JP6560486B2 - Weekday / non-weekday estimation device and weekday / non-weekday estimation method - Google Patents

Description

  The present invention relates to a technique for estimating weekdays / non-weekdays.

  2. Description of the Related Art Conventionally, there has been proposed a technique for automatically inferring a user's action purpose and presenting guidance information without the user inputting the action purpose. For example, when predicting a destination by referring to past travel history based on the departure time and day of the week, if the departure time is a weekday, refer to the travel history from Monday to Friday, and if the departure time is a Saturday and Sunday The prediction is performed by referring to the movement history of (Patent Document 1).

  Also, based on the movement history, the frequency of arrival at each day of the week (appearance frequency) is calculated, the action class is calculated based on the arrival frequency and appearance frequency, and the action class corresponding to the current date and time information is calculated. A destination prediction method that predicts a destination using the included movement history has also been proposed (Patent Document 2).

JP 2005-156350 A JP 2009-36594 A

  In Patent Document 1, classification is performed on Monday to Friday as weekdays and Saturdays and Sundays as holidays. In other words, it was applicable only to this limited lifestyle.

  In the case of Patent Document 2, the appearance frequency for each day of the week is obtained for each destination, and it is not possible to discriminate between weekdays / non-weekdays (holidays).

  Therefore, an object of the present invention is to make it possible to estimate weekdays / non-weekdays based on the movement history of a moving object.

The weekday / non-weekday estimation device according to the present invention is:
A movement information acquisition unit for acquiring movement information of the moving body;
A history storage unit that associates the movement information of the mobile body with the date and time, and stores the movement information as movement history information;
Divide a day into multiple time zones, determine the frequency of movement of each time zone from the travel history for each day of the week, classify the data for each day of the week into two groups according to the degree of similarity, and classify into groups with a large number of elements The estimated day of the week is a weekday and the day of the week classified into a group with a small number of elements is a non-weekday;
Have

  The weekday / non-weekday estimation device may determine, as the movement frequency, a frequency at which the arrival time of the moving body belongs to each time zone based on the movement history.

In the weekday / non-weekday estimation device, the history storage section divides an area where the moving body moves into a plurality of areas, determines an area where the moving body is located based on the position information, and the moving body Holding history information representing the transition of the area where the moving body is located with the movement of
The estimation unit obtains a visit frequency for each area based on the history information, estimates the area having the highest visit frequency as an area to which a home belongs, and relates to the area to which the home belongs in the history information. History information may be extracted, and the weekdays or non-weekdays may be estimated based on the extracted history information.

Further, the weekday / non-weekday estimation method according to the present invention is:
Acquiring movement information of the moving object;
Associating movement information of the mobile object with date and time, and storing it as movement history information;
Divide a day into multiple time zones, determine the frequency of movement of each time zone from the travel history for each day of the week, classify the data for each day of the week into two groups according to the degree of similarity, and classify into groups with a large number of elements Estimating a selected day of the week as a weekday and a day of the week classified into a low-element group as a non-weekday;
Is executed by the computer.

  The weekday / non-weekday estimation method may obtain the frequency of arrival in each time zone from the travel history as a boarding frequency for each arrival time for each day of the week, and perform the estimation according to data for each day of the week.

  The contents described in the means for solving the problems can be combined as much as possible without departing from the problems and technical ideas of the present invention. The contents of the means for solving the problems can be provided as a device such as a computer or a system including a plurality of devices, a method executed by the computer, or a program executed by the computer. Further, a recording medium that holds the program may be provided.

  According to the present invention, weekdays / non-weekdays can be estimated based on the movement history of the moving object.

