JP2019220100A - Estimation device - Google Patents

Abstract

To provide an estimation device capable of estimating occurrence of a user's life event on the basis of information which can be easily acquired while improving precision.SOLUTION: An estimation device 10 is a device which estimates a user's life event, and comprises: an acquisition part 11 which acquires input information of each user including contract information associated with at least a contract of a portable terminal; a feature quantity generation part 12 which generates a feature quantity for inputting to a life event estimation model for estimating the user's life event on the basis of the input information; a model generation part 13 which generates a life event estimation model inputting the feature quantity and outputting an estimated value associated with occurrence of a predetermined life event on the basis of the feature quantity; and a model output part 14 which outputs the life event estimation model.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a life event estimation device.

  2. Description of the Related Art There is known a technology for recommending products and services to a customer based on information such as a customer's action history and purchase history. For example, Patent Literature 1 discloses a technique of estimating an event in a user's life based on a response to a questionnaire by a customer and purchase information of a product, and sending advertising information related to the estimated event to the customer. I have.

JP 2004-206161 A

  Conventionally, a life event, which is an event in the life of a user, has been estimated based on information such as a user's answer to a questionnaire and a purchase history of a product of the user. That is, since information based on the active behavior of the user is used, it is not easy to obtain such information. Further, it has been required to improve the accuracy of life event estimation without forcing the user to provide specific information.

  Therefore, the present invention has been made in view of the above problems, and provides an estimating apparatus that can estimate the occurrence of a life event of a user based on information that can be easily obtained while improving accuracy. With the goal.

  In order to solve the above problem, an estimation device according to an aspect of the present invention is an estimation device for estimating a life event of a user, and acquires input information for each user including at least contract information on a contract of a mobile terminal. An acquisition unit, a feature amount generation unit that generates a feature amount for input to a life event estimation model for estimating a user's life event based on input information, and a feature amount based on the feature amount. , A model generation unit that generates a life event estimation model that outputs an estimated value related to the occurrence of a predetermined life event, and a model output unit that outputs the generated life event estimation model.

  According to the above aspect, since the information related to the contract of the mobile terminal is used for generating the feature amount, the feature amount to be input to the life event estimation model can be obtained based on the easily obtainable information. Since the information related to the contract of the mobile terminal is information that reflects the aspect of the user's life, a model that outputs an estimated value related to the occurrence of a life event using a feature amount generated based on such information as an input is used. By generating, it is possible to accurately estimate the life event of the user.

  Estimation of the occurrence of a user's life event can be performed based on information that can be easily acquired while improving accuracy.

FIG. 2 is a diagram illustrating a functional configuration of an estimation device according to the embodiment. It is a hardware block diagram of an estimation device. FIG. 4 is a diagram illustrating an example of a feature amount to be input to a life event estimation model. It is a flowchart which shows the generation process of a life event estimation model. FIG. 14 is a diagram illustrating an example of the output probability of a life event. It is a flowchart which shows the process which estimates the occurrence probability of a predetermined life event using a life event estimation model. FIG. 3 is a diagram illustrating generation of a model for estimating a probability of occurrence of a life event in consideration of a life stage. It is a figure explaining processing of estimating the occurrence probability of a life event in consideration of a life stage. It is a figure showing composition of an estimation program.

  An embodiment of an estimation device according to the present invention will be described with reference to the drawings. Note that, where possible, the same portions are denoted by the same reference numerals, and redundant description will be omitted.

  FIG. 1 is a diagram illustrating a functional configuration of an estimation device 10 according to the present embodiment. The estimation device 10 is a device that estimates a life event of a user. A life event is a milestone in the lifetime of a person (user). In addition, the estimation device 10 may give a recommendation such as a product and a service to the user based on the estimated value regarding the occurrence of the life event.

  In the example illustrated in FIG. 1, the estimation device 10 includes an acquisition unit 11, a feature amount generation unit 12, a model generation unit 13, a model output unit 14, an estimation unit 15, and a proposal unit 16. Each functional unit included in the estimation device 10 of the present embodiment may be configured in one device, or may be configured to be distributed to a plurality of devices.

  In addition, the estimation device 10 includes storage units such as an input information storage unit 20, a model storage unit 30, and a proposal information storage unit 40. Each storage unit included in the estimation device 10 of the present embodiment may be configured in the estimation device 10 or may be configured in another device that can communicate with the estimation device 10.

  Further, the estimation device 10 may be configured by a minimum function unit for generating a model. That is, the estimation device 10 may be configured to include the acquisition unit 11, the feature amount generation unit 12, the model generation unit 13, and the model output unit 14.

  Further, the estimation device 10 may be configured with a minimum function unit for outputting an estimated value related to the occurrence of the life event using the generated model. That is, the estimation device 10 may be configured to include the acquisition unit 11, the feature amount generation unit 12, and the estimation unit 15.

  Note that the block diagram shown in FIG. 1 shows blocks in functional units. These functional blocks (components) are realized by any combination of hardware and / or software. Further, the means for realizing each functional block is not particularly limited. That is, each functional block may be realized by one device physically and / or logically coupled, or two or more devices physically and / or logically separated from each other directly and / or indirectly. (For example, wired and / or wireless), and may be realized by the plurality of devices.

  For example, the estimation device 10 according to an embodiment of the present invention may function as a computer. FIG. 2 is a diagram illustrating an example of a hardware configuration of the estimation device 10 according to the present embodiment. The estimation device 10 may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, and the like.

