. 9, at the same age, the user who often watches the “business program” has a higher household asset score than the user who often watches the “midnight animation”.

  That is, the basic trend of household annual income, etc., by age and gender from FIG. 6 and FIG. 7, the trend of higher household annual income, etc. for men than for women, and the age range from 20 to 70 years There is a tendency that is consistent with the common sense that household income increases as the value increases.

  8 and 9, it is clear that users who often watch “business programs” (for example, businessmen) are more likely to earn household annual income than users who often watch “late-night animation” (for example, students and freelancers). A tendency to coincide with common sense such as an increase is shown, and it can be seen that the model for smartphone users generated using the user attributes of the existing model for PC users is functioning.

As described above, by estimating the same user between the PC user and the smartphone user from the login history, and setting the user attribute of the model for the PC user of the same user as the user attribute of the model for the smartphone user of the same user A smartphone model can be generated without conducting a huge questionnaire survey.
Moreover, although it is not the same user, when the action history is the same or similar, a smartphone model can be generated by estimating that the user attribute is the same or similar.

  In the description of the embodiment, a model for a smartphone user is generated using a user attribute of an existing model for a PC user. However, a new attribute is generated using a user attribute of a generated model for a smartphone user. It is also possible to generate a model for a PC user.

A model for a smartphone user may be generated based on a newly generated model for a PC user. That is, the process of generating the other model based on one model can be repeated. Thus, by repeating the process of generating a model, the consistency of the model is improved and the reliability of the model is increased.
Also, a model for a device other than a smartphone can be generated based on a model for a PC user or a model for a smartphone user.

  In the description of the embodiment, information is stored in the PC login history storage unit 6, the smartphone login history storage unit 7, the PC user attribute storage unit 8, and the smartphone action history storage unit 9, respectively. 2 may store all information.

  In the description of the embodiment, information is stored in the PC login history storage unit 6, the smartphone login history storage unit 7, the PC user attribute storage unit 8, and the smartphone action history storage unit 9, respectively. The server may manage these data collectively or in a distributed manner, control the communication unit 5, and acquire information as necessary.

  In the description of the embodiment, when the login history of the PC and the login history of the smartphone are matched, and the same user ID is used for the predetermined period, the PC and the smartphone are logged in. Although it is estimated that the user and the smartphone user are the same person, the same user may be estimated by other existing estimation methods.

100 Model generation apparatus 1 Control unit (estimating means, specifying means, model generating means)
2 storage unit 3 display unit 4 operation unit 5 communication unit 6 PC login history storage unit 7 smartphone login history storage unit 8 PC user attribute storage unit 9 smartphone action history storage unit

Claims (8)

In a model generation device that generates a first device model for estimating a user attribute from an action history on a first device of a user who uses the first device,
Estimating means for estimating the same user between a user of the first device and a user of a second device of a different type from the first device;
Using the existing second device model for estimating the user attribute from the action history on the second device of the user who uses the second device, the user who is estimated to be the same user by the estimating means A specifying means for specifying a user attribute of the second device model corresponding to an action history on the second device;
Model generating means for generating the first device model by using the user attribute specified by the specifying means as the user attribute of the estimated user device model of the first device. Model generation device.   The model generation unit may perform the estimation when the estimated action history of the first user on the first device matches or resembles the action history of the user other than the estimated user on the first device. 2. The model generation according to claim 1, wherein a user attribute of the second device model corresponding to the determined user is a user attribute of the first device model of a user other than the estimated user. apparatus.   The model generation unit estimates that a user of the first device and a user of the second device are the same user when logging in with the same ID during a predetermined period. Or the model production | generation apparatus of 2.   The model generation unit repeats a process of generating a second device model based on the generated first device model after generating the first device model based on the second device model. The model generation apparatus according to any one of claims 1 to 3, wherein   5. The model generation apparatus according to claim 1, wherein the first device is a portable terminal, and the second device is a personal computer.   5. The model generation apparatus according to claim 1, wherein the first device is a personal computer, and the second device is a mobile terminal. 6. A model generation method for generating a first device model for estimating a user attribute from an action history on a first device of a user who uses the first device using a computer,
Estimating the same user between a user of the first device and a user of a second device of a different type from the first device;
On the second device of the user who is estimated to be the same user using the existing second device model for estimating the user attribute from the action history on the second device of the user who uses the second device Identifying a user attribute of the second device model corresponding to the behavior history of
Generating the first device model by using the identified user attribute as the user attribute of the estimated model for the first device of the user. Computer
Estimating means for estimating the same user between a user of a first device and a user of a second device of a different type from the first device;
Using the existing second device model for estimating the user attribute from the behavior history of the user using the second device on the second device, the user who is estimated to be the same user by the estimating means A specifying means for specifying a user attribute of the second device model corresponding to an action history on the second device;
A program that functions as a model generation unit that generates the first device model by using the user attribute specified by the specifying unit as the user attribute of the estimated first device model of the user.

Images