FIG. 1 is a functional block diagram of a weekday / non-weekday estimation apparatus. FIG. 2 is a diagram illustrating an example of movement history information stored in the history storage unit. FIG. 3 is a diagram illustrating an example of data for each day of the week that the estimation unit obtains based on movement history information. FIG. 4 is an apparatus configuration diagram illustrating an example of a computer (information processing apparatus). FIG. 5 is a diagram illustrating an example of processing for storing a movement history. FIG. 6 is a diagram illustrating an example of the estimation process. FIG. 7 is an explanatory diagram of a process for obtaining the boarding rate for each time zone. FIG. 8 is an explanatory diagram of processing for obtaining the Euclidean distance between the feature vectors of each day of the week. FIG. 9 is a diagram illustrating an example of data of a user who commute by car. FIG. 10 is a diagram illustrating a clustering result of the data in FIG. FIG. 11 is a diagram illustrating a data example of a user who does not use a car for commuting. FIG. 12 is a diagram illustrating an example of a feature vector based on the data of FIG. FIG. 13 is a diagram illustrating a clustering result of the data in FIG. FIG. 14 is a diagram illustrating a modification of the first embodiment. FIG. 15 is a functional block diagram of the weekday / non-weekday estimation apparatus according to the second embodiment. FIG. 16 is an explanatory diagram of a regional mesh. FIG. 17 is a diagram illustrating an example of movement history information held by the history storage unit. FIG. 18 is a diagram illustrating an example of a feature amount obtained by the estimation unit based on movement history information. FIG. 19 is an explanatory diagram of the observation date when obtaining the feature amount. FIG. 20 is a diagram illustrating a calculation example of the arrival time information amount. FIG. 21 is a diagram illustrating an example of a process for storing a movement history. FIG. 22 is a diagram illustrating an example of the estimation process. FIG. 23 is a diagram illustrating an example of home / work area estimation processing. FIG. 24 is a diagram illustrating an example of a home area estimation process. FIG. 25 is a functional block diagram of the area attribute estimation device. FIG. 26 is a diagram illustrating an example of a feature amount obtained by the estimation unit based on history information. FIG. 27 is an explanatory diagram of the observation date when the feature amount is obtained. FIG. 28 is a diagram illustrating an example of processing for storing a movement history. FIG. 29 is a diagram illustrating an example of the estimation process. FIG. 30 is a diagram illustrating an example of a home area estimation process. FIG. 31 is a diagram illustrating an example of work area estimation processing. FIG. 32 is a diagram illustrating an example of a home area estimation process according to Modification 3-1. FIG. 33 is a diagram illustrating an example of an estimation process according to Modification 3-2.

Embodiment 1
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, the following embodiment is an illustration of this invention and this invention is not limited to the following structure.
<Functional explanation>
FIG. 1 is a functional block diagram of a weekday / non-weekday estimation apparatus according to the present embodiment. The weekday / non-weekday estimation device 1 of this embodiment is a device that moves together with a moving body such as a vehicle or a user, and is mounted on a vehicle such as an automobile (passenger car, truck, motorcycle, etc.), a motorbike, a light vehicle, or the like. And a user terminal (mobile device) such as a mobile phone used by a user, a smartphone, a tablet terminal (slate PC (Personal Computer)), a notebook PC, and a portable game machine. The weekday / non-weekday estimation device 1 includes a movement information acquisition unit 11, a history storage unit 12, and an estimation unit 13.

The movement information acquisition unit 11 is a sensor that detects the movement of a moving object.
Positioning system) Acquires position information using a receiver. That is, the movement of the moving body is detected by the transition of the position information. Further, the detection of movement is not limited to the position information, but may be detected based on the vehicle speed, acceleration, or the state of the ignition switch.

  The history storage unit 12 associates the movement information of the moving object with the date and time and stores it as movement history information. For example, when the movement information acquisition unit 11 detects that the moving body has started moving, this time is stored as the departure time. Further, the time when the vehicle speed becomes zero from the state where the vehicle is moving and the ignition switch is turned off is stored as the arrival time. In the case of a user terminal, the movement information acquisition unit 11 periodically acquires a position, the time when the movement information acquisition unit 11 detects that the movement information acquisition unit 11 has moved to a position separated by a predetermined distance or more is set as a departure time. The time when the vehicle starts to stagnate within a predetermined range is stored as the arrival time.

  The estimation unit 13 divides a day into a plurality of time zones, obtains the movement frequency of each time zone from the movement history for each day of the week, classifies the data for each day of the week into two groups according to the degree of similarity, The days of the week classified into a group with a large number of days are estimated as weekdays, and the days of the week classified into a group with a small number of elements are estimated as non-weekdays.

For example, distance functions such as Euclidean distance, Manhattan distance, Canberra distance, and Minkowski distance, cosine similarity, and Pearson similarity can be used as the similarity of data for each day of the week. The cosine similarity and the Pearson similarity can be used as distances by converting them into distance = 1−cosine similarity and distance = 1−Pearson similarity.

  And, as a method of classifying into two groups according to the similarity such as distance, for example, the Ward method, the nearest neighbor method (also called the shortest distance method or the single connection method), the farthest neighbor method (the farthest distance method) Method, also called the perfect connection method), county average method, center of gravity method, median method, and McQuitty method.

  FIG. 2 is a diagram illustrating an example of movement history information stored in the history storage unit 12. As shown in FIG. 2, the movement history information stores departure date / time and arrival date / time in association with each other. That is, the movement from the departure date to the arrival date is performed.

  FIG. 3 is a diagram illustrating an example of data for each day of the week that the estimation unit 13 obtains based on the movement history information. In the example of FIG. 3, one day is divided into 24 time zones from 1 o'clock to 23:00, and the number of boarding times in each time zone is counted based on the travel history, and the number of days of each day (number of appearances) ) To get the boarding rate (movement frequency).

