JP7157681B2 - ATTRIBUTE ESTIMATION APPARATUS, METHOD AND PROGRAM - Google Patents

Description

The present invention relates to technology for estimating attributes of residents.

Conventionally, customer attribute information (for example, residence type, family composition, presence or absence of children/elderly, employment type, frequency of going out, etc.) is useful information necessary for various marketing and service provision, but acquisition It wasn't always easy.

Therefore, there is a technique for estimating customer attribute information. For example, Patent Literature 1 discloses estimating attribute information of a household based on information indicating the power consumption of the household. Specifically, when deep learning is performed on the information indicating the power usage of each of multiple households, the power usage feature values obtained by deep learning and the power usage of each household attribute information (see (step S708) in paragraph [0022] of Patent Document 1). Then, based on this association, it is possible to estimate the attribute information associated with the feature quantity of the power consumption of the household whose attribute information is unknown (see (step S1306) of paragraph [0043] of Patent Document 1). .

Japanese Patent Application Laid-Open No. 2018-132802

However, in Patent Literature 1, all attribute information is estimated. Specifically, the amount of electricity usage is associated with all attribute information of the household from which the electricity usage is obtained, and based on the electricity usage alone, Estimate attribute information.

Currently, some companies that supply energy such as gas and electricity (hereinafter also referred to as energy suppliers) have only partial attribute information of their customers. However, in Patent Literature 1, even if some of the attribute information has already been acquired, some of the acquired attribute information is not utilized.

Accordingly, an object of the present invention is to estimate unacquired attribute information based on the usage amount of energy such as gas and electricity (hereinafter also referred to as energy usage amount) and some attribute information that has already been acquired. do.

One aspect of the present invention includes a usage acquisition unit that acquires energy usage, an attribute information acquisition unit that acquires first attribute information, an energy usage and first attribute information, and second attribute information. an estimation unit that estimates the second attribute information based on the obtained energy usage amount and the obtained first attribute information, and the correspondence relationship; Prepare.

According to the present invention, unacquired attribute information can be estimated based on the amount of energy used and some acquired attribute information.

1 is an overall configuration diagram including an attribute estimation device according to an embodiment of the present invention; FIG. 1 is a functional block diagram of an attribute estimation device according to one embodiment of the present invention; FIG. It is an example of pre-estimation data stored in a pre-estimation data storage unit according to one embodiment of the present invention. FIG. 4 is a diagram for explaining estimation according to an embodiment of the present invention; FIG. It is a figure for demonstrating the correspondence which concerns on one Embodiment of this invention. It is an example of post-estimation data stored in a post-estimation data storage unit according to one embodiment of the present invention. 4 is a flowchart of attribute information estimation processing according to an embodiment of the present invention; 6 is a flowchart of list output processing according to one embodiment of the present invention. It is a block diagram which shows the hardware constitutions of the attribute estimation apparatus which concerns on one Embodiment of this invention.

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below based on the drawings.

FIG. 1 is an overall configuration diagram including an attribute estimation device 100 according to one embodiment of the present invention. The overall configuration diagram may include an attribute estimation device 100, one or more measuring instruments 101, one or more resident terminals 102, one or more worker terminals 103, and one or more list output terminals 104. . Each of these will be described below.

The attribute estimation device 100 is a device for estimating attribute information of customers (hereinafter also referred to as residents) who use energy such as gas and electricity. Specifically, the attribute estimation device 100 calculates the attribute information that has not been acquired based on the energy usage amount acquired from the measuring device 101 and some attribute information acquired from the resident terminal 102 and the worker terminal 103. It is a device for estimating. The attribute estimation device 100 consists of one or more computers. The attribute estimation device 100 can transmit and receive data to and from the measuring device 101 via any network 105 . Also, the attribute estimation device 100 can transmit and receive data to and from the resident terminal 102 via any network 105 . Also, the attribute estimation device 100 can transmit and receive data to and from the worker terminal 103 via an arbitrary network 105 . Also, the attribute estimation device 100 can transmit and receive data to and from the list output terminal 104 via any network 105 . The attribute estimation device 100 will be described in detail later with reference to FIG.