  In the following description, the term “apparatus” can be read as a circuit, a device, a unit, or the like. The hardware configuration of the estimation device 10 may be configured to include one or more devices illustrated in FIG. 2 or may be configured without including some devices.

  The functions of the estimation device 10 are performed by reading predetermined software (program) on hardware such as the processor 1001 and the memory 1002, so that the processor 1001 performs an arithmetic operation. This is realized by controlling the reading and / or writing of data in.

  The processor 1001 controls the entire computer by operating an operating system, for example. The processor 1001 may be configured by a central processing unit (CPU) including an interface with a peripheral device, a control device, an arithmetic device, a register, and the like. Further, the processor 1001 may include a GPU (Graphics Processing Unit). For example, the functional units 11 to 16 illustrated in FIG. 1 may be realized by the processor 1001.

  In addition, the processor 1001 reads a program (program code), a software module, and data from the storage 1003 and / or the communication device 1004 to the memory 1002, and executes various processes according to these. As the program, a program that causes a computer to execute at least a part of the operation described in the above embodiment is used. For example, each of the functional units 11 to 16 of the estimation device 10 may be realized by a control program stored in the memory 1002 and operated by the processor 1001. Although it has been described that the above-described various processes are executed by one processor 1001, the processes may be executed simultaneously or sequentially by two or more processors 1001. Processor 1001 may be implemented with one or more chips. Note that the program may be transmitted from a network via a telecommunication line.

  The memory 1002 is a computer-readable recording medium, and includes, for example, at least one of a ROM (Read Only Memory), an EPROM (Erasable Programmable ROM), an EEPROM (Electrically Erasable Programmable ROM), and a RAM (Random Access Memory). May be done. The memory 1002 may be called a register, a cache, a main memory (main storage device), or the like. The memory 1002 can store a program (program code), a software module, and the like that can be executed to execute the shelf allocation information generation method according to one embodiment of the present invention.

  The storage 1003 is a computer-readable recording medium, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, and a magneto-optical disk (for example, a compact disk, a digital versatile disk, a Blu-ray). (Registered trademark) disk, smart card, flash memory (for example, card, stick, key drive), floppy (registered trademark) disk, magnetic strip, or the like. The storage 1003 may be called an auxiliary storage device. The storage medium described above may be, for example, a database including the memory 1002 and / or the storage 1003, a server, or any other suitable medium.

  The communication device 1004 is hardware (transmission / reception device) for performing communication between computers via a wired and / or wireless network, and is also referred to as, for example, a network device, a network controller, a network card, a communication module, or the like.

  The input device 1005 is an input device that receives an external input (for example, a keyboard, a mouse, a microphone, a switch, a button, a sensor, and the like). The output device 1006 is an output device that performs output to the outside (for example, a display, a speaker, an LED lamp, and the like). Note that the input device 1005 and the output device 1006 may have an integrated configuration (for example, a touch panel).

  Each device such as the processor 1001 and the memory 1002 is connected by a bus 1007 for communicating information. The bus 1007 may be configured by a single bus, or may be configured by a different bus between devices.

  The estimating apparatus 10 includes hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array). Alternatively, some or all of the functional blocks may be implemented by the hardware. For example, the processor 1001 may be implemented by at least one of these hardware.

  With reference to FIG. 1 again, each functional unit of the estimation device 10 will be described. The acquisition unit 11 acquires input information for each user. The input information is obtained for generating a model. In the estimation device 10 of the present embodiment, a life event estimation model for estimating a life event of a user is generated. The life event estimation model inputs a feature amount generated based on input information and outputs an estimated value related to occurrence of a predetermined life event. The estimated value output by the life event estimation model is not limited as long as it is a value relating to the occurrence of a predetermined life event. In the present embodiment, an example in which the life event estimation model outputs the occurrence probability of a predetermined life event is described. explain.

  When the life event estimation model is generated by supervised learning, the acquisition unit 11 acquires input information and event information associated with the input information. The event information is used as correct information in supervised learning.

  The input information is a group of information representing characteristics of a certain user at a certain point in time, and includes at least contract information relating to a contract of a mobile terminal. Such contract information is information that can be easily obtained without forcing the user to perform any action for providing special information. The event information is an event related to the user that has occurred at a certain point in time, and includes, for example, life event information and other information that can be interpreted as a life event. Life event information is information indicating a life event that has occurred in the user at a certain point in time. Information that can be interpreted as a life event is, for example, a predetermined name change procedure (such as contract information of a mobile terminal) by a user. In the present embodiment, an example where the event information is life event information will be described.

  The input information storage unit 20 is a storage unit that stores input information. Further, the input information storage unit 20 may store data used for learning the model, such as life event information associated with the input information.

  As described above, the input information includes contract information related to the contract of the mobile terminal, and includes, for example, information such as a charge plan, a relationship between customers in the contract contents, a terminal type, a charged amount, and a charge storage state. Further, the input information may include information on the contract status of various services such as services tailored to the age, such as for children and the elderly, and services for distributing contents such as moving images. Further, the input information may include information such as the number of work days and travel days estimated from the position information and the time spent in various facilities. Further, the input information may include statistical information based on a census or the like in addition to the information unique to the user.

  The life event information is information indicating whether or not a life event has occurred in the user at the time when the associated input information is associated. As described above, a life event is a milestone event in the lifetime of a person (user), and is, for example, an event such as marriage, birth of a child, and purchase of a house.