  In this example, the number of times of boarding is counted as 1 even if there are a plurality of times of boarding in one time zone of the day. And if the target period for calculating this boarding rate is 21 days and each day of the week appears for 3 days, the number of boardings in each time zone is divided by the number of days of each day (3 days). The number of boarding is obtained as a boarding rate.

  At this time, since holidays and long vacation periods often take different actions, it may not be valid data, so if there are holidays or long vacation periods during this period, exclude the data for that day To do. The long vacation period is New Year's holiday or Golden Week, and may be arbitrarily set.

<Device configuration>
FIG. 4 is an apparatus configuration diagram illustrating an example of a computer (information processing apparatus). The weekday / non-weekday estimation device 1 is, for example, a computer as shown in FIG. 4 includes a CPU (Central Processing Unit) 1001, a main storage device 1002, an auxiliary storage device (external storage device) 1003, a communication IF (Interface) 1004, an input / output IF (Interface) 1005, a drive device 1006, A communication bus 1007 is provided. The CPU 1001 performs processing and the like according to this embodiment by executing a program. The main storage device 1002 caches programs and data read by the CPU 1001 and develops a work area of the CPU. Specifically, the main storage device is a RAM (Random Access Memory), a ROM (Read Only Memory), or the like. The auxiliary storage device 1003 stores programs executed by the CPU 1001, position information, and the like. Specifically, the auxiliary storage device 1003 is an HDD (Hard-disk Drive), an SSD (Solid State Drive), an eMMC (embedded Multi-Media Card), a flash memory, or the like. The main storage device 1002 and the auxiliary storage device 1003 function as the history storage unit 12 of the weekday / non-weekday estimation device 1.

The communication IF 1004 transmits / receives data to / from other computers. The weekday / non-weekday estimation apparatus 1 is connected to a network via a communication IF 1004. The communication IF 1004 is specifically a wired or wireless network card or the like. The input / output IF 1005 is connected to the input / output device and accepts input from the user or outputs information to the user. Specifically, the input / output device is a keyboard, a mouse, a display, a touch panel, an acceleration sensor, or the like. The drive device 1006 reads data recorded on a storage medium such as a magnetic disk, a magneto-optical disk, and an optical disk, and writes data to the storage medium. The above components are connected by a communication bus 1007. A plurality of these components may be provided, or some of the components (for example, the drive device 1006) may not be provided. Further, the input / output device may be integrated with the computer. In addition, the program executed in this embodiment is provided via a portable storage medium readable by the drive device 1006, a portable auxiliary storage device 1003 such as a flash memory, a communication IF 1004, and the like. It may be. Then, when the CPU 1001 executes the program, the computer as shown in FIG. 4 is operated as the weekday / non-weekday estimation device 1 shown in FIG.

<Estimation method>
Next, details of processing executed by the weekday / non-weekday estimation apparatus 1 according to the program will be described. The weekday / non-weekday estimation device 1 stores a movement history as the moving body moves, and performs weekday / non-weekday estimation based on the movement history.

  FIG. 5 is a diagram illustrating an example of processing for storing a movement history. The weekday / non-weekday estimation device 1 starts the process of FIG. 5 at a predetermined timing. For example, the processing of FIG. 5 is periodically started every hour and every 15 minutes. In the case of an in-vehicle device, the processing in FIG. 5 may be performed when the vehicle starts moving or stops. In the case of a user terminal (mobile device), the process of FIG. 5 may be performed when movement or stagnation is detected from the acquired position information, or when movement or stagnation is detected by an acceleration sensor.

  When the processing of FIG. 5 is started, the weekday / non-weekday estimation device 1 first acquires the movement information of the moving body from the position information or the vehicle speed sensor by the GPS receiver as the movement information acquisition unit 11 (step S10).

  Further, the weekday / non-weekday estimation device 1 stores the departure time and the arrival time based on the acquired movement information and sets it as movement history information (step S20).

  FIG. 6 is a diagram illustrating an example of the estimation process. The estimation unit 13 of the weekday / non-weekday estimation apparatus 1 executes the process of FIG. 6 at a predetermined timing. For example, the processing of FIG. 6 may be started periodically every 6 hours or every 24 hours, or when a request is received from another application or system, or when the user selects activation, FIG. The process may be started.

