JP2019028488A - Extraction apparatus, extraction method, extraction program, learning data and model - Google Patents

Abstract

To provide an extraction apparatus, an extraction method, an extraction program, a learning data and a model for appropriately extracting information on a user satisfying regularity.SOLUTION: An extraction apparatus 100 includes an acquisition unit 131 and an extraction unit 132. The acquisition unit 131 acquires a plurality of positional information of a user by a function such as a GPS sensor included in a terminal apparatus of the user in a predetermined period. The extraction unit 132 extracts a set of positional information satisfying a predetermined regularity with regard to a change of a position among the plurality of positional information acquired by the acquisition unit.SELECTED DRAWING: Figure 4

Description

  The present invention relates to an extraction device, an extraction method, an extraction program, learning data, and a model.

  Conventionally, a technique for extracting information related to user behavior has been provided. For example, there is provided a technique for indirectly estimating a facility where a user has stopped or stayed while using history information obtained intermittently with the user's behavior or movement.

JP 2010-244433 A

  However, in the above-described conventional technology, it may be difficult to appropriately extract user information that satisfies regularity. For example, it may be difficult to appropriately extract, for example, information related to a user's continuous behavior by using history information obtained intermittently with the user's behavior and movement.

  The present application has been made in view of the above, and an object thereof is to provide an extraction device, an extraction method, an extraction program, learning data, and a model that appropriately extract user information satisfying regularity.

  The extraction device according to the present application has an acquisition unit that acquires a plurality of pieces of position information of a user detected in a predetermined period, and a predetermined regularity regarding a change in position among the plurality of pieces of position information acquired by the acquisition unit. And an extraction unit for extracting a position information group to be satisfied.

  According to one aspect of the embodiment, it is possible to appropriately extract user information that satisfies regularity.

FIG. 1 is a diagram illustrating an example of an extraction process according to the embodiment. FIG. 2 is a diagram illustrating an example of the extraction process according to the embodiment. FIG. 3 is a diagram illustrating a configuration example of the extraction system according to the embodiment. FIG. 4 is a diagram illustrating a configuration example of the extraction device according to the embodiment. FIG. 5 is a diagram illustrating an example of a user information storage unit according to the embodiment. FIG. 6 is a diagram illustrating an example of the behavior information storage unit according to the embodiment. FIG. 7 is a diagram illustrating an example of the position information storage unit according to the embodiment. FIG. 8 is a diagram illustrating an example of a model information storage unit according to the embodiment. FIG. 9 is a flowchart illustrating an example of the extraction process according to the embodiment. FIG. 10 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the extraction device.

  Hereinafter, an extraction apparatus, an extraction method, an extraction program, learning data, and a form for implementing a model (hereinafter referred to as “embodiment”) according to the present application will be described in detail with reference to the drawings. In addition, the extraction apparatus, the extraction method, the extraction program, the learning data, and the model according to the present application are not limited by this embodiment. In the following embodiments, the same portions are denoted by the same reference numerals, and redundant description is omitted.

(Embodiment)
[1. (Extraction process)
An example of the extraction process according to the embodiment will be described with reference to FIGS. 1 and 2. 1 and 2 are diagrams illustrating an example of an extraction process according to the embodiment. The example of FIGS. 1 and 2 shows a case where a position information group satisfying a predetermined regularity regarding a change in position is extracted from sensor information (position information) detected by the terminal device 10 used by the user. Moreover, the extraction process shown in FIGS. 1 and 2 includes a process of generating a model using the position information group and information indicating the presence or absence of the user's predetermined action. Note that the extraction process may not include the process of generating a model.

[1-1. (Configuration of extraction system)
First, prior to the description of FIGS. 1 and 2, the extraction system 1 shown in FIG. 3 will be described. As illustrated in FIG. 3, the extraction system 1 includes a terminal device 10 and an extraction device 100. The terminal device 10 and the extraction device 100 are connected via a predetermined network N so as to be communicable by wire or wirelessly. FIG. 3 is a diagram illustrating a configuration example of the extraction system according to the embodiment. Note that the extraction system 1 illustrated in FIG. 3 may include a plurality of terminal devices 10 and a plurality of extraction devices 100.

  The terminal device 10 is an information processing device used by a user. The terminal device 10 is realized by, for example, a smartphone, a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), or the like. FIG.1 and FIG.2 shows the case where the terminal device 10 is a smart phone.

  The terminal device 10 has a function such as a GPS (Global Positioning System) sensor, and detects and acquires user position information (sensor information). Further, the terminal device 10 may estimate and acquire the position information of the user using the position information of the base station with which communication is performed or the radio wave of WiFi (registered trademark) (Wireless Fidelity). Hereinafter, the position information may be simply referred to as “position”.

  Moreover, the terminal device 10 may detect various sensor information not only with a GPS sensor etc. but with various sensors. For example, the terminal device 10 has a function of an acceleration sensor, and may be able to detect and acquire acceleration information (sensor information) in the movement of the user. Further, the terminal device 10 may have various functions such as a temperature sensor and an atmospheric pressure sensor, and may be able to detect and acquire environmental information where the user is placed such as temperature and atmospheric pressure. Moreover, the terminal device 10 may have various functions such as a heart rate sensor, and may be able to detect and acquire a user's biological information. For example, a user using the terminal device 10 may acquire a user's own context information by the terminal device 10 by wearing a wearable device that can communicate with the terminal device 10. For example, a user who uses the terminal device 10 wears a wristband type wearable device that can communicate with the terminal device 10, and the terminal device 10 acquires information on the user's own heartbeat (pulse) by the terminal device 10. It may be possible.

  Moreover, the terminal device 10 receives an operation by a user. In addition, the terminal device 10 transmits various information to the extraction device 100. For example, the terminal device 10 transmits position information, behavior information, and the like to the extraction device 100. Hereinafter, the terminal device 10 may be referred to as a user. That is, hereinafter, the user can be read as the terminal device 10.

  Hereinafter, the user identified by the user ID “U1” may be referred to as “user U1”. Thus, hereinafter, when “user U * (* is an arbitrary numerical value)” is described, this indicates that the user is a user identified by the user ID “U *”. For example, when “user U2” is described, the user is a user identified by the user ID “U2”. For example, the extraction system 1 illustrated in FIG. 3 includes a large number of terminal devices 10 such as the terminal device 10 used by the user U1 and the terminal device 10 used by the user U2.

  A terminal device 10 illustrated in FIG. 1 is an information processing device used by a user U1. The terminal device 10 illustrated in FIG. 2 is an information processing device used by the user U2. The terminal device 10 receives various operations by the user. The terminal device 10 may be referred to as a user. That is, the user can be read as the terminal device 10.

  The extraction apparatus 100 is an information processing apparatus that extracts position information (position information group) that satisfies a predetermined regularity regarding a change in position, from among a plurality of pieces of position information of a user detected in a predetermined period. Further, the extraction device 100 is used to predict whether or not a predetermined action has occurred based on learning data including a position information group and action information corresponding to the position information group (hereinafter also referred to as “target action information”). Generate the model

[1-2. About movement (with position change)
In the example of FIG. 1, the extraction apparatus 100 extracts a position information group that satisfies a predetermined regularity regarding a change in position from among a plurality of pieces of position information of the user detected at dates dt11 to dt18 for the user U1. . Note that the user U1 is a user who uses the terminal device 10, and the extraction device 100 acquires the position of the terminal device 10 by a function such as a GPS sensor included in the terminal device 10. The map information MP1 shown in FIG. 1 schematically shows the position of the user U1. In the map information MP1, the positions of the terminal device 10 (the user who owns the terminal device 10) are indicated by positions LC11 to LC18. If the positions LC11 to LC18 are not distinguished, the position LC may be used. Moreover, although the case where the extraction apparatus 100 acquires the position of the user U1 and the action information of the user U1 in real time will be described below, the extraction apparatus 100 acquires the position of the user U1 and the action information of the user U1 at an appropriate timing. The extraction process may be performed. For example, the extraction device 100 may acquire the position and behavior information of each user periodically (for example, every other day) and perform the extraction process.

  First, the extraction device 100 acquires the position LC11 of the user U1 at the date and time dt11. Then, the extraction device 100 acquires a position LC12 indicating that the user U1 is in the XX park (the point of interest POI1) at the date dt12 after the date dt11. For example, the extraction apparatus 100 may be a point of interest such as a location where various facilities (point of interest) such as XX park (point of interest POI1) shown in FIG. 1 or YY movie theater (point of interest POI2) shown in FIG. It is assumed that various information (hereinafter also referred to as “POI information”) regarding (POI: Point of Interest) is included.