  The feature amount generation unit 12 generates a feature amount based on the input information. Specifically, the feature amount generation unit 12 generates, by processing the input information, a feature amount that is data to be input to the model in order to generate the life event estimation model and that indicates a feature of each input information.

  An example of generation of a feature by the feature generator 12 will be described below. Based on the input information, the characteristic amount generation unit 12 converts the numerical value of the contract information or the like included in the input information, a total value obtained by totaling the numerical values, or information indicating the attribute at one time point or a predetermined period in the time series into the characteristic amount. Is generated as Specifically, the feature value generation unit 12 may generate, for example, a numerical value, a total value (average value, maximum value, minimum value, and the like) and an attribute value in the latest predetermined period as the feature value.

  In addition, based on the input information, the feature amount generation unit 12 includes information indicating a numerical value, a total value, or an attribute of contract information or the like included in the input information at the first time point or the first period in the time series; The information indicating a change from the numerical value, the aggregate value, or the information indicating the attribute in the second time point before the time point or in the second time period before the first time period may be generated as the feature amount. More specifically, the feature amount generation unit 12 stores information indicating a change in a numerical value, a total value (average value, maximum value, minimum value, and the like), an attribute value, and the like between two latest time points and two periods. It may be generated as a feature amount.

  FIG. 3 is a diagram illustrating an example of a feature amount generated by the feature amount generation unit 12. As shown in FIG. 3, the feature amount generation unit 12 includes the feature amounts X1 (charged amount), X2 (terminal type), X3 (use period),... Of the users A, B, C,. .., Xn are generated. These feature amounts X1 to Xn are data input to the life event estimation model, and constitute explanatory variables. Further, life event information Y as correct answer information corresponding to each feature amount is associated with the feature amount of each user. The life event information Y forms a target variable.

  In addition, the feature amount generation unit 12 estimates a group to which the user belongs based on the input information, and outputs at least one of the information indicating the group and the predetermined input information regarding other users belonging to the same group. It may be generated as a part of the feature amount.

  Specifically, the feature amount generation unit 12 may extract identification information indicating a group to which the user belongs based on the input information, and generate the extracted identification information as a part of the feature amount. Further, the feature amount generation unit 12 may output input information of another user belonging to the same group as one user as a part of the feature amount of the one user.

  Since the occurrence of a life event may be influenced by the group to which the user belongs and other users in the group, the feature amount is generated in this manner, so that information about the group and information about other users in the same group are characterized. Since it is included in the quantity, it is possible to generate a model with high estimation accuracy of occurrence of a life event.

  The model generation unit 13 generates a life event estimation model that outputs an estimated value related to the occurrence of a predetermined life event based on the feature amount based on the feature amount generated by the feature amount generation unit 12. When the life event estimation model is generated by supervised learning, the model generation unit 13 generates a life event estimation model based on the feature amount and the event information. In the present embodiment, as described above, the model generation unit 13 uses the feature amount as an input (explanatory variable) based on the feature amount of each user and the life event information associated with each feature amount, and A case where a life event estimation model that outputs an estimated value related to occurrence is generated by supervised learning will be described. The life event estimation model is represented by, for example, a calculation formula and other logic, and is generated by a well-known learning method for generating a model. In the present embodiment, the model generation unit 13 describes an example of generating a life event estimation model by machine learning using feature amounts as explanatory variables and life event information as target variables, but the learning method is not limited to machine learning. . The model generation unit 13 may generate a life event estimation model by unsupervised learning. In this case, since the model generation unit 13 generates the life event estimation model based on the feature amount generated by the feature amount generation unit 12, the acquisition unit 11 does not need to acquire the event information.

  Note that generation of a life event estimation model is not limited to a technique based on a certain machine learning. For example, the life event estimation model may be generated without using machine learning, or may be generated using either supervised or unsupervised machine learning. For example, when a life event estimation model is generated by supervised machine learning, the model generation unit 13 generates a life event estimation model by applying a method such as regression, classification, and clustering. More specifically, the model generation unit 13 uses the explanatory variable and the objective variable shown in FIG. 3 as an input and an output, respectively, for example, a decision tree, a linear regression, a logistic regression, a support vector machine, a random forest, a neural network, The life event estimation model may be generated by a technique such as gradient boosting.

  The model output unit 14 outputs the life event estimation model generated by the model generation unit 13. Specifically, the model output unit 14 causes the model storage unit 30 to store, for example, a life event estimation model. When the estimation device 10 outputs an estimated value related to the occurrence of a life event, the model output unit 14 stores the life event estimation model in the model storage unit 30 in order to provide the estimation processing. Further, the model output unit 14 may output the model output to a device that estimates an estimated value related to the occurrence of a life event.

  FIG. 4 is a flowchart illustrating a process of generating a life event estimation model according to the present embodiment. In step S1, the acquiring unit 11 acquires input information for each user and life event information associated with the input information. When a life event estimation model is generated by unsupervised learning, the acquisition unit 11 acquires input information in step S1.

  In step S2, the feature value generation unit 12 generates a feature value for generating a model based on the input information acquired in step 1.

  In step S3, the model generation unit 13 generates a predetermined life event based on the characteristic amount based on the characteristic amount generated by the characteristic amount generation unit 12 and the life event information associated with the characteristic amount for each user. A life event estimation model that outputs probabilities is generated by machine learning. When the life event estimation model is generated by unsupervised learning, in step S3, the model generation unit 13 generates the life event estimation model based on the feature amount without using the life event information.