The measuring device 101 is a device (for example, a smart meter or the like) installed in a resident's house for measuring the amount of energy used. The measuring device 101 can transmit and receive data to and from the attribute estimation device 100 via any network 105 . Specifically, the measuring device 101 transmits data of the measured energy usage 111 to the attribute estimation device 100 .

A resident terminal 102 is a terminal used by a resident. The resident terminal 102 is a computer such as a personal computer, tablet, or smart phone. The resident terminal 102 can transmit and receive data to and from the attribute estimation device 100 via any network 105 . Specifically, the resident terminal 102 transmits data of the resident's attribute information 121 input by the resident to the resident terminal 102 to the attribute estimation device 100 . For example, a resident inputs attribute information of the resident through a questionnaire, membership registration, or the like.

The worker terminal 103 is a terminal used by a person (for example, a worker of an energy supplier, etc.) who has a chance to come into contact with a resident by visiting the resident's house. The worker terminal 103 is a computer such as a personal computer, tablet, or smart phone. The worker terminal 103 can transmit and receive data to and from the attribute estimation device 100 via any network 105 . Specifically, the worker terminal 103 transmits data of the resident's attribute information 131 input to the worker terminal 103 by the worker or the like to the attribute estimation device 100 . For example, a worker or the like inputs attribute information of a resident that he or she has seen or heard when visiting the resident's house.

The list output terminal 104 outputs a list of residents having desired attribute information from a database including some attribute information acquired from the resident terminal 102 and the worker terminal 103 and attribute information estimated by the attribute estimation device 100. A terminal for output. The list output terminal 104 is a computer such as a personal computer, tablet, or smart phone. The list output terminal 104 can transmit and receive data to and from the attribute estimation device 100 via any network 105 . Specifically, the list output terminal 104 transmits data of desired attribute information to the attribute estimation device 100 . The list output terminal 104 also receives the data of the list of residents having desired attribute information from the attribute estimation device 100 .

FIG. 2 is a functional block diagram of the attribute estimation device 100 according to one embodiment of the invention. The attribute estimation device 100 includes a usage acquisition unit 201 , an attribute information acquisition unit 202 , an estimation unit 203 , a pre-estimation data storage unit 205 , a correspondence storage unit 206 and a post-estimation data storage unit 207 . Note that the attribute estimation device 100 can also be configured to include the list output unit 204 . Also, the attribute estimation device 100 functions as a usage amount acquiring unit 201, an attribute information acquiring unit 202, and an estimating unit 203 by executing programs. Note that the attribute estimation device 100 can also be configured to function as the list output unit 204 by executing a program. Each of these will be described below.

The usage acquisition unit 201 acquires energy usage data of residents. Specifically, the usage acquisition unit 201 can acquire energy usage data of the resident from the measuring device (for example, smart meter or the like) 101 installed in the resident's house. In addition, the usage acquisition unit 201 causes the pre-estimation data storage unit 205 to store the acquired energy usage data of the resident.

It should be noted that the usage acquisition unit 201 can also be configured to acquire data on the amount of energy used by the residents from a computer in which the energy supplier manages the amount of energy used.

The attribute information acquisition unit 202 acquires data of resident attribute information. Specifically, the attribute information acquisition unit 202 acquires attribute information data of the resident from at least one of the resident terminal 102 and the worker terminal 103 . In addition, the attribute information acquisition unit 202 causes the pre-estimation data storage unit 205 to store the acquired attribute information data of the resident.

The attribute information acquisition unit 202 acquires the attribute information of the resident from a computer other than the resident terminal 102 and the worker terminal 103 (for example, acquires the attribute information of the resident in the external database). can also

The pre-estimation data storage unit 205 associates and stores the energy usage data of the resident acquired by the usage acquisition unit 201 and the attribute information data of the resident acquired by the attribute information acquisition unit 202. ing. The pre-estimation data 300 stored in the pre-estimation data storage unit 205 will be described in detail with reference to FIG.

FIG. 3 is an example of pre-estimation data 300 stored in the pre-estimation data storage unit 205 according to one embodiment of the present invention. As shown in FIG. 3, the pre-estimation data 300 includes, for example, data on items such as residents, energy consumption, and attribute information (for example, residence type, family composition, children, elderly people, employment type, frequency of going out). include.