  For example, the extraction device 100 estimates that the user U1 is heading to the XX park by comparing the POI information with the position LC of the user U1. For example, the extraction apparatus 100 acquires a position LC12 indicating that the user U1 is in a position different from the position LC11 at a date and time dt12 later than the date and time dt11 and is in the XX park. For example, the extraction apparatus 100 estimates that the user U1 is heading to the XX park by comparing the POI information with the change in the position LC of the user U1. In the example of FIG. 1, the extraction device 100 estimates that the user U1 is moving toward the XX park (from the bottom to the top in FIG. 1) based on the positions LC11 and LC12. Note that the extraction device 100 may estimate the moving direction of the user based on the acceleration detected by the terminal device 10. Further, the extraction device 100 may estimate movement from the point where the position information was last detected based on the acceleration detected by the terminal device 10 by appropriately using various conventional techniques.

  Then, the extraction device 100 acquires a position LC13 indicating that the user U1 is at a position different from the position LC12 in the XX park at a date and time dt13 later than the date and time dt12. In the example of FIG. 1, the extraction apparatus 100 estimates that the user U1 is moving in the XX park based on the positions LC12 and LC13. Further, the extraction device 100 estimates the moving speed of the user U1 based on the positions LC12 and LC13. In addition, the extraction device 100 may estimate the moving speed of the user based on the acceleration detected by the terminal device 10.

  Then, the extraction device 100 acquires a position LC14 indicating that the user U1 is at a position different from the position LC13 in the XX park at a date and time dt14 later than the date and time dt13. In the example of FIG. 1, the extraction apparatus 100 estimates that the user U1 is moving in the XX park based on the positions LC13 and LC14. Further, the extraction device 100 estimates the moving speed of the user U1 based on the positions LC13 and LC14.

  Then, the extraction device 100 acquires a position LC15 indicating that the user U1 is at a position different from the position LC14 in the XX park at a date and time dt15 later than the date and time dt14. In the example of FIG. 1, the extraction device 100 estimates that the user U1 is moving in the XX park based on the positions LC14 and LC15. Further, the extraction device 100 estimates the moving speed of the user U1 based on the positions LC14 and LC15.

  Then, the extraction apparatus 100 acquires a position LC16 indicating that the user U1 is at a position different from the position LC15 in the XX park at a date and time dt16 after the date and time dt15. In the example of FIG. 1, the extraction apparatus 100 estimates that the user U1 is moving in the XX park based on the positions LC15 and LC16. Further, the extraction device 100 estimates the moving speed of the user U1 from the positions LC15 and LC16.

  Then, the extraction device 100 acquires a position LC17 indicating that the user U1 has left the XX park at a date dt17 later than the date dt16. Then, the extraction device 100 acquires a position LC18 indicating that the user U1 is at the A convenience store (shop SP1) at the date dt18 after the date dt17. In addition, the extraction device 100 acquires action information AC11 indicating that the user U1 purchased W water at the A convenience store at the date dt18.

  The extraction apparatus 100 extracts a position information group satisfying a predetermined regularity regarding a change in position from the plurality of position information LC11 to LC18 of the user detected at the date and time dt11 to dt18 (step S11). For example, the extraction apparatus 100 extracts a position information group that satisfies a predetermined regularity corresponding to “behavior pattern PT1 (running)” as shown in the model information storage unit 124 (see FIG. 8). For example, the extraction device 100 extracts a position information group in which the change in the position information of the user U1 is changing at a speed corresponding to running, such as 10 km / h. In the example of FIG. 1, the extraction device 100 extracts a position information group when the user U1 is running in the XX park (point of interest POI1). For example, the extraction apparatus 100 extracts a position information group in which the change in the position information of the user U1 is moving around the XX park at, for example, 10 km / h.

  In the example of FIG. 1, in the XX park, the movement from the position LC14 to the position LC15 passes through the position LC12 in the circulation direction of the user U1 (the clockwise direction in the example of FIG. 1). The extraction device 100 estimates that the user U1 is running in the XX park based on the movement progress of the user U1 and the speed of the user U1 as described above. Therefore, the extraction device 100 extracts the position information LC12 to LC16 in which the user U1 is moving around the XX park at the date and time dt12 to dt16. In the example of FIG. 1, the extraction device 100 extracts position information LC12 to LC16 in which the user U1 is moving around the XX park at the date and time dt12 to dt16.

  Then, the extraction apparatus 100 extracts an action satisfying a predetermined condition (hereinafter, also referred to as “target action”) from the user action outside the target period, which is the period in which the position information group is detected (step S12). . For example, the extraction apparatus 100 includes target behavior information that is behavior information satisfying a predetermined condition among user behavior information outside the target period (dates dt12 to dt16) in which the position information group (position information LC12 to LC16) is detected. To extract. In the example of FIG. 1, the extraction apparatus 100 includes target behavior information that is behavior information satisfying a predetermined condition among user behavior information within a predetermined period (period PD1) from the end of the target period (date and time dt16). To extract. The extraction apparatus 100 may extract target behavior information that is behavior information satisfying a predetermined condition among user behavior information within a predetermined range from the position LC16 at the end of the target period (date and time dt16). In the example of FIG. 1, the extraction apparatus 100 has behavior information indicating that the user U1 has purchased W water at the A convenience store at a date and time dt18 within a predetermined period (period PD1) from the end of the target period (date and time dt16). AC11 is extracted as target action information. Further, the extraction apparatus 100 uses the behavior information AC11 indicating that the user U1 has purchased W water at the A convenience store within a predetermined range (for example, within 1 km) from the position LC16 at the end of the target period (date dt16). Extract as information. In the example of FIG. 1, an example in which only the information of the user U1 is extracted is shown, but it is assumed that the same information is extracted for a large number of users (for example, 10,000 people).

  Then, the extraction device 100 generates a model based on the position information group and the target behavior information (Step S13). In the example of FIG. 1, the extraction device 100 generates a model M1 based on the position information LC12 to LC16 and the behavior information AC11. For example, the extraction device 100 generates a model M1 that predicts whether the user from whom the position information group has been extracted purchases a beverage. For example, the extraction device 100 generates a model M1 that predicts whether a user purchases a beverage after running. In the example of FIG. 1, the extraction apparatus 100 generates a model M1 using the position information group extracted for a large number of users (for example, 10,000 people) including the user U1.

  For example, the extraction apparatus 100 uses the combination of the position information group extracted as satisfying the regularity of the behavior pattern PT1 and information indicating that the beverage has been purchased before and after the behavior as a positive example, as a model M1. Is generated. For example, the extraction device 100 generates a model M1 by performing a learning process so as to output a score “1” when a position information group as a positive example is input. For example, when the position information group is input, the extraction device 100 outputs a model M1 indicating that the more likely a beverage is purchased after the action of the action pattern PT1 is output, the closer the score is to “1”. Generate. In addition, for example, the extraction device 100 may generate the model M1 by using, as a negative example, a user position information group detected before and after a beverage is not purchased. For example, the extraction apparatus 100 generates a model M1 by performing a learning process so as to output a score “0” when a position information group serving as a negative example is input. For example, when the position information group is input, the extraction device 100 generates a model M1 indicating that the more likely a beverage is purchased, the more a score close to “0” is output. For example, the extraction apparatus 100 uses, as a negative example, a combination of a position information group extracted as satisfying the regularity of the action pattern PT1 and information indicating that a beverage is not purchased before and after the action. M1 may be generated.

  Each model may be generated using various conventional techniques relating to machine learning as appropriate. For example, the model generation may be performed using a technique related to machine learning of supervised learning such as SVM (Support Vector Machine). Further, for example, the model generation may be performed using a technique related to machine learning of unsupervised learning. For example, the generation of the model may be performed using a deep learning technique. For example, the model generation may be performed by appropriately using various deep learning techniques such as DNN (Deep Neural Network), RNN (Recurrent Neural Network), and CNN (Convolutional Neural Network). The description relating to the generation of the model is merely an example, and the generation of the model may be performed by a learning method appropriately selected according to information that can be acquired.

[1-3. About stay (no change in position)
In the example of FIG. 2, the extraction apparatus 100 extracts a position information group that satisfies a predetermined regularity related to a change in position among a plurality of pieces of position information of the user detected at dates dt21 to dt26 for the user U2. . Note that the user U <b> 2 is a user who uses the terminal device 10, and the extraction device 100 acquires the position of the terminal device 10 by a function such as a GPS sensor included in the terminal device 10. The map information MP2 shown in FIG. 2 schematically shows the position of the user U2. In the map information MP2, the positions of the terminal device 10 (the user who owns the terminal device 10) are indicated by positions LC21 to LC23. If the positions LC21 to LC23 are not distinguished, the position LC may be used. Moreover, although the case where the extraction apparatus 100 acquires the position of the user U2 and the action information of the user U2 in real time will be described below, the extraction apparatus 100 acquires the position of the user U2 and the action information of the user U2 at an appropriate timing. The extraction process may be performed. For example, the extraction apparatus 100 may acquire the position and behavior information of each user periodically (for example, every hour) and perform the extraction process.

  First, the extraction device 100 acquires the position LC21 of the user U2 at the date and time dt21. Then, the extraction device 100 acquires a position LC22 indicating that the user U2 is in the YY movie theater (the point of interest POI2) at a date dt22 later than the date dt21.