  In step S4, the model output unit 14 outputs the life event estimation model generated by the model generation unit 13.

  Next, the function of each functional unit in the phase of estimating the probability of occurrence of a life event using the life event estimation model in the estimation device 10 will be described.

  The acquisition unit 11 acquires input information on a user whose life event occurrence probability is to be estimated. The content of the input information is the same as the input information obtained in the model generation phase.

  The feature value generation unit 12 generates a feature value to be input to the life event estimation model based on the input information acquired by the acquisition unit 11. That is, the feature amount generation unit 12 processes the input information into a format for inputting to the life event estimation model.

  The estimating unit 15 uses the life event estimation model generated by the model generating unit 13 to output an estimated value related to the occurrence of a predetermined life event related to the user based on the feature amount generated by the feature amount generating unit 12. . In the present embodiment, as an example, the estimating unit 15 estimates the occurrence probability of a predetermined life event related to the user. FIG. 5 is a diagram illustrating an example of a life event occurrence probability estimated by the estimation unit 15. As illustrated in FIG. 5, based on the feature amount of the user A, the estimation unit 15 determines that the probability that the event “marriage” occurs in the user A is “0.2” and the probability that the event “birth of the child” occurs is “0”. .8 "and the probability" 0.1 "that the event" buy a house "will occur is estimated. Similarly, the estimating unit 15 calculates the probability “0.9” that the event “marriage” occurs in the user B, the probability “0.3” that the event “birth of a child” occurs, based on the feature amount of the user B, And the probability “0.2” of the occurrence of the event “purchase of a house” is estimated.

  When the life event estimation model generated by the model generation unit 13 is stored in the model storage unit 30, the estimation unit 15 refers to the model storage unit 30 and determines the life event estimation model used for the estimation. get.

  The proposing unit 16 proposes a product or a product to be proposed to the user based on the association between the life event stored in advance and the predetermined product or service in accordance with the probability of occurrence of the predetermined life event estimated by the estimation unit 15. Output information about services.

  In the present embodiment, for example, the proposal information storage unit 40 stores, in advance, proposal information that is information in which information on a product or the like proposed to a user is associated with a predetermined event and the occurrence probability of the event. The proposal unit 16 refers to the proposal information stored in the proposal information storage unit 40 and corresponds to the occurrence probability of a certain user's predetermined life event estimated by the estimation unit 15 (for example, see FIG. 5). Extract information about products and the like. Then, the suggestion unit 16 transmits the information on the extracted product or the like to the user (the terminal of the user or the like) as recommendation information. Such processing of the estimating unit 15 enables appropriate recommendation of a product or the like to the user according to the estimated life event occurrence probability.

  FIG. 6 is a flowchart illustrating a process of estimating a probability of occurrence of a life event using a life event estimation model. In step S11, the obtaining unit 11 obtains input information regarding a user whose life event occurrence probability is to be estimated.

  In step S12, the feature amount generation unit 12 generates a feature amount to be input to the model based on the input information acquired in step S11.

  In step S13, the estimation unit 15 estimates the occurrence probability of a predetermined life event for the user based on the feature amount generated in step S12, using the life event estimation model generated by the model generation unit 13.

  In step S14, the estimating unit 15 outputs the occurrence probability of the life event for each user, which is estimated in step S14. The output of the life event occurrence probability may be, for example, storage in a predetermined storage unit, display by a predetermined display device, and output of recommendation information by the suggestion unit 16.

  Next, an example of generation of a life event estimation model in consideration of information of a life stage to which a user belongs and estimation of a life event occurrence probability will be described with reference to FIGS. 7 and 8. The life stage is a stage of the living environment divided by a life event. For example, the pre-stage separated by the life event of “marriage” is a life stage of “unmarried”, and the post-stage of “marriage” is a life stage of “married”.

  First, generation of a life event estimation model in consideration of a life stage will be described. FIG. 7 is a diagram illustrating generation of a life event estimation model for outputting an estimated value related to the occurrence of a life event in consideration of a life stage.

  First, prior to generation of a model by the model generation unit 13, the acquisition unit 11 acquires input information for each user and life stage information associated with the input information. The life stage information is information indicating a life stage to which a user corresponding to the life stage information belongs, and is used as correct answer information (objective variable) in generation of a life stage estimation model described later.

  In the output of the estimated value related to the occurrence of the life event in consideration of the life stage, the model generation unit 13 outputs the life stage estimation model, the first conditional life event estimation model, and the second conditional life event estimation model as the life event. Generate as an estimation model.

  The life stage estimation model is a model in which a feature amount of a user is input and an estimated value regarding whether or not the user belongs to a predetermined life stage is output. In the present embodiment, the life stage estimation model outputs a probability that a user belongs to a predetermined life stage. Note that the estimated value regarding whether or not the user belongs to the life stage is not limited to the probability that the user belongs to the predetermined life stage. For example, it is determined whether the user belongs to the predetermined life stage or not. May be the value indicated by.

  Specifically, the model generation unit 13 uses the feature amount group F1 for each user generated based on the input information as an explanatory variable, and the life stage information LS as correct information as an objective variable, and uses a predetermined method to determine the life. A stage estimation model M10 is generated. The method of generating the life stage estimation model M10 is not limited, but the model generation unit 13 generates the life stage estimation model M10 by, for example, predetermined machine learning.