  When the processing of FIG. 6 is started, the weekday / non-weekday estimation device 1 first reads movement history data for a predetermined number of days as a target period from the storage device (step S110), and each time zone divided into a plurality of times as shown in FIG. For each day of the week (step S120). For example, as shown in the movement history data 71 of FIG. 7, the departure time on July 8 (Tuesday) is 8:50, the arrival time is 9:30, and the departure time on July 15 (Tuesday) is 9:01. When the arrival time is recorded as 10:10, as shown in the boarding frequency data 72, the number of boarding at 8 o'clock on Tuesday is 1, the number of boarding at 9 o'clock on Tuesday is 2, and the number of boarding at 10 o'clock on Tuesday is Count as one. Then, when the number of appearances on Tuesday is 4 days, the number of times of boarding is divided by the number of appearances (4 days), and as shown in the boarding rate data 73, the boarding rate (movement frequency) at 8 o'clock on Tuesday Is 0.25, the occupancy rate at 9 o'clock on Tuesday is 0.5, and the occupancy rate at 10 o'clock on Tuesday is 0.25.

When calculating the boarding rate (moving frequency), rather than counting the entire period of travel vaguely, pay attention to the arrival time, count the number of boarding times (moving times) in the time zone to which the arrival time belongs, The rate (movement frequency) may be obtained. In this case, the time zone of 1 hour before and after may be counted. For example, as shown in the movement history data 71 of FIG. 7, when the arrival time on Tuesday, July 8 is 9:30, in addition to the time zone of 9 o'clock, the time of 8 o'clock and 10 o'clock Count the number of rides on the belt as well. Similarly, July 1 shown in the movement history data 71 of FIG.
If the arrival time on the 5th (Tuesday) is 10:10, in addition to the time zone of 10 o'clock as shown in the boarding frequency data 74, the number of times of boarding is counted in the time zone of 9 o'clock and 11 o'clock. As shown in the boarding frequency data 74, the number of boarding at 8 o'clock on Tuesday is 1, the number of boarding at 9 o'clock is 2, the number of boarding at 10 o'clock is 2, and the number of boarding at 11 o'clock is 1.

  Then, when the number of appearances on Tuesday is 4 days, the number of times of boarding is divided by the number of appearances (4 days), and as shown in the boarding rate data 73, the boarding rate (movement frequency) at 8 o'clock on Tuesday 0.25, Tuesday 9 o'clock boarding rate 0.5, Tuesday 10 o'clock boarding rate 0.5, Tuesday 11 o'clock boarding rate 0.25. This is to prevent the frequency (number of times of boarding) from being divided near the boundary of the time zone.

Next, the weekday / non-weekday estimation device 1 clusters the data of each day of the week into two groups according to the similarity. For example, the Euclidean distance is obtained for the occupancy rate for each time zone on each day of the week, and data having a close distance is defined as one cluster by the Ward method, and this is repeated to form two clusters (groups) (step S130). .
For example, when the data for each day of the week is shown as a feature vector as shown in FIG. 8A, the Euclidean distance between the feature vectors is obtained as a round robin as shown in FIG. Is a cluster. When obtaining the Euclidean distance, specifically, for the feature vector of each day of the week, the occupancy rate at 0 o'clock is A, the occupancy rate at 1 o'clock is B, the occupancy rate at 2 o'clock is C ... 23:00 The boarding rate is indicated by X, and the distance is obtained by taking the difference for each time zone as shown in FIG. As a result, data with a small difference in boarding rate for each time zone, that is, data with a similar boarding tendency can be obtained as data having a short distance.

  Then, the weekday / non-weekday estimation apparatus 1 selects a day of the week classified into a cluster (group) having a large number of elements (number of data for each day) from two clusters (groups) on weekdays and a cluster (group) having a small number of elements. ) Are estimated as non-weekdays (step S140).

  FIG. 9 is a diagram showing an example of data of a user who commute by car. In FIG. 9, the number of boarding times from 0 o'clock to 23 o'clock counted for each day of the week based on movement history information, and each day of the week The number of occurrences of is shown. The boarding rate (movement frequency) for each day of the week shown in FIG. 3 is obtained by dividing the number of times of boarding for each day of the week in FIG. 9 by the number of appearances of each day of the week.

  Then, the boarding rate for each day of the week in FIG. 3 is used as a feature vector, and clustering is performed according to the distance between the feature vectors as described above, so that it can be classified into two clusters (groups) as shown in FIG. In the example of FIG. 10, data of Monday, Tuesday, Wednesday, Thursday, and Friday are classified into one cluster, and data of Saturday and Sunday are classified into the other cluster. Accordingly, it can be estimated that Monday, Tuesday, Wednesday, Thursday, and Friday are weekdays and Saturday and Sunday are non-weekdays for the user.

  In this way, in the case of a user who commute by car, the occupancy rate (movement frequency) in a specific time zone increases because the user moves to work at the start time on weekdays, so the occupancy rate for each time zone By classifying similar data, weekdays and non-weekdays can be distinguished.

  FIG. 11 is a diagram showing an example of data of a user who is a so-called weekend driver who does not use a car for commuting. 11 also shows the number of times of boarding for each day of the week and the number of appearances of each day of the week based on the movement history information. The boarding rate (movement frequency) for each day of the week shown in FIG. 12 is obtained by dividing the number of times of boarding for each day of the week in FIG. 11 by the number of appearances of each day of the week.