  For example, the extraction apparatus 100 estimates that the user U2 is heading to the YY movie theater by comparing the POI information with the position LC of the user U2. For example, the extraction apparatus 100 acquires a position LC22 indicating that the user U2 is in a position different from the position LC21 and is in the YY movie hall at a date and time dt22 later than the date and time dt21. For example, the extraction apparatus 100 estimates that the user U2 is heading to the YY movie theater by comparing the POI information with the change in the position LC of the user U2. In the example of FIG. 2, the extraction device 100 estimates that the user U2 is moving toward the YY movie theater (from the bottom to the top in FIG. 2) based on the positions LC21 and LC22. Note that the extraction device 100 may estimate the moving direction of the user based on the acceleration detected by the terminal device 10. Further, the extraction device 100 may estimate movement from the point where the position information was last detected based on the acceleration detected by the terminal device 10 by appropriately using various conventional techniques.

  Then, the extraction device 100 acquires a position LC22 indicating that the user U2 is in the YY movie hall at a date and time dt23 later than the date and time dt22. In the example of FIG. 2, the extraction apparatus 100 estimates that the user U2 is staying in the YY movie hall based on the position LC22 acquired at the date dt22 and the date dt23.

  Then, the extraction device 100 acquires the position LC22 indicating that the user U2 is in the YY movie hall at the date dt24 after the date dt23. In the example of FIG. 2, the extraction apparatus 100 estimates that the user U2 is staying in the YY movie hall from the position LC22 acquired from the date dt22 to the date dt24. Further, the extraction device 100 acquires the position LC22 indicating that the user U2 is in the YY movie hall at the date dt25 after the date dt24. In the example of FIG. 2, the extraction apparatus 100 estimates that the user U2 is staying in the YY movie hall from the position LC22 acquired from the date dt22 to the date dt25. Further, for example, the extraction apparatus 100 estimates that the user U2 is watching a movie in the YY movie hall based on the time between the date dt22 and the date dt25 when the position LC22 of the user U2 is continuously detected. Good. For example, when the time between the date and time dt22 to the date and time dt25 is equal to or greater than a predetermined threshold (for example, 1 hour 30 minutes), the extraction device 100 may estimate that the user U2 is watching a movie in the YY movie hall. . The position LC22 detected in each of the date and time dt22 to the date and time dt25 may be a position within a predetermined range (error range) from the position LC22 as long as it can be estimated that the user U2 is located in the YY movie hall.

  Then, the extraction device 100 acquires a position LC23 indicating that the user U2 has left the YY movie theater at a date dt26 later than the date dt25. In the example of FIG. 2, the extraction apparatus 100 acquires a position LC23 indicating that the user U2 is at the B restaurant (shop SP2) at the date and time dt26. In addition, the extraction device 100 acquires action information AC21 indicating that the user U2 has eaten or eaten at the restaurant B at the date and time dt26.

  The extraction apparatus 100 extracts a position information group satisfying a predetermined regularity regarding a change in position from the plurality of pieces of position information LC21 to LC23 of the user detected at the dates dt21 to dt26 (step S21). For example, the extraction apparatus 100 extracts a position information group that satisfies a predetermined regularity corresponding to “behavior pattern PT2 (stay)” as shown in the model information storage unit 124 (see FIG. 8). For example, the extraction device 100 extracts a position information group in which the change in the position information of the user U2 has not changed over, for example, a predetermined time. Here, the position that has not been changed includes a position that has changed within the error range. In the example of FIG. 2, the extraction device 100 extracts a position information group when the user U2 is staying at the YY movie theater (the point of interest POI2). For example, the extraction apparatus 100 extracts a position information group in which the change in the position information of the user U2 is in the YY movie hall, for example, over a predetermined time.

  In the example of FIG. 2, in the YY movie hall, since the position LC22 is continuously detected between the dates dt22 to dt25, it is estimated that the user U2 is staying in the YY movie hall. Therefore, the extraction apparatus 100 extracts the position information LC22 corresponding to each of the dates dt22 to dt25 indicating that the user U2 is staying in the YY movie hall at the dates dt22 to dt25. In the example of FIG. 2, the extraction apparatus 100 extracts a plurality of pieces of position information LC22 respectively corresponding to the dates dt22 to dt25 indicating that the user U2 is staying in the YY movie hall at the dates dt22 to dt25.

  And the extraction apparatus 100 extracts the target action which satisfy | fills predetermined conditions among the user's action out of the target period which is the period when the positional information group was detected (step S22). For example, the extraction device 100 includes target behavior information that is behavior information satisfying a predetermined condition among user behavior information outside the target period (dates dt22 to dt25) in which the position information group (four position information LC22) is detected. To extract. In the example of FIG. 2, the extraction device 100 includes target behavior information that is behavior information that satisfies a predetermined condition among user behavior information within a predetermined period (period PD2) from the end of the target period (date and time dt25). To extract. For example, the extraction apparatus 100 may extract target behavior information that is behavior information satisfying a predetermined condition among user behavior information within a predetermined range from the position LC22 at the end of the target period (date and time dt25). . In the example of FIG. 2, the extraction apparatus 100 includes behavior information AC21 indicating that the user U2 has eaten or eaten at the B restaurant from the end of the target period (date and time dt25) to a date and time dt26 within a predetermined period (period PD2). Is extracted as target behavior information. In addition, the extraction apparatus 100 uses the behavior information AC21 indicating that the user U2 has eaten or eaten at the restaurant B within a predetermined range (for example, within 2 km) from the position LC22 at the end of the target period (date dt25). Extract as In the example of FIG. 2, an example in which only the information on the user U2 is extracted has been described. However, it is assumed that the same information is extracted for a large number of users (for example, 100,000 people).

  And the extracting device 100 produces | generates a model based on a positional information group and object action information (step S23). In the example of FIG. 2, the extraction device 100 generates a model M2 based on the four pieces of position information LC22 and the behavior information AC21. For example, the extraction device 100 generates a model M2 that predicts whether the user from whom the position information group has been extracted will eat or drink. For example, the extraction device 100 generates a model M2 that predicts whether the user will eat or drink after staying. In the example of FIG. 2, the extraction apparatus 100 generates a model M2 using the position information group extracted for a large number of users (for example, 100,000 people) including the user U2.

  For example, the extraction apparatus 100 generates a model M2 using a combination of a position information group extracted as satisfying the regularity of the behavior pattern PT2 and information indicating that eating and drinking has been performed before and after the behavior as a positive example. To do. For example, the extraction apparatus 100 generates a model M2 by performing a learning process so as to output a score “1” when a position information group as a positive example is input. For example, when the position information group is input, the extraction apparatus 100 generates the model M2 indicating that the possibility of eating and drinking after the action of the action pattern PT2 is higher as the score closer to “1” is output. . In addition, for example, the extraction device 100 may generate the model M2 by using, as a negative example, a user position information group detected before and after no eating or drinking is performed. For example, the extraction apparatus 100 generates a model M2 by performing a learning process so as to output a score “0” when a position information group serving as a negative example is input. For example, when the position information group is input, the extraction device 100 generates a model M2 indicating that the possibility of eating and drinking is lower as the score closer to “0” is output. For example, the extraction apparatus 100 takes a model M2 as a negative example of a combination of a positional information group extracted as satisfying the regularity of the behavior pattern PT2 and information indicating that eating and drinking is not performed before and after the behavior. It may be generated.

  As described above, the extraction apparatus 100 can appropriately extract user information that satisfies regularity. In the example of FIG. 1, when the user is performing an action such as going around the XX park, the extraction apparatus 100 extracts the position information group assuming that the position of the user satisfies a predetermined regularity regarding the change in position. . Further, the extraction apparatus 100 extracts a target action outside the target period, which is a period in which the position information group is detected. Thereby, the extraction device 100 can extract the positional information group indicating that the user is running and the target behavior of the user before and after the position information group in association with each other. Further, in the example of FIG. 2, when the user is staying at a predetermined point of interest (YY movie theater or the like), the extraction apparatus 100 assumes that the position of the user satisfies a predetermined regularity regarding the change in position. Extract information group. Thereby, the extraction device 100 can extract the position information group indicating that the user is staying at a predetermined position and the user's target behavior before and after the position information group in association with each other. In addition, the extraction apparatus 100 can generate a model that can appropriately estimate whether the user's target action has occurred by generating a model using the position information group extracted in this way. As described above, the extraction device 100 can generate a model that can appropriately estimate the user's behavior.