  More specifically, for example, the model generation unit 13 uses the life stage information LS as the correct answer information indicating whether the user is “unmarried” or “married”, so that the user is “unmarried”. It is possible to generate a life stage estimation model M10 including a calculation formula that outputs a probability.

  On the other hand, in the generation of the first / second conditional life event estimation models M11 and M12, first, the model generation unit 13 determines the feature amount group F1 for each user based on the life stage information LS as a determination target. Judgment is made to one of the feature amount group F11 of the user belonging to the predetermined life stage and the feature amount group F12 of the user belonging to the predetermined life stage to be judged (D1).

  More specifically, for example, the model generation unit 13 uses the life stage information LS as the correct answer information indicating whether the user is “unmarried” or “married”, and thereby sets the feature amount group F1 to “ The feature amount group F11 of the “unmarried” user and the feature amount group F12 of the “married” user can be determined.

  Subsequently, the model generation unit 13 uses a feature amount group F11 of the user belonging to the target predetermined life stage as an explanatory variable, uses life event information LI as correct information as an objective variable, and uses a well-known method such as predetermined machine learning. As a result, a first conditional life event estimation model M11 that outputs an estimated value related to the occurrence of a predetermined life event when it belongs to the target life stage is generated. In the present embodiment, the first conditional life event estimation model M11 outputs a probability of occurrence of a predetermined life event when belonging to a target life stage.

  The model generation unit 13 uses a feature amount group F12 of a user who does not belong to a predetermined life stage as an explanatory variable, uses life event information LI as correct information as an objective variable, and uses a known method such as predetermined machine learning. As a result, a second conditional life event estimation model M12 that outputs an estimated value relating to the occurrence of a predetermined life event when the life event does not belong to the target life stage is generated. In the present embodiment, the second conditional life event estimation model M12 outputs the probability of occurrence of a predetermined life event when it does not belong to the target life stage.

  More specifically, for example, the model generation unit 13 outputs the feature amount group F11 of the “unmarried” user and the life event as correct information indicating whether or not the life event of “home purchase” has occurred in the user. Based on the information LI, it is possible to generate a first conditional life event estimation model M11 composed of a calculation formula that outputs the probability of occurrence of a life event of “house purchase” in a “married” user. In addition, the model generation unit 13 calculates the feature amount group F12 of the user “married” and the life event information LI as the correct answer information indicating whether the life event “purchase home” has occurred in the user. It is possible to generate a second conditional life event estimation model M12 composed of a calculation formula that outputs the probability of occurrence of a life event of “house purchase” for a “married” user.

  In this way, in view of the fact that the occurrence of a life event is greatly affected by the life stage to which the user belongs, the life stage to which the user belongs is estimated. Since the event estimation model is generated, it is possible to estimate the occurrence of an appropriate life event taking into account the life stage to which the user belongs.

  Next, the output of the estimated value regarding the occurrence of the life event in consideration of the life stage will be described. FIG. 8 is a diagram illustrating the calculation of the probability of occurrence of a life event in consideration of a life stage.

  First, the estimating unit 15 acquires a feature amount group F2 generated based on input information of a user whose life event is to be estimated. Then, the estimating unit 15 uses the life stage estimation model M10 that outputs the probability that the user belongs to the target life stage, and based on the feature amount group F2 of the user, the probability p ( S) is estimated (P10).

  More specifically, for example, based on the feature amount group F2 of the target user, the estimating unit 15 generates a life stage estimation model M10 including a calculation formula that outputs the probability that the user belongs to the life stage of “unmarried”. The probability p (S) that the user is “unmarried” can be estimated using the above.

  In addition, the estimating unit 15 uses the first conditional life event estimation model M11 that outputs the probability of occurrence of a predetermined life event when belonging to the target life stage, based on the user's feature amount group F2, Estimate the first occurrence probability p (E | S) of a predetermined life event in the case where P belongs to the target life stage (P11).

  More specifically, the estimating unit 15 includes a first calculation formula that outputs a probability of occurrence of a life event of “home purchase” for a “unmarried” user based on the feature amount group F2 of the target user. Using the conditional life event estimation model M11, it is possible to estimate the occurrence probability p (E | S) of the life event of "buying a house" in a "unmarried" user.

  In addition, the estimation unit 15 uses the second conditional life event estimation model M12 that outputs the probability of occurrence of a predetermined life event when the life event does not belong to the target life stage, based on the feature amount group F2 of the user. A second occurrence probability p (E | ¬S) of a predetermined life event when the user belongs to the target life stage is estimated (P12).

  More specifically, the estimating unit 15 is based on a feature amount group F2 of the target user, and is based on a calculation formula that outputs the probability of occurrence of a life event of “home purchase” for a “married” user. Using the conditional life event estimation model M12, it is possible to estimate the occurrence probability p (E | ¬S) of the life event “buy house” in the “married” user.

Subsequently, the estimating unit 15 determines, based on the probability p (S) that the user belongs to the predetermined life stage, the first occurrence probability p (E | S), and the second occurrence probability p (E | ¬S). The probability p (E) of occurrence of a predetermined life event in the user is estimated (P20). Specifically, the occurrence probability p (E) is calculated by the following equation (1).
p (E) = p (E | S) * p (S) + p (E | ¬S) * p (¬S) (1)
Then, the estimating unit 15 outputs the estimated occurrence probability p (E) (P30).