Then, the boarding ratio for each day of the week in FIG. 12 is used as a feature vector, and clustering is performed according to the distance between the feature vectors as described above, so that it can be classified into two clusters (groups) as shown in FIG. Also in the example of FIG. 13, the data for Monday, Tuesday, Wednesday, Thursday, and Friday are classified into one cluster, and the data for Saturday and Sunday are classified into the other cluster. Accordingly, it can be estimated that Monday, Tuesday, Wednesday, Thursday, and Friday are weekdays and Saturday and Sunday are non-weekdays for the user.

  In this way, for users who do not commute by car, the daytime boarding rate is low on weekdays, the boarding rate is high only at night, the daytime boarding rate is high on holidays, and the difference between day and night tends to be small. By classifying data with similar occupancy rates for each time zone, weekdays and non-weekdays can be distinguished.

  As described above, in the present embodiment, the data on the boarding rate for each day of the week is clustered by the Ward method according to the Euclidean distance, so that the data having similar boarding rates for each time zone is classified into two. Thus, the data for each day of the week is classified into a weekday cluster (group) and a non-weekday cluster (group). Not limited to this, the weekday / non-weekday estimation apparatus 1 of the present embodiment may use another distance function or a clustering technique.

  According to the present embodiment, weekdays / non-weekdays can be discriminated based on the tendency of the occupancy rate (movement frequency) for each day of the week, so that weekdays / non-weekdays are set appropriately without fixedly setting weekends as non-weekdays. it can. For example, a weekday / non-weekday can be appropriately estimated even in a lifestyle where non-weekdays are non-weekdays, such as when Monday and Wednesday are non-weekdays or only Thursday is a non-weekday.

<Modification>
In the above embodiment, weekdays / non-weekdays are determined except for the holiday data, but the occupancy rate (movement frequency) is obtained for the holiday data in the same way as the data for one day of the week, and the holiday is included in the data for the seven days of the week. Clustering may be performed for eight data.
For example, as shown in FIG. 14, one column of holiday data is added to the data for each day of the week in seven columns to obtain eight columns of data. The time zone separator for each row is the same as the data for other days of the week.
In this way, holiday data is added and clustered into two clusters (groups), and the day of the week or holiday classified as a cluster (group) with a large number of elements is classified into a cluster (group) with a small number of elements on weekdays. The day of the week or holiday is estimated as a non-weekday.
This makes it possible to determine whether the holiday is a weekday or a non-weekday.

<< Embodiment 2 >>
In the second embodiment, in addition to the above-described estimation process, a process for estimating the area to which the user's home belongs and the area to which the work belongs is performed. For example, based on the movement history when moving from the home to the work, An example of estimating weekdays is shown.

<Functional explanation>
FIG. 15 is a functional block diagram of the weekday / non-weekday estimation apparatus according to the second embodiment. The weekday / non-weekday estimation device 100 of the second embodiment is a device (on-vehicle device) mounted on a vehicle such as an automobile (passenger car, truck, motorcycle, etc.), a motorbike, a light vehicle, etc., and a position information acquisition unit 11, a history storage unit 120, and an estimation unit 130.

The position information acquisition unit 11 is a sensor capable of acquiring position information such as a GPS (Global Positioning System) receiver, and receives position information of a vehicle on which the weekday / non-weekday estimation device 100 is mounted from a plurality of GPS satellites. Acquire based on the signal. Further, the position information acquisition unit 11 may acquire the position not only by GPS but also by autonomous navigation using an acceleration sensor.

  The history storage unit 120 divides an area in which the vehicle moves into a plurality of areas, determines an area in which the vehicle is located based on position information, and represents movement of the area in which the vehicle is located as the vehicle moves Holds history information.

  The estimation unit 130 includes an area attribute estimation unit 13A and a weekday / non-weekday estimation unit 13B. The area attribute estimation unit 13A obtains the variation in visit frequency and arrival time for each area based on the movement history information, and the area to which the home belongs is the area where the variation in arrival time is the largest among the areas where the visit frequency is equal to or greater than the threshold. (Home area). Moreover, the estimation part 13 estimates the area where the dispersion | variation in arrival time is the smallest among the areas whose visit frequency is more than a threshold as an area to which the destination belongs. The destination in this example is a place where a user (driver) visits habitually with the vehicle, for example, the work place of the user. The destination may be an employee place engaged in a business of a store or company managed by the user. Further, it may be a place where a club activity or volunteer activity is performed, a school destination, a hospital destination, or the like. In the present application, a destination where these users regularly go is also referred to as a work place for convenience, and an area to which the destination belongs is also referred to as a work area.