[1-4. (Target action (time) and information provision)
In the example described above, the case where the user performs the target action after the target period that is the period in which the position information group is detected has been described as an example. However, the target action of the user may be performed before the target period. . For example, in the example of FIG. 1, when the user U1 purchases X beverage at a Z convenience store at a time point (for example, date dt10) before the target period (dates dt12 to dt16), the extraction device 100 is a Z convenience store. You may extract purchase of X drink as object action. For example, the extraction device 100 uses the behavior information indicating that the user U1 has purchased the X beverage at the Z convenience store within a predetermined range (for example, within 1 km) from the position LC12 at the start of the target period (date dt12). Extract as As described above, the extraction device 100 may extract the target behavior of the user before the target period, and store the positional information group and the target behavior of the user before the target period in association with each other. In addition, the extraction device 100 may provide information that associates the position information group with the target behavior of the user before the target period to an external information processing device. For example, the extraction apparatus 100 may provide information that associates the position information group with the target action of the user before the target period to a business operator who is researching the vicinity of the area including the XX park. For example, the extraction device 100 may provide information that associates the position information group with the target behavior of the user before the target period to a business operator planning to open a store near the area including the XX park.

  For example, in the example of FIG. 2, when the user U2 eats and drinks at the C restaurant at a time point (for example, the date and time dt20) before the target period (date and time dt22 to dt25), the extraction device 100 uses the food and drink at the C restaurant. May be extracted as the target action. For example, the extraction device 100 uses, as the target behavior information, behavior information indicating that the user U2 has eaten or eaten at the C restaurant within a predetermined range (for example, within 2 km) from the position LC22 at the start of the target period (date and time dt22). Extract. For example, the extraction device 100 may provide information that associates the position information group with the target action of the user before the target period to a business operator who is researching the vicinity of the area including the YY movie theater. For example, the extraction apparatus 100 may provide information that associates the position information group with the target behavior of the user before the target period to a business operator planning to open a store near the area including the YY movie theater.

  Note that the extraction apparatus 100 may generate a model based on the position information group and the target behavior before the target period. For example, the extraction apparatus 100 may generate a model that estimates whether the user was eating or drinking before the position information group was extracted. For example, the extraction apparatus 100 generates a model that estimates whether the user has eaten or eaten before staying.

[1-5. (Target action (space) and information provision)
In the example described above, the case where the user performs the target action in the real space has been described as an example. However, the target action of the user is not limited to the real space, and may be performed in, for example, the Internet space (on the web).

  For example, in the example of FIG. 1, when the user U1 performs a search using the query “convenience store” by the terminal device 10 at a date and time (for example, date and time dt17) after the target period (date and time dt12 to dt16), the extraction device 100 May extract a search using the query “convenience store” by the user U1 as the target action. As described above, the extraction apparatus 100 may extract the target behavior of the user on the web after the target period, and store the positional information group and the target behavior of the user on the web in association with each other. In addition, the extraction device 100 may provide information that associates the position information group with the target behavior of the user on the web to an external information processing device. For example, the extraction apparatus 100 may provide information that associates the position information group with the target behavior of the user on the web to a business operator who is researching the vicinity of the area including the XX park. For example, the extraction apparatus 100 may provide information that associates the position information group with the target behavior of the user on the web to a business operator planning to open a store near an area including the XX park.

  For example, in the example of FIG. 2, when the user U2 performs a search using the query “restaurant” by the terminal device 10 at the date and time (for example, date and time dt26) after the target period (date and time dt22 to dt25), the extraction device 100 May extract a search using the query “restaurant” by the user U2 as the target action. For example, the extraction device 100 may provide information that associates the position information group with the target behavior of the user on the web to a business operator who is researching the vicinity of the area including the YY movie theater. For example, the extraction device 100 may provide information that associates the position information group with the target behavior of the user on the web to a business operator planning to open a store near an area including a YY movie theater. As described above, the extraction apparatus 100 may provide information for predicting demand to an external information processing apparatus based on information in which the position information group and the user's target behavior on the web are associated with each other.

[1-6. (Use of model)
The extraction apparatus 100 may estimate whether or not a predetermined action of the user has occurred using the models M1 and M2 generated in the above-described example.

  For example, the extraction apparatus 100 may predict whether or not the user U5 performs a predetermined action by inputting the position information group extracted for the user U5 to the model M1. For example, when the extraction apparatus 100 inputs the position information group extracted for the user U5 located around the ZZ park to the model M1 and the output score is equal to or greater than a predetermined threshold, the user U5 performs a purchase action. It may be predicted.

  For example, the extraction apparatus 100 may predict whether or not the user U10 performs a predetermined action by inputting the position information group extracted for the user U10 to the model M2. For example, the extraction apparatus 100 inputs a position information group extracted for the user U10 whose position does not change for a predetermined period to the model M2, and predicts that the user U10 will eat and drink when the output score is equal to or greater than a predetermined threshold. May be.

[1-7. (Extraction on the terminal device side)
In the above-described example, the example in which the extraction apparatus 100 extracts the position information group satisfying the predetermined regularity regarding the change of the position from the plurality of position information acquired from the terminal apparatus 10 has been described. Among a plurality of pieces of position information detected by, a position information group satisfying a predetermined regularity regarding a change in position may be extracted. In this case, the terminal device 10 may be an extraction device. In this case, the extraction system may include a terminal device 10 as an extraction device and a service providing device that collects information on regularity extracted by the terminal device 10 and provides a service. This point will be described below.

  The terminal device 10 used by each user extracts a position information group satisfying a predetermined regularity regarding a change in the position of each user from the detected plurality of position information. Thereby, the terminal device 10 extracts a position information group satisfying the regularity of each user. Further, the terminal device 10 may generate (calculate) information related to the regularity extracted for each individual user who uses the terminal device 10 and the context regarding the position, and transmit the result (information) to the service providing device. Good.

  For example, the terminal device 10 extracts a position information group satisfying a predetermined regularity corresponding to the “behavior pattern PT1 (running)”, information on the regularity generated based on the position information group (for example, circular movement), etc. And information indicating running or context information indicating running may be transmitted to the service providing apparatus. Further, for example, the terminal device 10 extracts a position information group satisfying a predetermined regularity corresponding to the “behavior pattern PT2 (stay)”, and information on the regularity generated based on the position information group (for example, position change) No)) or context information indicating that the user is staying for a certain period of time may be transmitted to the service providing apparatus. Note that the above is an example, and the terminal device 10 may generate information indicating various regularities and information indicating context. Then, the terminal device 10 may transmit the generated information together with information for identifying the terminal device 10 and information for identifying a user who uses the terminal device 10. The service providing apparatus that has acquired information from the terminal device 10 may store information indicating regularity or information indicating context and the user in association with each other, and provide the service to each user based on the information.

  In such an extraction system, the terminal device 10 transmits information generated on the client side (terminal device 10) to the service providing device without transmitting position information to the server side (service providing device), and the service providing device. Can provide services. In this case, for example, the service providing apparatus provides a service based on the list information (POI list) of the points of interest and the information on the regularity and the context acquired from the terminal device 10. For example, when information indicating the same regularity is acquired in the vicinity of a predetermined POI, the service providing apparatus may provide a service by predicting that the same action will be performed after the action indicating the regularity. Good. For example, when information indicating regularity PTX is acquired in the vicinity of a predetermined POI, the service providing apparatus has a predetermined number or more of information indicating an action pattern of purchasing the product X after the behavior indicating the regularity PTX If acquired, information prompting the purchase of the product X may be provided to the terminal device 10 from which information indicating the regularity PTX is acquired in the vicinity of a predetermined POI. Note that the above is an example, and the service providing apparatus may provide various services based on the POI information, information indicating the regularity of the user's behavior, and information indicating the context. As described above, the service providing apparatus can provide a service based on information having a higher abstraction level than the position information acquired from the terminal apparatus 10. Thereby, in the extraction system, it is possible to provide a service to the user while considering the privacy of each user.

[2. (Extractor configuration)
Next, the configuration of the extraction apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 4 is a diagram illustrating a configuration example of the extraction device according to the embodiment. As illustrated in FIG. 4, the extraction device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The extraction device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator of the extraction device 100, and a display unit (for example, a liquid crystal display) for displaying various types of information. May be.

(Communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to a network (for example, the network N in FIG. 3) by wire or wireless, and transmits / receives information to / from the terminal device 10.

(Storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. As illustrated in FIG. 4, the storage unit 120 according to the embodiment includes a user information storage unit 121, a behavior information storage unit 122, a position information storage unit 123, and a model information storage unit 124. Although not shown in FIG. 4, the storage unit 120 stores various types of points of interest such as locations where various facilities (points of interest) such as XX park (point of interest POI1) shown in FIG. 1 are located. Information is stored.

(User information storage unit 121)
The user information storage unit 121 according to the embodiment stores various types of information regarding the user. For example, the user information storage unit 121 stores various types of information regarding user attributes and various types of information regarding user behavior. FIG. 5 is a diagram illustrating an example of a user information storage unit according to the embodiment. The user information storage unit 121 illustrated in FIG. 5 includes items such as “user ID”, “age”, and “sex”.

  “User ID” indicates identification information for identifying a user. For example, the user identified by the user ID “U1” corresponds to the user U1 illustrated in the example of FIG. “Age” indicates the age of the user identified by the user ID. The “age” may be a specific age of the user identified by the user ID, such as 35 years old. “Gender” indicates the gender of the user identified by the user ID.