  More specifically, the estimation unit 15 calculates the product of the probability p (E | S) of the occurrence of the life event “purchase of home” for the user “unmarried” and the probability p (S) that the user is “unmarried”. The value obtained by adding the value of the product of the probability p (E | ¬S) of occurrence of the life event “home purchase” for the user “married” and the probability p (¬S) that the user is “married” to the value , Can be calculated as the probability p (E) of occurrence of a life event “purchase house” in the target user.

  In this way, in view of the fact that the occurrence of a life event is greatly affected by the life stage to which the user belongs, the probability that the user belongs to the predetermined life stage is estimated, and when the user belongs to the predetermined life stage, Since the probability of occurrence of a predetermined life event is estimated based on each life event estimation model in the case where there is no life event, it is possible to estimate the probability of occurrence of an appropriate life event in consideration of the life stage to which the user belongs.

  In the example described with reference to FIGS. 7 and 8, when the target life stage is an object separated by the life event to be estimated, the feature amount group F12 and the life event information LI in FIG. Of the second conditional life event estimation model M12 based on the above, calculation of the second occurrence probability P12 using the second conditional life event estimation model M12 in FIG. 8, and occurrence probability of the life event It is possible to omit the second term (p (E | ¬S) * p (¬S)) of the equation (1) for calculating the occurrence probability p (E) in the calculation P20 of (1).

  More specifically, for example, when the life event to be estimated is “marriage” and the life stage used for estimation is “unmarried” / “married”, when the life stage is “married” Since the probability of occurrence of the life event "marriage" can be approximated to zero, the above omission is possible.

  Next, an estimation program for causing a computer to function as the estimation device 10 of the present embodiment will be described. FIG. 9 is a diagram showing a configuration of the estimation program P1.

  The estimation program P1 includes a main module m10, an acquisition module m11, a feature amount generation module m12, a model generation module m13, a model output module m14, an estimation module m15, and a proposal module m16 that comprehensively control the estimation processing in the estimation device 10. It is composed. Each of the modules m11 to m16 implements the functions of the acquisition unit 11, the feature amount generation unit 12, the model generation unit 13, the model output unit 14, the estimation unit 15, and the proposal unit 16 in the estimation device 10. The estimation program P1 may be transmitted via a transmission medium such as a communication line, or may be stored in a recording medium M1 as shown in FIG.

  In the estimating apparatus 10 according to the present embodiment described above, since the information regarding the contract of the mobile terminal is used for generating the feature, it is possible to obtain the feature for inputting to the model based on the easily obtainable information. it can. Since the information related to the contract of the mobile terminal is information that reflects the aspect of the user's life, a model that outputs an estimated value related to the occurrence of a life event using a feature amount generated based on such information as an input is used. By generating, it is possible to accurately estimate the life event of the user.

  Further, in the estimation device according to another aspect, the feature amount generation unit may include a feature amount generation unit configured to calculate a numerical value of the contract information at one time point or a predetermined period in a time series based on the input information. Information including at least one of information indicating a value and an attribute, and information indicating a numerical value, an aggregate value, or an attribute of contract information in a first time point or a first period in a time series; Generating, as a feature value, information including at least one of a numerical value, a total value, and information indicating a change from information indicating an attribute in a second time point before or in a second period before the first period. It may be.

  According to the above aspect, the value included in the contract information at a certain time or a certain period and information indicating a change in the value included in the contract information between two time or two periods are input to the model. Used as Therefore, it is possible to obtain a feature amount in which the manner of the user's life and the change in the manner are appropriately reflected.

  In the estimation device according to another aspect, the feature amount generation unit estimates a group to which the user belongs based on the input information, and outputs information indicating the group, and predetermined input information regarding other users belonging to the same group. May be generated as a part of the feature amount of the user.

  The occurrence of a life event may be affected by the group to which the user belongs and other users in the group. According to the above aspect, the group to which the user belongs is estimated, and the information on the group and the information on the other users in the group are included in the feature amount. Therefore, it is possible to generate a model with high estimation accuracy of the occurrence of the life event. .

  In the estimation device according to another aspect, the acquisition unit acquires life stage information indicating a life stage that is a stage in a living environment separated by a life event in association with the input information and the user, and the model generation unit includes: As a life event estimation model, based on the feature amount and the life stage information, a life stage estimation model is generated in which a user's feature amount is input and an estimated value regarding whether or not the user belongs to a predetermined life stage is output. Generating a first conditional life event estimation model that outputs a feature value of a user belonging to a predetermined life stage as an input and outputs an estimated value related to occurrence of a predetermined life event based on the feature value; Based on the characteristics of a user who does not belong to a predetermined life stage, It may generate the second condition life event estimation model with a to output the estimate for preparative generation.

  The occurrence of a life event is greatly affected by the life stage to which the user belongs. According to the above aspect, the life stage to which the user belongs is estimated, and a life event estimation model is generated for each of the case where the user belongs to the predetermined life stage and the case where the user does not belong. It is possible to estimate the occurrence of a life event.

  Further, in the estimation device according to another aspect, the acquisition unit acquires event information that is information associated with the input information and is an event relating to the user that has occurred at a certain time, and the model generation unit A life event estimation model may be generated based on feature amounts and event information by supervised learning as correct information.

  According to the above aspect, it is possible to generate a life event estimation model that outputs an accurate estimated value regarding the occurrence of a life event.