In the second embodiment, a regional mesh defined by Japanese Industrial Standard JIS X0410 is used as this area. FIG. 16 is an explanatory diagram of a regional mesh. The regional meshes are a primary mesh (also referred to as “first regional partition”), a secondary mesh (also referred to as “secondary regional partition”, “integrated regional mesh”), and a tertiary mesh (“third region”). The size of each step of the standard area mesh such as “next area section” and “reference area mesh”, and the divided area mesh such as the half area mesh, the quarter area mesh, and the eighth area mesh. It is composed of meshes and is specified by a mesh code indicating each mesh.

  The primary mesh is obtained by dividing an area into a substantially rectangular area having a side length of about 80 km, and the primary mesh code is an upper two-digit number representing latitude and a lower 2 representing longitude. It is expressed as a 4-digit number consisting of a numeric value. The secondary mesh is obtained by dividing the primary mesh into eight equal parts in the east-west direction and the north-south direction, respectively, and the secondary mesh code is the first one digit representing the latitude direction and the last one digit representing the longitude direction. This is expressed in a format in which a two-digit number consisting of the following numbers is connected to the primary mesh code by “-(hyphen)”. The tertiary mesh is obtained by dividing the secondary mesh into 10 equal parts in the east-west direction and the north-south direction, respectively. The tertiary mesh code is a first digit representing the latitude direction and a last digit representing the longitude direction. This is expressed in a form in which a two-digit number consisting of a number of 1 is connected to a primary mesh code and a secondary mesh code. The half-region mesh is the third mesh divided into two equal parts in the east-west direction and north-south direction. The half-region mesh code is 1 for the southwest region, 2 for the southeast region, and northwest. The area is 3 and the area in the northeast is 4 and is expressed in a form connected to the primary to tertiary mesh codes. The quarter region mesh is a halve of the half region mesh in the east-west direction and the north-south direction. The quarter region mesh code is 1 for the southwest region and 2 for the southeast region. , The northwest region is 3 and the northeast region is 4, and is expressed in a format connected to the primary to half-region mesh codes. One-eighth region mesh is a quarter-region mesh divided into two equal parts in the east-west direction and north-south direction. The eighth-region mesh code is 1 for the southwest region and 2 for the southeast region. , The northwest region is 3 and the northeast region is 4, and is expressed in a format connected to the primary to quarter region mesh codes.

  The weekday / non-weekday estimation unit 13B extracts the movement history related to the area specified by the area attribute estimation unit 13A, and performs the estimation of the weekday / non-weekday based on the extracted movement history. The weekday / non-weekday estimation process by the weekday / non-weekday estimation unit 13B is the same as the estimation process by the estimation unit 13 of the first embodiment.

The history storage unit 120 obtains the mesh code of the mesh where the weekday / non-weekday estimation device 100 is located from the position information including the latitude and longitude acquired by the position information acquisition unit 11. That is, position information including latitude and longitude is converted into a regional mesh code. Note that the process of converting the position information into the regional mesh code can be performed using an existing algorithm, and thus details are omitted. Further, the area in the present embodiment is not limited to an existing area mesh, and a mesh that is divided into a mesh shape having a position and size uniquely determined for the area may be used.

  By using the regional mesh code in this way, it is possible to reduce the cost for updating and managing the map information as compared with the case of using the map information.

  FIG. 17 is a diagram illustrating an example of movement history information held by the history storage unit 120. As shown in FIG. 17, the movement history information stores arrival date / time, arrival area, departure date / time, destination area, and the like in association with each other. The arrival date and time is information indicating the date and time of arrival in the arrival area. The arrival area is a mesh code indicating the area that has arrived. The departure date and time is information indicating the date and time of movement from the arrival area. The movement destination area is a mesh code indicating the area after movement.

  FIG. 18 is a diagram illustrating an example of a feature amount obtained by the estimation unit 13A based on the movement history information, and FIG. 19 is an explanatory diagram of an observation date when the feature amount is obtained. As illustrated in FIG. 18, the estimation unit 13A of the present embodiment obtains a boarding frequency (boarding rate), a visit frequency (visit rate), an average staying time, and an arrival time information amount (arrival time entropy) as feature amounts. Yes.