  For example, in the example illustrated in FIG. 5, the age of the user identified by the user ID “U1” is “30s”, and the gender is “male”. Further, for example, in the example illustrated in FIG. 5, the age of the user identified by the user ID “U2” is “20s”, and the gender is “female”.

  The user information storage unit 121 is not limited to the above, and may store various types of information according to the purpose. For example, the user information storage unit 121 may store other demographic attribute information and psychographic attribute information in addition to age and sex. For example, the user information storage unit 121 may store information such as name, home, work location, interest, family structure, income, lifestyle, and the like.

(Behavior information storage unit 122)
The behavior information storage unit 122 according to the embodiment stores various types of information regarding the user's behavior. FIG. 6 is a diagram illustrating an example of the behavior information storage unit according to the embodiment. For example, the behavior information storage unit 122 stores various behavior information such as the behavior of each user. The action information storage unit 122 illustrated in FIG. 6 includes items such as “user ID”, “action ID”, “date / time”, “type”, and “content”.

  “User ID” indicates identification information for identifying a user. For example, the user identified by the user ID “U1” corresponds to the user U1 illustrated in the example of FIG. The “behavior ID” indicates information for identifying the user's behavior. “Date and time” indicates the date and time when the corresponding user action was performed. In the example of FIG. 6, “dt18” or the like is illustrated, but “date and time” may store a specific date and time such as “July 6, 2017 19:31:52”. The “type” indicates information related to the type of action of the corresponding user. Further, “content” indicates content that is targeted in the corresponding user's action.

  For example, in the example illustrated in FIG. 6, the user (user U1) identified by the user ID “U1” has performed the action (behavior AC18) identified by the action ID “AC18”. For example, in the example illustrated in FIG. 6, the user U1 indicates that W water has been purchased (action AC18) at the A convenience store at the date dt18.

  For example, in the example illustrated in FIG. 6, the user identified by the user ID “U2” (user U2) indicates that the behavior identified by the behavior ID “AC26” (behavior AC26) has been performed. For example, in the example illustrated in FIG. 6, the user U2 indicates that the settlement of eating and drinking at the restaurant B (action AC26) is performed at the date and time dt26. That is, the action AC26 suggests that the user U2 has eaten or eaten at the B restaurant.

  The behavior information storage unit 122 is not limited to the above, and may store various types of information according to the purpose. Moreover, although the case where action information was memorize | stored in the action information storage part 122 for every user ID was shown in FIG. 6, action information may be memorize | stored not only for every user ID but in order of time, for example.

(Position information storage unit 123)
The position information storage unit 123 according to the embodiment stores various types of information regarding the position. For example, the position information storage unit 123 stores various types of information related to the position detected by the sensor. FIG. 7 is a diagram illustrating an example of the position information storage unit according to the embodiment. The position information storage unit 123 illustrated in FIG. 7 includes items such as “user ID” and “position information”. “Position information” includes items such as “date and time” and “position”.

  “User ID” indicates identification information for identifying a user. “Position information” indicates a history of user position information. “Date and time” indicates the date and time when each piece of position information was acquired. In the example of FIG. 7, “date and time” is illustrated by an abstract code “dt11” and the like, but “date and time” is a specific date and time such as “July 6, 2017 18:59:48”. May be stored. “Position” indicates position information of the corresponding user at the corresponding date and time. In the example of FIG. 7, “position” is illustrated by an abstract code “LC11” or the like, but specific latitude and longitude information or the like may be stored in “position”.

  For example, in the example illustrated in FIG. 7, the user position “LC11” identified by the user ID “U1” is acquired at the date “dt11”. For example, the position “LC11” of the terminal device 10 used by the user identified by the user ID “U1” is acquired at the date “dt11”. For example, in the example illustrated in FIG. 7, the user position “LC12” identified by the user ID “U1” is acquired at the date “dt12”. For example, the position “LC13” of the terminal device 10 used by the user identified by the user ID “U1” is acquired at the date “dt13”.

  Note that the position information storage unit 123 is not limited to the above, and may store various types of information according to the purpose. Further, the position information storage unit 123 may store information (flag) indicating that the position information group is extracted by the extraction process. The position information storage unit 123 may store information related to the user context.

(Model information storage unit 124)
The model information storage unit 124 according to the embodiment stores information about the model. For example, the model information storage unit 124 stores model information (model data) generated by the generation process. FIG. 8 is a diagram illustrating an example of a model information storage unit according to the embodiment. The model information storage unit 124 illustrated in FIG. 8 includes items such as “model ID”, “behavior pattern”, and “model data”. In FIG. 8, only the model M1 and the model M2 are illustrated, but a plurality of model information (models M3, M4, etc.) corresponding to each behavior pattern may be stored.

  “Model ID” indicates identification information for identifying a model. For example, the model identified by the model ID “M1” corresponds to the model M1 illustrated in the example of FIG. “Behavior pattern” indicates a behavior pattern of a corresponding model. The “behavior pattern” stores conditions corresponding to each behavior pattern. For example, the “behavior pattern PT1 (running)” stores conditions (for example, movement at a speed of 5 to 30 km per hour) for the user behavior pattern to correspond to running. “Model data” indicates data of a corresponding model associated with the model data. For example, “model data” includes information including nodes in each layer, functions adopted by the nodes, connection relationships between the nodes, and connection coefficients set for connections between the nodes.

  For example, in the example shown in FIG. 8, the model (model M1) identified by the model ID “M1” has the action pattern “PT1 (running)”, and the target action information before and after the action pattern PT1 is performed. It is used to predict what will happen. The model data of the model M1 is model data MDT1.

  The model M1 (model data MDT1) is a user position information group, and an input layer to which a position information group satisfying a predetermined regularity regarding a change in position is input, an output layer, an input layer to an output layer A first element belonging to any layer other than the output layer, and a second element whose value is calculated based on the first element and the weight of the first element, and is input to the input layer The sensor information is used to predict the occurrence of a predetermined action by performing an operation based on the first element and the weight of the first element, with each element belonging to each layer other than the output layer as the first element. This is a model for causing a computer to function so as to output score values from an output layer.

Here, it is assumed that the models M1, M2, etc. are realized by a regression model represented by “y = a ₁ * x ₁ + a ₂ * x ₂ +... + A _i * x _i ”. In this case, for example, the first element included in the model M1 corresponds to input data (xi) such as x1 or x2. The weight of the first element corresponds to the coefficient ai corresponding to xi. Here, the regression model can be regarded as a simple perceptron having an input layer and an output layer. When each model is regarded as a simple perceptron, the first element can correspond to any node of the input layer, and the second element can be regarded as a node of the output layer.

  Further, it is assumed that the models M1, M2, etc. are realized by a neural network having one or a plurality of intermediate layers, such as DNN (Deep Neural Network). In this case, for example, the first element included in the model M1 corresponds to any node of the input layer or the intermediate layer. The second element corresponds to the next node, which is a node to which a value is transmitted from the node corresponding to the first element. The weight of the first element corresponds to a connection coefficient that is a weight considered for a value transmitted from a node corresponding to the first element to a node corresponding to the second element.

  The model information storage unit 124 is not limited to the above, and may store various model information according to the purpose.

(Control unit 130)
Returning to the description of FIG. 4, the control unit 130 is a controller, and is stored in a storage device inside the extraction apparatus 100 by, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. Various programs (corresponding to an example of an extraction program or a generation program) are implemented by using the RAM as a work area. The control unit 130 is a controller, and is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA). The control unit 130 is a position information group of the user that satisfies a predetermined regularity regarding a change in position by information processing according to the model M1, the model M2, or the like stored in the model information storage unit 124. Is input based on the input layer, the output layer, the first element belonging to any layer other than the output layer from the input layer to the output layer, and the weight of the first element and the first element A first element and a weight of the first element, with each element belonging to each layer other than the output layer as a first element for sensor information input to the input layer. By performing the calculation based on the above, the score value used for predicting the occurrence of the predetermined action is output from the output layer.

  For example, the extraction device 100 calculates a usage score and a bid score using a model having an arbitrary structure such as the above-described regression model or neural network. Specifically, the models M1, M2, and the like are user position information groups, and position information groups satisfying predetermined regularity related to position changes (that is, each element used for calculating the above-described score) are input. In such a case, the coefficient is set so as to output a score relating to the occurrence or non-occurrence of the predetermined action (that is, a score indicating how likely the user is to perform the predetermined action). The information providing apparatus 10 calculates the score of each user in the corresponding behavior pattern using such models M1, M2, and the like.

  In the above example, when the models M1, M2, etc. are user position information groups, and a position information group satisfying a predetermined regularity related to a change in position is input, a score related to the occurrence of a predetermined action The example which is the model which outputs is shown. However, the models M1, M2, and the like according to the embodiment may be models generated based on results obtained by repeating data input / output. For example, the model is a user's position information group, and a model (model M1b) that is learned to receive a use score output from the model M1 with a position information group that satisfies a predetermined regularity regarding a change in position as an input. ). Alternatively, the model may be a user's position information group that has been learned so that a position information group that satisfies a predetermined regularity regarding a change in position is input and an output value of the model M1b is output. .