"Resident" is information for identifying a resident (for example, a customer number given to a resident by an energy supplier, a resident name, etc.). Energy usage and attribute information are linked by information for identifying residents.

"Energy Usage" is the energy usage of the resident. Energy usage may be energy usage in any period (eg, monthly energy usage, yearly energy usage, etc.). Also, a plurality of energy usage amounts (for example, for n months, for n years, etc.) may be stored. Also, a plurality of types of energy usage (for example, gas usage and electricity usage) may be stored.

It should be noted that an energy charge value may be used instead of the amount of energy used.

"Attribute information" is attribute information of the resident. Attribute information may be any information about the resident. For example, attribute information includes housing type (e.g., detached house or apartment complex), family structure (e.g., number of family members), presence/absence of children, presence/absence of elderly people, employment status (e.g., double-income or single-income), frequency of going out, Presence or absence of pets, housing facilities (for example, presence or absence of predetermined gas appliances, etc.), age (for example, age of each resident), and the like.

In addition, in FIG. 3, the shaded portions (that is, resident A's "elderly", resident B's "outing frequency", resident C's "children" and "elderly", resident D's "family composition" “Employment status” and “residential status” of resident E) are unacquired attribute information. These unacquired attribute information are estimated.

Return to FIG. Estimation unit 203, based on the energy usage amount acquired by usage amount acquisition unit 201 and the attribute information acquired by attribute information acquisition unit 202 (that is, attribute information of a part of the resident), attribute information that has not been acquired to estimate Specifically, the estimating unit 203 estimates unacquired attribute information based on the energy usage amount, part of the attribute information, and the correspondence in the correspondence storage unit 206 . First, estimation and correspondence will be described with reference to FIGS. 4 and 5. FIG.

The "attribute information acquired by the attribute information acquiring unit 202" is also called first attribute information, and the estimated "unacquired attribute information" is also called second attribute information. Also, the resident attribute information can include first attribute information and second attribute information. That is, the resident's attribute information may or may not include attribute information other than the first attribute information and the second attribute information.

FIG. 4 is a diagram for explaining estimation according to one embodiment of the present invention.
- First, pre-estimation data 300 (that is, energy usage 111 and some attribute information 121, 131) are acquired.
・Next, the estimation unit 203 estimates attribute information other than some attribute information (that is, unacquired attribute information) 400 based on the pre-estimation data 300 and the correspondence in the correspondence storage unit 206. do.
- Next, post-estimation data 600 including pre-estimation data 300 (that is, energy usage 111 and some attribute information 121 and 131) and attribute information 400 other than some attribute information is generated.

FIG. 5 is a diagram for explaining correspondence relationships according to an embodiment of the present invention. The correspondence storage unit 206 stores the correspondence between "energy usage and partial attribute information" and "attribute information other than the partial attribute information". The correspondence relationship may be a learned model that inputs "the amount of energy used and some attribute information" and outputs "attribute information other than the part of the attribute information". Alternatively, the correspondence relationship may be defined by a rule base for deriving "attribute information other than the partial attribute information" from "the energy usage amount and the partial attribute information".

Hereinafter, the estimation unit 203 will be described separately for a machine learning method (<machine learning> below) and a statistical method (<statistical method> below), also referring to FIG.

<Machine learning>
For example, machine learning creates a model that learns the correspondence between variables and infers unknown variables based on each resident's energy usage and their attribute information.

The estimation unit 203 can estimate unacquired attribute information by inputting a resident's energy usage and a part of the resident's attribute information into the model.

<Statistical method>
The estimation unit 203 can estimate "attribute information other than the partial attribute information" from the "energy usage amount and the partial attribute information" by a statistical technique such as regression analysis instead of machine learning. Specifically, for example, a logistic model with arbitrary attribute information as the objective variable and energy consumption as the explanatory variable is created for each group segmented by the known attribute information of the residents, and the attribute information is estimated. It is possible.