  The boarding frequency (boarding rate) is the frequency that the user gets on the vehicle, and is a value (unit: days) obtained by dividing the number of days (boarding days) in the observation period by the number of days (observation days) in the observation period. In this example, since the number of times of boarding is counted every day to obtain the number of boarding days, the number of boarding days is counted as 1 even when boarding a plurality of times a day. The number of observation days is the number of days to acquire data for calculating the feature quantity, and the data is acquired by going back a day having valid data from the starting day as shown in FIG. 5 by a predetermined number of days. . FIG. 19 shows a case where the area attribute is determined on July 25. Since the data for that day is not yet prepared, the data is acquired retroactively for a predetermined number of days (for example, seven days) from July 25. To do. At this time, holidays and long vacation periods are often excluded because they often take different actions and may not be valid data. Note that the long vacation period is New Year's holiday or Golden Week, and is arbitrarily set. In the example of FIG. 19, since July 21 is a holiday, it is excluded, and seven days from July 17 to 20 and 22 to 24, which are valid data, are set as the observation period. And since the 17th, 18th, 22nd to 24th of July during this observation period are the days of boarding, the number of boarding days (5 days) / number of observation days (7 days) = boarding rate (0. 714...

  The visit frequency is a value (unit: days) obtained by dividing the number of days of visiting each area (mesh) by the number of observation days. In this example, since the number of visits is counted every day, the number of days visited is counted as 1 even when visiting multiple times a day. As shown in FIG. 19, if the number of observation days is 7, and the number of visit days is 4, for example, the number of visit days (4 days) / the number of observation days (7 days) = visit frequency (0.571...).

  The average stay time is a value indicating the stay time in each area, and in this example, two types of first stay time T1 and second stay time T2 are used. The first stay time T1 is calculated by counting the difference between the departure date and arrival date and time (departure date and time-arrival date and time) for each mesh code as the time spent in the area (mesh) and dividing by the number of visit days. The second stay time T2 is calculated by counting the difference between the departure date and arrival date (departure date-arrival date) for each mesh code as the time spent in the area (mesh) and dividing by the number of observation days.

The arrival time information amount is the information amount of time distribution arriving at the area (mesh) (unit: bit)
It is also a value indicating the variation in arrival time, and takes a large value if the arrival time variation is large, and takes a small value if the arrival time variation is small. For example, a day is divided into a predetermined number of time zones (in this example, 24 time zones in increments of 1 hour), and the number of times of arrival in this area after leaving another area not adjacent to each time zone is counted. The distribution was determined.
In other words, the frequency distribution was obtained by excluding cases of movement within an area or movement between adjacent areas. In calculating the frequency distribution, a pseudo count may be added to the value of each time zone. For example, in order to be able to quickly reflect in the estimation results that the home or work place has changed due to moving or moving, it is necessary to set the observation period to be short and to be able to estimate from the data within this short observation period. There is. However, if the observation period is set to be short, the appearance frequency of an event with a low appearance frequency becomes almost zero, which is not correctly reflected in the estimation result, and a so-called zero frequency problem occurs. Therefore, the pseudo count is added to the value of each time zone to correct the appearance frequency of the event having a low appearance frequency, thereby avoiding the zero frequency problem. Note that the value of the pseudo count may be changed according to the length of the observation period, the range (width) of the feature amount used for estimation, variation, and the like. It may be obtained and stored as a relational expression or a data table, and a pseudo count value corresponding to an observation period or a feature amount may be used during estimation processing.

  FIG. 6 is a diagram illustrating an example of calculating the arrival time information amount. In the example of FIG. 6, the observation days are 20 days, the number of arrivals is counted in 24 time zones from 0 to 23:00, incremented by 1, and the logarithm of the logarithm is calculated using equation (1). The calculation result when the base is 2 is obtained as the arrival time information amount. The arrival probability Pt is a value obtained by dividing the number of arrivals including the pseudo count by the total. Further, the departure time information amount can be similarly obtained using Equation 1.

<Device configuration>
The weekday / non-weekday estimation apparatus 100 of the second embodiment is a computer as shown in FIG. 4, for example, as in the first embodiment.

<Estimation method>
Next, details of processing executed by the weekday / non-weekday estimation apparatus 100 according to a program will be described. The weekday / non-weekday estimation apparatus 100 stores a movement history as the user moves, and estimates a home area, a work area, and the like based on the movement history.

  FIG. 21 is a diagram illustrating an example of a process for storing a movement history. The weekday / non-weekday estimation apparatus 100 starts the process of FIG. 21 at a predetermined timing. For example, the processing in FIG. 21 may be started periodically every hour or every 15 minutes, or when the vehicle starts moving or stops, the processing in FIG. 21 may be performed. For example, when the ignition switch is turned on from off or when the vehicle speed exceeds a predetermined value, the mesh code at the current position and the departure time are stored in association with each other, and when the ignition switch is turned off from on The mesh code at the current position and the arrival time are stored in association with each other.

  When the processing of FIG. 21 is started, the weekday / non-weekday estimation apparatus 100 first uses the GPS receiver as the position information acquisition unit 11 to determine the latitude and longitude of the position where the weekday / non-weekday estimation apparatus 100 exists as vehicle position information. Obtain (step S210).