  Further, when the extraction apparatus 100 performs a prediction process using GAN (Generative Adversarial Networks), the models M1, M2, and the like may be models that constitute part of the GAN.

  As illustrated in FIG. 4, the control unit 130 includes an acquisition unit 131, an extraction unit 132, a generation unit 133, a prediction unit 134, and a provision unit 135, and functions and operations of information processing described below. Realize or execute. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 4, and may be another configuration as long as information processing described later is performed.

(Acquisition part 131)
The acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires various types of information from an external information processing device such as the terminal device 10. For example, the acquisition unit 131 acquires various types of information from the storage unit 120 such as the user information storage unit 121, the behavior information storage unit 122, the position information storage unit 123, and the model information storage unit 124. For example, the acquisition unit 131 acquires various types of information detected by the terminal device 10. For example, the acquisition unit 131 acquires position information and acceleration information detected by the terminal device 10.

  For example, the acquisition unit 131 acquires a plurality of pieces of user position information detected in a predetermined period. For example, the acquisition unit 131 acquires user behavior information outside the target period, which is a period in which the position information group is detected.

  In the example of FIG. 1, the acquisition unit 131 acquires the position LC11 of the user U1 at the date and time dt11. And the acquisition part 131 acquires position LC12 which shows that the user U1 exists in XX park (interesting point POI1) in the date dt12 after the date dt11. For example, the acquisition unit 131 acquires a position LC12 indicating that the user U1 is in a position different from the position LC11 at a date and time dt12 later than the date and time dt11 and is in the XX park. For example, the acquisition unit 131 acquires a position LC13 indicating that the user U1 is in a position different from the position LC12 in the XX park at a date and time dt13 later than the date and time dt12. For example, the acquisition unit 131 acquires a position LC14 indicating that the user U1 is in a position different from the position LC13 in the XX park at a date dt14 later than the date dt13.

  For example, the acquisition unit 131 acquires a position LC15 indicating that the user U1 is at a position different from the position LC14 in the XX park at a date and time dt15 later than the date and time dt14. For example, the acquiring unit 131 acquires a position LC16 indicating that the user U1 is at a position different from the position LC15 in the XX park at a date and time dt16 later than the date and time dt15. For example, the acquisition unit 131 acquires a position LC17 indicating that the user U1 has left the XX park at a date and time dt17 later than the date and time dt16.

  In the example of FIG. 1, the acquisition unit 131 acquires a position LC18 indicating that the user U1 is at the A convenience store (shop SP1) at a date dt18 later than the date dt17. For example, the acquisition unit 131 acquires action information AC11 indicating that the user U1 purchased W water at the A convenience store at the date dt18.

  In the example of FIG. 2, the acquisition unit 131 acquires the position LC21 of the user U2 at the date and time dt21. Then, the acquiring unit 131 acquires a position LC22 indicating that the user U2 is in the YY movie theater (point of interest POI2) at a date and time dt22 later than the date and time dt21. For example, the acquisition unit 131 acquires a position LC22 indicating that the user U2 is at a position different from the position LC21 and is in the YY movie hall at a date and time dt22 later than the date and time dt21. For example, the acquisition unit 131 acquires the position LC22 indicating that the user U2 is in the YY movie hall at the date dt23 later than the date dt22. For example, the acquisition unit 131 acquires the position LC22 indicating that the user U2 is in the YY movie hall at the date dt24 after the date dt23. For example, the acquisition unit 131 acquires the position LC22 indicating that the user U2 is in the YY movie hall at the date dt25 after the date dt24. For example, the acquiring unit 131 acquires a position LC23 indicating that the user U2 has left the YY movie theater at a date dt26 later than the date dt25.

  In the example of FIG. 2, the acquisition unit 131 acquires a position LC23 indicating that the user U2 is at the B restaurant (shop SP2) at the date and time dt26. For example, the acquisition unit 131 acquires action information AC21 indicating that the user U2 has eaten or eaten at the restaurant B at the date and time dt26.

(Extractor 132)
The extraction unit 132 extracts various information. For example, the extraction unit 132 extracts various types of information from the storage unit 120 such as the user information storage unit 121, the behavior information storage unit 122, the position information storage unit 123, and the model information storage unit 124.

  For example, the extraction unit 132 extracts a position information group that satisfies a predetermined regularity regarding a change in position from among a plurality of pieces of position information acquired by the acquisition unit 131. For example, the extraction unit 132 extracts a plurality of pieces of position information detected continuously as a position information group.

  For example, the extraction unit 132 extracts target behavior information that is behavior information satisfying a predetermined condition from user behavior information. For example, the extraction unit 132 extracts target behavior information after the target period. For example, the extraction unit 132 extracts the target behavior information within a predetermined range from the position corresponding to the position information at the end of the target period. For example, the extraction unit 132 extracts target behavior information before the target period. For example, the extraction unit 132 extracts the target behavior information within a predetermined range from the position corresponding to the position information at the start of the target period.

  In the example of FIG. 1, the extraction unit 132 extracts a position information group that satisfies a predetermined regularity related to a change in position, from among a plurality of pieces of position information LC11 to LC18 of the user detected at dates dt11 to dt18. For example, the extraction unit 132 extracts a position information group that satisfies a predetermined regularity corresponding to the “behavior pattern PT1 (running)” as shown in the model information storage unit 124 (see FIG. 8). For example, the extraction unit 132 extracts a position information group in which the change in the position information of the user U1 is changing at a speed corresponding to running such as 10 km / h. For example, the extraction unit 132 extracts a position information group when the user U1 is running in the XX park (the point of interest POI1). For example, the extraction unit 132 extracts a position information group in which the change in the position information of the user U1 is moving around the XX park at, for example, 10 km / h. For example, the extraction unit 132 extracts the position information LC12 to LC16 where the user U1 is moving around the XX park at the date and time dt12 to dt16.

  In the example of FIG. 1, the extraction unit 132 extracts target behavior information that is behavior information that satisfies a predetermined condition from user behavior information outside the target period, which is a period in which the position information group is detected. For example, the extraction unit 132 includes target behavior information that is behavior information satisfying a predetermined condition among user behavior information outside the target period (dates dt12 to dt16) in which the position information group (position information LC12 to LC16) is detected. To extract. For example, the extraction unit 132 extracts target behavior information that is behavior information satisfying a predetermined condition from user behavior information within a predetermined period (period PD1) from the end of the target period (date and time dt16). For example, the extraction unit 132 may extract target behavior information that is behavior information satisfying a predetermined condition among user behavior information within a predetermined range from the position LC16 at the end of the target period (date dt16). . For example, the extraction unit 132 uses the behavior information AC11 indicating that the user U1 has purchased W water at the A convenience store at the date and time dt18 within the predetermined period (period PD1) from the end of the target period (date and time dt16). Extract as information.

  In the example of FIG. 2, the extraction unit 132 extracts a position information group that satisfies a predetermined regularity related to a change in position from the plurality of pieces of position information LC21 to LC23 of the user detected at the dates dt21 to dt26. For example, the extraction unit 132 extracts a position information group that satisfies a predetermined regularity corresponding to the “behavior pattern PT2 (stay)” as shown in the model information storage unit 124 (see FIG. 8). For example, the extraction unit 132 extracts a position information group in which the change in the position information of the user U2 has not changed over, for example, a predetermined time. For example, the extraction unit 132 extracts a position information group when the user U2 is staying at a YY movie theater (interesting point POI2). For example, the extraction unit 132 extracts a position information group in which the change in the position information of the user U2 is, for example, in the YY movie hall over a predetermined time. For example, the extraction unit 132 extracts a plurality of pieces of position information LC22 indicating that the user U2 is staying in the YY movie hall at the dates dt22 to dt25.

  In the example of FIG. 2, the extraction unit 132 extracts target behavior information that is behavior information satisfying a predetermined condition from user behavior information outside the target period, which is a period in which the position information group is detected. For example, the extraction unit 132 includes target behavior information that is behavior information that satisfies a predetermined condition among user behavior information outside the target period (dates dt22 to dt25) in which the position information group (four position information LC22) is detected. To extract. For example, the extraction unit 132 extracts target behavior information that is behavior information satisfying a predetermined condition from user behavior information within a predetermined period (period PD2) from the end of the target period (date and time dt25). For example, the extraction unit 132 may extract target behavior information that is behavior information satisfying a predetermined condition among user behavior information within a predetermined range from the position LC22 at the end of the target period (date dt25). . For example, the extraction unit 132 uses the behavior information AC21 indicating that the user U2 has eaten or eaten at the restaurant B at the date and time dt26 within the predetermined period (period PD2) from the end of the target period (date and time dt25). Extract as

(Generator 133)
The generation unit 133 generates various information. For example, the generation unit 133 generates a model from information stored in the user information storage unit 121, the behavior information storage unit 122, the position information storage unit 123, and the like. For example, the generation unit 133 generates a model used for predicting the occurrence of a predetermined action based on learning data including a position information group and target action information corresponding to the position information group. For example, the generation unit 133 predicts whether or not a predetermined action has occurred based on learning data including the position information group stored in the position information storage unit 123 and the target action information stored in the action information storage unit 122. Generate the model used for.