In this way, the estimation unit 203 can estimate based on a learned model (correspondence relationship) generated by machine learning, or can estimate based on a correspondence relationship defined by a statistical method. can.

The estimation unit 203 can be configured to indicate the estimated value with the highest probability (that is, to indicate one attribute information that has been estimated), or to indicate the estimated value probabilistically (that is, to indicate a plurality of estimated attributes). attribute information is shown together with the probability).

Return to FIG. As described above, the correspondence storage unit 206 stores the correspondence between "energy usage and partial attribute information" and "attribute information other than the partial attribute information".

Post-estimation data storage unit 207 stores pre-estimation data 300 (that is, energy usage 111 and partial attribute information 121 and 131) and attribute information 400 other than the partial attribute information estimated by estimation unit 203. It's linked and stored. The post-estimation data 600 stored in the post-estimation data storage unit 207 will be described in detail with reference to FIG.

FIG. 6 is an example of post-estimation data 600 stored in the post-estimation data storage unit 207 according to one embodiment of the present invention. As shown in FIG. 6, the shaded portion in FIG. 3 (that is, unacquired attribute information) is estimated.

The list output unit 204 outputs a list of residents having desired attribute information in response to a request from the list output terminal 104 . Specifically, the list output unit 204 receives data of desired attribute information from the list output terminal 104 . The list output unit 204 also extracts a list of residents having desired attribute information from the post-estimation data 600 . The list output unit 204 also transmits the extracted list data to the list output terminal 104 .

FIG. 7 is a flowchart of attribute information estimation processing according to one embodiment of the present invention.

At step 701 (S701), the usage acquisition unit 201 acquires energy usage data of the resident. The usage acquisition unit 201 causes the pre-estimation data storage unit 205 to store the acquired energy usage data of the resident.

At step 702 (S702), the attribute information acquisition unit 202 acquires attribute information data of the resident. The attribute information acquisition unit 202 stores the acquired attribute information data of the resident in the pre-estimation data storage unit 205 . The order of S701 and S702 may be reversed.

At step 703 (S703), the estimation unit 203 estimates unacquired attribute information based on the energy usage amount acquired at S701 and some attribute information acquired at S702.

FIG. 8 is a flowchart of list output processing according to one embodiment of the present invention.

At step 801 ( S801 ), the list output unit 204 acquires desired attribute information from the list output terminal 104 .

At step 802 (S802), the list output unit 204 outputs a list of residents having the attribute information acquired at S801.

<Updating/adding attribute information>
In one embodiment of the invention, attribute information can be updated. For example, when there is a change in energy usage (for example, a sudden increase or decrease in energy usage), the attribute information can be estimated and updated based on the energy usage after the change. In this way, in one embodiment of the present invention, attribute information is estimated based on the "energy usage amount" that can be reliably acquired periodically, so attribute information that changes over time can be estimated and updated. .

In one embodiment of the invention, attribute information can be added. For example, when an energy supplier or the like sends direct mail, e-mail, etc. to residents based on the list output by the list output unit 204, these results (e.g., presence or absence of conversion, opening/clicking ) can be added to the post-estimation data 600 . Residents who have similar reactions to direct mail, e-mail, etc. can be presumed to have the same attribute information. Thus, in one embodiment of the present invention, new attribute information (ie, responses to direct mail, e-mail, etc.) can be acquired and added.

FIG. 9 is a block diagram showing the hardware configuration of the attribute estimation device 100 according to one embodiment of the invention. The attribute estimation device 100 has a CPU (Central Processing Unit) 1 , a ROM (Read Only Memory) 2 and a RAM (Random Access Memory) 3 . CPU1, ROM2, and RAM3 form a so-called computer.

The attribute estimation device 100 also has an auxiliary storage device 4 , a display device 5 , an operation device 6 , an I/F (Interface) device 7 and a drive device 8 . Each piece of hardware of the attribute estimation device 100 is interconnected via a bus 9 .

The CPU 1 is an arithmetic device that executes various programs installed in the auxiliary storage device 4 .