  Further, the estimation device according to an aspect of the present invention is an estimation device for estimating a life event of a user, the input information including at least contract information on a contract of a mobile terminal, and a user for whom the life event is estimated. An acquisition unit that acquires input information related to the input unit; a feature amount generation unit that generates a feature amount to be input to a life event estimation model for estimating a life event of a user based on the input information; Using a life event estimation model that outputs an estimated value regarding the occurrence of a predetermined life event as an output, based on the feature amount generated by the feature amount generation unit, outputting an estimated value regarding the occurrence of the predetermined life event regarding the user And a unit.

  According to the above aspect, since the information on the contract of the mobile terminal is used for generating the feature amount, the feature amount to be input to the life event estimation model can be obtained based on the easily obtainable information. Since the information related to the contract of the mobile terminal is information that reflects the aspect of the user's life, by using a model in which a feature amount generated based on such information is used as an input, the life event of the user is It is possible to accurately estimate the occurrence probability.

  In the estimation device according to another aspect, the life event estimation model outputs the estimated value regarding whether or not the user belongs to a life stage representing a stage in a living environment divided by the life event, with the feature amount being an input. A life stage estimation model, a first conditional life event estimation model that receives a feature amount of a user belonging to a predetermined life stage as an input, and outputs a first estimated value related to occurrence of a predetermined life event, and a predetermined life And a second conditional life event estimation model that outputs a feature value of a user who does not belong to the stage and outputs a second estimated value related to the occurrence of a predetermined life event. To determine whether the user belongs to a predetermined life stage based on the user's feature value. The first estimated value for the user is calculated based on the feature amount of the user using the first conditional life event estimation model, and the second estimated life event estimation model is calculated using the second conditional life event estimation model. Calculating a second estimated value for the user based on the feature amount of the user, and calculating an estimated value for whether the user belongs to a predetermined life stage, a first estimated value for the user, and a second estimated value for the user. Based on this, an estimated value related to the occurrence of a predetermined life event for the user may be output.

  The occurrence of a life event is greatly affected by the life stage to which the user belongs. According to the above aspect, the probability that the user belongs to the predetermined life stage is estimated, and the generation of the predetermined life event is performed based on the life event estimation models when the user belongs to the predetermined life stage and when the user does not belong to the predetermined life stage. Since the probability is estimated, it is possible to estimate an appropriate probability of occurrence of a life event taking into account the life stage to which the user belongs.

  In addition, an estimation device according to another aspect is configured such that, based on an association between a life event stored in advance and a predetermined product or service, a user is provided with an estimated value related to the occurrence of the predetermined life event estimated by the estimation unit. May be further provided with a suggestion unit that outputs information on a product or service to be proposed to the user.

  According to the above embodiment, it is possible to appropriately recommend a product or the like to the user in accordance with the estimated life event occurrence probability.

  Although the present embodiment has been described in detail, it is obvious to those skilled in the art that the present embodiment is not limited to the embodiment described in this specification. This embodiment can be implemented as modified and changed aspects without departing from the spirit and scope of the present invention defined by the description of the claims. Therefore, the description in the present specification is for the purpose of illustrative explanation, and does not have any restrictive meaning to the present embodiment.

  The processing procedure, sequence, flowchart, and the like of each aspect / embodiment described in this specification may be interchanged as long as there is no inconsistency. For example, the methods described herein present elements of various steps in an exemplary order, and are not limited to the specific order presented.

  Information and the like can be output from an upper layer (or lower layer) to a lower layer (or upper layer). Input and output may be performed via a plurality of network nodes.

  The input and output information and the like may be stored in a specific place (for example, a memory) or may be managed by a management table. Information that is input and output can be overwritten, updated, or added. The output information or the like may be deleted. The input information or the like may be transmitted to another device.

  The determination may be made based on a value represented by 1 bit (0 or 1), a Boolean value (Boolean: true or false), or a comparison of numerical values (for example, a predetermined value). Value).

  Each aspect / embodiment described in the present specification may be used alone, may be used in combination, or may be used by switching with execution. In addition, the notification of the predetermined information (for example, the notification of “X”) is not limited to explicitly performed, and is performed implicitly (for example, not performing the notification of the predetermined information). Is also good.

  Software, regardless of whether it is called software, firmware, middleware, microcode, a hardware description language, or any other name, instructions, instruction sets, codes, code segments, program codes, programs, subprograms, software modules , Applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, functions, and the like.

  In addition, software, instructions, and the like may be transmitted and received via a transmission medium. For example, the software may use a wired technology such as coaxial cable, fiber optic cable, twisted pair and digital subscriber line (DSL) and / or a web site, server, or other using wireless technology such as infrared, wireless and microwave. When transmitted from a remote source, these wired and / or wireless technologies are included within the definition of transmission medium.

  The information, signals, etc. described herein may be represented using any of a variety of different technologies. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc., that can be referred to throughout the above description are not limited to voltages, currents, electromagnetic waves, magnetic or magnetic particles, optical or photons, or any of these. May be represented by a combination of

  Note that terms described in this specification and / or terms necessary for understanding this specification may be replaced with terms having the same or similar meaning.

  As used herein, the terms “system” and “network” are used interchangeably.

  Further, the information, parameters, and the like described in this specification may be represented by an absolute value, may be represented by a relative value from a predetermined value, or may be represented by another corresponding information. .