  For example, the generation unit 133 generates models M1, M2, etc., and stores the generated models M1, M2, etc. in the model information storage unit 124. Note that the generation unit 133 may generate the models M1, M2, and the like using any learning algorithm. For example, the generation unit 133 generates models M1, M2, and the like using a learning algorithm such as a neural network, a support vector machine, clustering, and reinforcement learning. As an example, when the generation unit 133 generates models M1, M2, etc. using a neural network, the models M1, M2, etc. include an input layer including one or more neurons, an intermediate layer including one or more neurons, and one And an output layer including the above neurons.

  In the example of FIG. 1, the generation unit 133 generates a model based on the position information group and the target behavior information. For example, the generation unit 133 generates the model M1 based on the position information LC12 to LC16 and the behavior information AC11. For example, the production | generation part 133 produces | generates the model M1 which estimates whether the user from which the positional information group was extracted will purchase a drink. For example, the generation unit 133 generates a model M1 that predicts whether the user purchases a beverage after running.

  In the example of FIG. 2, the generation unit 133 generates a model based on the position information group and the target behavior information. For example, the generation unit 133 generates a model M2 based on the four pieces of position information LC22 and the behavior information AC21. For example, the generation unit 133 generates a model M2 that predicts whether the user from whom the position information group has been extracted eats or drinks. For example, the generation unit 133 generates a model M2 that predicts whether the user will eat or drink after staying.

(Prediction unit 134)
The prediction unit 134 predicts various types of information. For example, the prediction unit 134 predicts the presence / absence of a predetermined action of the user based on the position information stored in the position information storage unit 123. For example, the prediction unit 134 predicts the presence / absence of a predetermined action of the user by inputting the position information stored in the position information storage unit 123 to the model stored in the model information storage unit 124.

  For example, the prediction unit 134 predicts user behavior before and after a predetermined period based on the position information group extracted by the extraction unit 132. For example, the prediction unit 134 is a model used for predicting the occurrence or non-occurrence of a predetermined action, and predicts the presence or absence of the user's predetermined action based on the output of the model to which the position information group is input.

  For example, the prediction unit 134 inputs, to the input layer, a position information group that satisfies a predetermined regularity regarding a change in position, which is a position information group of the user, by information processing according to models M1, M2, and the like. Then, the predicting unit 134 causes the input data to propagate to the intermediate layer and the output layer, thereby outputting a score relating to the presence or absence of the predetermined action from the output layer.

  In the example of FIG. 1, the prediction unit 134 estimates that the user U1 is heading to the XX park by comparing the POI information with the position LC of the user U1. For example, the prediction unit 134 may estimate the moving direction of the user based on the acceleration detected by the terminal device 10. For example, the prediction unit 134 may estimate the movement from the point where the position information was last detected based on the acceleration detected by the terminal device 10 by appropriately using various conventional techniques.

  In the example of FIG. 1, the prediction unit 134 estimates that the user U1 is heading toward the XX park by comparing the POI information with the change in the position LC of the user U1. For example, the prediction unit 134 estimates that the user U1 is moving toward the XX park (from the bottom to the top in FIG. 1) based on the positions LC11 and LC12. For example, the prediction unit 134 estimates that the user U1 is moving in the XX park based on the positions LC12 and LC13. For example, the prediction unit 134 estimates the moving speed of the user U1 based on the positions LC12 and LC13. For example, the prediction unit 134 may estimate the moving speed of the user based on the acceleration detected by the terminal device 10. For example, the prediction unit 134 estimates that the user U1 is moving in the XX park based on the positions LC13 and LC14. For example, the prediction unit 134 estimates the moving speed of the user U1 based on the positions LC13 and LC14. For example, the prediction unit 134 estimates that the user U1 is moving in the XX park based on the positions LC14 and LC15. For example, the prediction unit 134 estimates the moving speed of the user U1 based on the positions LC14 and LC15. For example, the prediction unit 134 estimates that the user U1 is moving in the XX park based on the positions LC15 and LC16. For example, the prediction unit 134 estimates the moving speed of the user U1 based on the positions LC15 and LC16. For example, the predicting unit 134 estimates that the user U1 is running in the XX park based on the movement progress of the user U1 and the speed of the user U1.

  In the example of FIG. 2, the prediction unit 134 estimates that the user U2 is heading for the YY movie theater by comparing the POI information with the position LC of the user U2. For example, the prediction unit 134 estimates that the user U2 is heading for the YY movie theater by comparing the POI information with the change in the position LC of the user U2. For example, the prediction unit 134 estimates that the user U2 is moving toward the YY movie theater (from bottom to top in FIG. 2) at the positions LC21 and LC22. For example, the prediction unit 134 estimates that the user U2 is staying in the YY movie hall based on the position LC22 acquired at the date dt22 and the date dt23. For example, the prediction unit 134 estimates that the user U2 is staying in the YY movie hall based on the position LC22 acquired from the date dt22 to the date dt24. For example, the prediction unit 134 estimates that the user U2 is staying in the YY movie hall based on the position LC22 acquired from the date dt22 to the date dt25.

(Providing unit 135)
The providing unit 135 provides various information. For example, the providing unit 135 provides various types of information to an external device such as the terminal device 10. For example, the providing unit 135 provides each model stored in the model information storage unit 124 to an external information processing apparatus. For example, the providing unit 135 provides information based on the prediction result to an external information processing apparatus. For example, the providing unit 135 may distribute an advertisement related to a convenience store near the XX park to the terminal device 10 of the user U1 estimated to be running in the XX park. Further, for example, the providing unit 135 may distribute an advertisement regarding a restaurant (restaurant) near the YY movie theater to the terminal device 10 of the user U2 who is estimated to be staying in the YY movie theater.

  For example, the providing unit 135 may provide information regarding the position information group extracted by the extracting unit 132 to a predetermined service provider. Further, for example, the providing unit 135 may provide information related to the model generated by the generating unit 133 to an external information processing apparatus. For example, the providing unit 135 may provide information output by the model to an external information processing apparatus. Note that the above is an example, and the providing unit 135 may provide various information based on the position information group extracted by the extracting unit 132 to an external information processing apparatus.

[3. (Extraction process flow)
Next, the procedure of extraction processing by the extraction system 1 according to the embodiment will be described with reference to FIG. FIG. 9 is a flowchart illustrating an example of the extraction process according to the embodiment. Further, the extraction process illustrated in FIG. 9 includes extraction of target behavior information and generation of a model.

  As illustrated in FIG. 9, the extraction apparatus 100 acquires a plurality of pieces of position information of a user detected in a predetermined period (step S101). For example, the extraction apparatus 100 acquires a history of position information as shown in the position information storage unit 123 (see FIG. 7). In the example of FIG. 1, the extraction device 100 acquires position information LC11 to LC18 detected at the date and time dt11 to dt18 from the terminal device 10.

  In addition, the extraction apparatus 100 extracts a position information group that satisfies a predetermined regularity regarding a change in position (step S102). For example, the extraction apparatus 100 extracts a position information group that satisfies the behavior pattern PT1 related to movement as shown in the model information storage unit 124 (see FIG. 8). In the example of FIG. 1, the extraction apparatus 100 extracts a position information group (position information LC12 to LC16) corresponding to a position circling in the XX park from the position information LC11 to LC18.

  Further, the extraction apparatus 100 extracts target behavior information satisfying a predetermined condition from the user behavior information (step S103). For example, the extraction device 100 extracts the target behavior information after the target period (dates dt12 to dt16) from the behavior history as shown in the behavior information storage unit 122 (see FIG. 6). In the example of FIG. 1, the extraction device 100 extracts behavior information AC11 indicating that W water has been purchased at the A convenience store at the date dt18.

  Then, the extraction apparatus 100 generates a model used for predicting the occurrence / non-occurrence of the predetermined behavior based on the position information group and the target behavior information (Step S104). For example, the extraction device 100 generates a model M1 corresponding to the behavior pattern PT1 as shown in the model information storage unit 124 (see FIG. 8). In the example of FIG. 1, the extraction device 100 generates the model M1 using the position information LC12 to LC16 and the action information AC11 corresponding to the position information LC12 to LC16 as correct answer information.

[4. effect〕
As described above, the extraction device 100 according to the embodiment includes the acquisition unit 131 and the extraction unit 132. The acquisition unit 131 acquires a plurality of pieces of position information of a user detected in a predetermined period. The extraction unit 132 extracts a position information group that satisfies a predetermined regularity regarding a change in position from among a plurality of pieces of position information acquired by the acquisition unit 131.