ROM2 is a non-volatile memory. The ROM 2 functions as a main storage device for storing various programs, data, etc. necessary for the CPU 1 to execute various programs installed in the auxiliary storage device 4 . Specifically, the ROM 2 functions as a main storage device that stores boot programs such as BIOS (Basic Input/Output System) and EFI (Extensible Firmware Interface).

The RAM 3 is a volatile memory such as a DRAM (Dynamic Random Access Memory) or an SRAM (Static Random Access Memory). The RAM 3 functions as a main storage device that provides a work area to be expanded when various programs installed in the auxiliary storage device 4 are executed by the CPU 1 .

The auxiliary storage device 4 is an auxiliary storage device that stores various programs and information used when various programs are executed.

The display device 5 is a display device that displays the internal state of the attribute estimation device 100 and the like.

The operation device 6 is an input device through which the administrator of the attribute estimation device 100 inputs various instructions to the attribute estimation device 100 .

The I/F device 7 is a communication device that connects to the network 105 and communicates with the measuring device 101 , the resident terminal 102 , the worker terminal 103 and the list output terminal 104 .

A drive device 8 is a device for setting the recording medium 10 . The recording medium 10 here includes media for optically, electrically or magnetically recording information, such as CD-ROMs, flexible disks, and magneto-optical disks. The recording medium 10 may also include a semiconductor memory that electrically records information such as an EPROM (Erasable Programmable Read Only Memory), a flash memory, or the like.

Various programs to be installed in the auxiliary storage device 4 are installed by, for example, setting the distributed recording medium 10 in the drive device 8 and reading out the various programs recorded in the recording medium 10 by the drive device 8. be done. Alternatively, various programs to be installed in the auxiliary storage device 4 may be installed by being downloaded from another network different from the network 105 via the I/F device 7 .

In this way, in the present invention, it is possible to estimate unacquired attribute information based on both "energy usage" without omissions and "already acquired attribute information" with omissions. . As a result, a database of all attribute information can be generated.

Although the embodiments of the present invention have been described in detail above, the present invention is not limited to the specific embodiments described above, and various modifications can be made within the scope of the gist of the present invention described in the claims.・Changes are possible.

100 Attribute estimation device 101 Measuring device 111 Energy usage 102 Resident terminal 121 Attribute information 103 Worker terminal 131 Attribute information 104 List output terminal 105 Network 201 Usage acquisition unit 202 Attribute information acquisition unit 203 Estimation unit 204 List output unit 205 Estimation Pre-data storage unit 206 Correspondence storage unit 207 Post-estimation data storage unit 300 Pre-estimation data 400 Attribute information other than some attribute information 500 Machine learning or rule base 501 Energy consumption 502 Partial attribute information 503 Partial attributes Attribute information other than information 600 Post-estimation data

Claims (5)

a usage acquisition unit that acquires energy usage;
an attribute information acquisition unit that acquires first attribute information;
A trained model that inputs the energy usage amount and the first attribute information and outputs second attribute information, or derives the second attribute information from the energy usage amount and the first attribute information a correspondence storage unit that stores correspondence defined by a rule base for
An attribute estimation device comprising: an estimation unit that estimates the second attribute information based on the acquired energy usage amount, the acquired first attribute information, and the correspondence relationship. 2. A list output unit for outputting a list of residents having predetermined attribute information from among the list of residents having the acquired first attribute information and the estimated second attribute information. Attribute estimation device according to. The attribute estimation device according to claim 1 or 2, wherein said estimation unit estimates said second attribute information with the highest probability, or estimates said second attribute information probabilistically. A computer-implemented method comprising: obtaining energy usage;
obtaining first attribute information;
A trained model that inputs the acquired energy usage amount and the acquired first attribute information, the energy usage amount and the first attribute information, and outputs second attribute information, or the energy usage estimating said second attribute information based on quantities and correspondences defined by a rule base for deriving said second attribute information from said first attribute information . Get energy usage,
Acquire the first attribute information,
A trained model that inputs the acquired energy usage amount and the acquired first attribute information, the energy usage amount and the first attribute information, and outputs second attribute information, or the energy usage for causing a computer to execute a process of estimating the second attribute information based on the amount and a correspondence defined by a rule base for deriving the second attribute information from the first attribute information; program.

Images