  The phrase "based on" as used herein does not mean "based solely on" unless stated otherwise. In other words, the description "based on" means both "based only on" and "based at least on."

  Where the designations "first," "second," etc., are used herein, any reference to the element does not generally limit the quantity or order of the element. These designations may be used herein as a convenient way to distinguish between two or more elements. Thus, a reference to a first and second element does not mean that only two elements may be employed therein, or that the first element must somehow precede the second element.

  As long as “include”, “including”, and variations thereof, are used in the present description or claims, these terms are equivalent to the term “comprising” It is intended to be comprehensive. Further, it is intended that the term "or", as used herein or in the claims, not be the exclusive OR.

  In this specification, a plurality of devices is also included unless the device is clearly or only one exists in context or technically.

  Throughout this disclosure, the singular encompasses the plural unless the context clearly dictates otherwise.

  DESCRIPTION OF SYMBOLS 10 ... Estimation apparatus, 11 ... Acquisition part, 12 ... Feature amount generation part, 13 ... Model generation part, 14 ... Model output part, 15 ... Estimation part, 16 ... Proposal part, 20 ... Input information storage part, 30 ... Model storage Unit, 40: proposal information storage unit, M1: recording medium, m10: main module, m11: acquisition module, m12: feature amount generation module, m13: model generation module, m14: model output module, m15: estimation module, m16 ... Suggestion module.

Claims (8)

An estimation device for estimating a life event of a user,
An acquisition unit that acquires input information for each user including contract information on at least a contract of the mobile terminal,
Based on the input information, a feature amount generation unit that generates a feature amount for input to a life event estimation model for estimating a life event of a user,
A model generation unit configured to generate the life event estimation model that outputs an estimated value related to the occurrence of a predetermined life event, based on the characteristic amount, using the characteristic amount as input;
A model output unit that outputs the generated life event estimation model,
An estimation device comprising: The feature amount generation unit includes:
Based on the input information,
Numerical value of the contract information at one time or a predetermined period in a time series, information including at least one of information indicating an aggregated value and an attribute of the numerical value, and
A numerical value of the contract information at the first time point or the first period in the time series, information indicating the total value or the attribute, and a second time point before the first time point or before the first time period. Generating, as the feature amount, information including at least one of the numerical value, the aggregation value, and information indicating a change from the information indicating the attribute in a second period;
The estimating device according to claim 1. The feature amount generation unit includes:
Estimating the group to which the user belongs based on the input information,
Information indicating the group, and at least one of predetermined input information regarding other users belonging to the same group, as a part of the feature amount of the user,
The estimation device according to claim 1. The acquisition unit acquires life stage information indicating a life stage that is a stage in a living environment separated by the life event in association with the input information and a user,
The model generation unit, as the life event estimation model,
Based on the feature value and the life stage information, the feature value of the user is input, and an estimated value regarding whether the user belongs to the predetermined life stage is output, and a life stage estimation model is generated.
On the basis of the feature amount, a first conditional life event estimation model is generated, in which a feature amount of a user belonging to a predetermined life stage is input and an estimated value related to occurrence of a predetermined life event is output,
Based on the characteristic amount, a characteristic amount of a user who does not belong to the predetermined life stage is input, and a second conditional life event estimation model that outputs an estimated value related to occurrence of the predetermined life event is generated.
The estimating device according to claim 1. The acquisition unit is information associated with the input information, and acquires event information that is an event related to the user that has occurred at a certain time,
The model generation unit generates the life event estimation model based on the feature amount and the event information by supervised learning using the event information as correct information.
The estimating device according to claim 1. An estimation device for estimating a life event of a user,
An acquisition unit that is at least input information including contract information on a contract of a mobile terminal, and acquires input information on a user whose life event is to be estimated,
Based on the input information, a feature amount generation unit that generates a feature amount to be input to a life event estimation model for estimating a life event of a user,
Using a life event estimation model that outputs an estimated value related to the occurrence of a predetermined life event with the characteristic amount as an input, based on the characteristic amount generated by the characteristic amount generation unit, a predetermined life event related to the user. An estimating unit that outputs an estimated value regarding occurrence of
An estimation device comprising: The life event estimation model,
A life stage estimation model that receives the feature amount as input, and outputs an estimated value regarding whether the user belongs to a life stage representing a stage in a living environment separated by the life event,
A first conditional life event estimation model in which a feature amount of a user belonging to a predetermined life stage is input and a first estimated value related to occurrence of a predetermined life event is output, and
A second conditional life event estimation model that receives a feature amount of a user who does not belong to the predetermined life stage as input and outputs a second estimated value related to occurrence of the predetermined life event,
The estimating unit,
Using the life stage estimation model, based on the feature amount of the user, output an estimated value regarding whether the user belongs to the predetermined life stage,
Using the first conditional life event estimation model, calculating the first estimated value for the user based on the feature amount of the user;
Using the second conditional life event estimation model, calculating the second estimated value for the user based on the feature amount of the user,
An estimated value related to the occurrence of the predetermined life event related to the user based on the estimated value related to whether the user belongs to the predetermined life stage or not, the first estimated value related to the user, and the second estimated value related to the user; Output
The estimating device according to claim 6. Based on an association between a life event and a predetermined product or service stored in advance, a product or service to be proposed to the user according to an estimated value related to the occurrence of the predetermined life event estimated by the estimation unit. The estimating device according to claim 6, further comprising a suggestion unit that outputs information.

Images