  As described above, the extraction apparatus 100 according to the embodiment appropriately extracts user information satisfying regularity by extracting a position information group satisfying predetermined regularity regarding a change in position from among a plurality of pieces of position information of the user. Can be extracted.

  In the extraction device 100 according to the embodiment, the extraction unit 132 extracts a plurality of pieces of position information detected continuously as a position information group.

  As described above, the extraction apparatus 100 according to the embodiment can appropriately extract user information satisfying regularity by extracting a plurality of pieces of position information detected continuously as a position information group.

  In addition, the extraction device 100 according to the embodiment includes a prediction unit 134. The prediction unit 134 predicts user behavior before and after a predetermined period based on the position information group extracted by the extraction unit 132.

  As described above, the extraction apparatus 100 according to the embodiment can appropriately predict the user's behavior by predicting the user's behavior before and after the predetermined period based on the extracted position information group.

  Further, in the extraction device 100 according to the embodiment, the prediction unit 134 is a model used for predicting whether or not a predetermined action has occurred, and based on the output of the model in which the position information group is input, Predict the presence or absence of action.

  As described above, the extraction apparatus 100 according to the embodiment is a model used for predicting whether or not a predetermined action has occurred, and based on the output of the model in which the position information group is input, whether or not the user has the predetermined action By predicting, it is possible to appropriately predict the user's behavior.

  In the extraction device 100 according to the embodiment, the acquisition unit 131 acquires user behavior information outside the target period, which is a period in which the position information group is detected. The extraction unit 132 extracts target behavior information that is behavior information satisfying a predetermined condition from the user behavior information.

  As described above, the extraction apparatus 100 according to the embodiment extracts target behavior information that is behavior information satisfying a predetermined condition from user behavior information outside the target period that is a period in which the position information group is detected. Thus, it is possible to appropriately extract user information satisfying regularity.

  In the extraction device 100 according to the embodiment, the extraction unit 132 extracts target behavior information after the target period.

  As described above, the extraction apparatus 100 according to the embodiment can appropriately extract user information satisfying regularity by extracting the target behavior information after the target period.

  In the extraction device 100 according to the embodiment, the extraction unit 132 extracts the target behavior information within a predetermined range from the position corresponding to the position information at the end of the target period.

  As described above, the extraction apparatus 100 according to the embodiment appropriately extracts user information satisfying regularity by extracting the target behavior information within a predetermined range from the position corresponding to the position information at the end of the target period. Can be extracted.

  In the extraction device 100 according to the embodiment, the extraction unit 132 extracts target behavior information before the target period.

  As described above, the extraction apparatus 100 according to the embodiment can appropriately extract user information satisfying regularity by extracting the target behavior information before the target period.

  In the extraction device 100 according to the embodiment, the extraction unit 132 extracts target behavior information within a predetermined range from a position corresponding to the position information at the start of the target period.

  As described above, the extraction apparatus 100 according to the embodiment appropriately extracts user behavior information that satisfies regularity by extracting target behavior information within a predetermined range from a position corresponding to position information at the start of the target period. Can be extracted.

  In addition, the extraction device 100 according to the embodiment includes a generation unit 133. The generation unit 133 generates a model used for predicting the occurrence or non-occurrence of a predetermined action based on learning data including a position information group and target action information corresponding to the position information group.

  As described above, the extraction apparatus 100 according to the embodiment generates a model used for predicting the occurrence of a predetermined action based on learning data including a position information group and target action information corresponding to the position information group. By doing so, it is possible to appropriately predict the user's behavior.

[5. Hardware configuration)
The extraction apparatus 100 according to the embodiment described above is realized by a computer 1000 having a configuration as shown in FIG. 10, for example. FIG. 10 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the extraction device. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM (Read Only Memory) 1300, an HDD (Hard Disk Drive) 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F). ) 1700.

  The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

  The HDD 1400 stores programs executed by the CPU 1100, data used by the programs, and the like. The communication interface 1500 receives data from other devices via the network N and sends the data to the CPU 1100, and transmits data generated by the CPU 1100 to other devices via the network N.

  The CPU 1100 controls an output device such as a display and a printer and an input device such as a keyboard and a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. In addition, the CPU 1100 outputs the generated data to the output device via the input / output interface 1600.

  The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

  For example, when the computer 1000 functions as the extraction device 100 according to the embodiment, the CPU 1100 of the computer 1000 executes a program or data (for example, models M1, M2, etc.) loaded on the RAM 1200, thereby controlling the control unit 130. Realize the function. The CPU 1100 of the computer 1000 reads and executes these programs or data (for example, models M1 and M2) from the recording medium 1800. As another example, these programs or data are acquired from other devices via the network N. May be.

  As described above, some of the embodiments of the present application have been described in detail with reference to the drawings. It is possible to implement the present invention in other forms with improvements.

[6. Others]
In addition, among the processes described in the above embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified. For example, the various types of information illustrated in each drawing is not limited to the illustrated information.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured.

  In addition, the processes described in the embodiments described above can be appropriately combined within a range in which the process contents are not contradictory.

  In addition, the “section (module, unit)” described above can be read as “means” or “circuit”. For example, the acquisition unit can be read as acquisition means or an acquisition circuit.

DESCRIPTION OF SYMBOLS 1 Extraction system 100 Extraction apparatus 121 User information storage part 122 Behavior information storage part 123 Position information storage part 124 Model information storage part 130 Control part 131 Acquisition part 132 Extraction part 133 Generation part 134 Prediction part 135 Provision part 10 Terminal apparatus N network

Claims (14)

An acquisition unit for acquiring a plurality of pieces of position information of users detected in a predetermined period;
An extraction unit that extracts a position information group that satisfies a predetermined regularity regarding a change in position among the plurality of position information acquired by the acquisition unit;
An extraction device comprising: The extraction unit includes:
The extraction apparatus according to claim 1, wherein a plurality of pieces of position information detected in succession are extracted as the position information group. A prediction unit that predicts the user's behavior before and after the predetermined period based on the position information group extracted by the extraction unit;
The extraction device according to claim 1, further comprising: The prediction unit
The model used for predicting the presence or absence of a predetermined action, wherein the presence or absence of the predetermined action of the user is predicted based on an output of the model to which the position information group is input. 3. The extraction device according to 3. The acquisition unit
Obtaining the user's behavior information outside the target period, which is a period in which the position information group is detected,
The extraction unit includes:
The extraction device according to claim 1, wherein target behavior information that is behavior information satisfying a predetermined condition is extracted from the user behavior information. The extraction unit includes:
The extraction apparatus according to claim 5, wherein the target behavior information after the target period is extracted. The extraction unit includes:
The extraction apparatus according to claim 6, wherein the target behavior information is extracted within a predetermined range from a position corresponding to the position information at the end of the target period. The extraction unit includes:
The extraction device according to any one of claims 5 to 7, wherein the target behavior information before the target period is extracted. The extraction unit includes:
The extraction device according to claim 8, wherein the target behavior information is extracted within a predetermined range from a position corresponding to the position information at the start of the target period. A generating unit that generates a model used for predicting the occurrence of a predetermined action based on learning data including the position information group and the target action information corresponding to the position information group;
The extraction apparatus according to claim 5, further comprising: An extraction method performed by a computer,
An acquisition step of acquiring a plurality of pieces of position information of users detected in a predetermined period;
An extraction step of extracting a position information group satisfying a predetermined regularity regarding a change in position among the plurality of position information acquired by the acquisition step;
The extraction method characterized by including. An acquisition procedure for acquiring a plurality of pieces of position information of a user detected in a predetermined period;
An extraction procedure for extracting a position information group satisfying a predetermined regularity related to a change in position among a plurality of pieces of position information acquired by the acquisition procedure;
An extraction program characterized by causing a computer to execute. Learning data including a position information group of users satisfying a predetermined regularity regarding a change in position, and correct information indicating whether or not a predetermined action by a user corresponding to the position information group has occurred,
A first element having an input layer and an output layer, the layer belonging to any layer from the input layer to the output layer other than the output layer, and weights of the first element and the first element A second element whose value is calculated on the basis of the above, and for each of the position information groups input to the input layer, each element belonging to each layer other than the output layer as the first element, Correct information corresponding to the position information group is output from the output layer of the model by being input to the input layer of the model that performs the calculation based on the weight of the first element. And learning based on the comparison between the output value and
Learning data to make the computer function. An input layer in which a position information group that is a user's position information group and satisfies a predetermined regularity regarding a change in position is input;
The output layer,
A first element belonging to any layer from the input layer to the output layer other than the output layer;
A second element whose value is calculated based on the first element and a weight of the first element;
The sensor information input to the input layer is subjected to a calculation based on the first element and the weight of the first element, with each element belonging to each layer other than the output layer as the first element. So as to output from the output layer the score value used for predicting the occurrence of the behavior of
A model generated by a computer function.

Images