  Further, the weekday / non-weekday estimation device 100 converts the acquired latitude and longitude into a mesh code by the function of the history storage unit 120, and specifies the area where the vehicle is located (step S220).

  Then, the weekday / non-weekday estimation device 100 associates the mesh code with the time by the function of the history storage unit 120 to obtain movement history information as shown in FIG. 17 and stores it in the storage device (step S230). Here, the history storage unit 120 stores the time of arrival at the position as the arrival date and time, and the time of movement from the position as the departure date and time.

  FIG. 22 is a diagram illustrating an example of the estimation process. The estimation unit 13A of the weekday / non-weekday estimation apparatus 100 executes the process of FIG. 22 at a predetermined timing. For example, the processing in FIG. 22 may be started periodically every 6 hours or every 24 hours, or when a request is received from another application or system, or when the user selects activation, as shown in FIG. The process may be started.

  When the process of FIG. 22 is started, the weekday / non-weekday estimation apparatus 100 first reads data for a predetermined number of observation days from the storage device and calculates the feature amount as shown in FIG. 18 (step S110).

  Next, the weekday / non-weekday estimation apparatus 100 determines whether or not the calculated boarding rate is equal to or greater than a predetermined threshold (step S320). If the weekday / non-weekday estimation device 100 determines that the boarding rate is equal to or higher than the threshold (step S320, Yes), the home area estimation and the work area estimation (step S320) are performed using the feature amount calculated in step S310. If the boarding rate is determined to be less than the threshold value (No at Step S320), the home area is estimated (Step S350) without estimating the work area. That is, if the occupancy rate is high, it is determined that the vehicle is used for commuting, and both the home area and the work area are estimated. If the occupancy rate is low, it is determined that the vehicle is not used for commuting. Estimate only the area.

  Then, when both the home area and the work area are estimated (step S320, Yes), the movement history related to the home area and the work area is extracted from the movement history information, and based on this movement history, Non-weekdays are estimated (step S340). The movement history related to the home area and the work area includes arrival time and departure time of the home area, arrival time and departure time of the work area, travel time from departure from the home area to arrival at the work area, This includes the travel time from departure from the work area to arrival at the home area. For example, regarding the travel time from the home area to the work area and the travel time from the work area to the home area, the number of boarding times in the time zone required for the travel time is counted as shown in FIG. 9 to obtain the boarding rate. Used for estimation processing on weekdays / non-weekdays. In addition, as for the arrival time when arriving from the home area and arriving at the work area and the arrival time when arriving from the work area and arrives at the home area, the arrival time belongs as shown in FIG. The number of times of boarding in the time zone may be counted, and the boarding rate may be obtained and used for the weekday / non-weekday estimation process. Since the weekday / non-weekday estimation procedure is the same as that in the first embodiment, the description thereof will be omitted.

  When only the home area area is estimated (step S320, No), the movement history related to the home area is extracted from the movement history information, and weekdays / non-weekdays are estimated based on the movement history (step S360). ). For example, for the travel time from the home area to another area and the travel time from the other area to the home area, the number of boarding times in the time zone required for the travel time is counted as shown in FIG. 9 to obtain the boarding rate. Used for estimation processing on weekdays / non-weekdays. Further, as for the arrival time when arriving from the home area and arriving at another area and the arrival time when arriving from the other area and arrived at the home area, the arrival time belongs as shown in FIG. The number of times of boarding in the time zone may be counted, and the boarding rate may be obtained and used for the weekday / non-weekday estimation process.

  FIG. 23 is a diagram showing details of the home area and work place estimation process (step S330). When the process of FIG. 23 is started, the weekday / non-weekday estimation apparatus 100 first determines whether or not the visit frequency of each area and the average stay times T1 and T2 exceed a predetermined threshold (step S410). ). In step S410 of this example, it is a condition that the visit frequency and average stay time T1, T2 of each area exceed the threshold, but not limited to this, the visit frequency of each area and the average stay time T1, It is good also as a condition that the visit frequency of each area, average stay time T2, or only visit frequency exceeds a threshold value.

Then, the weekday / non-weekday estimation apparatus 100 estimates the area having the maximum arrival time information amount as the home area among the areas satisfying the condition of step S410 (step S420). Furthermore, the weekday / non-weekday estimation apparatus 100 estimates an area having the smallest amount of arrival time information as an office area among the areas that satisfy the condition of step S410 (step S425). The visit rate is extremely high in the home area and the work area, and the amount of arrival time information is small at a work place that often goes to work at a fixed time, and is large at a home where the arrival time is not fixed. For this reason, an area where the visit frequency is equal to or greater than the threshold and the arrival time information amount is maximum is estimated as a home area, and an area where the visit frequency is equal to or greater than the threshold and the arrival time information amount is minimum is estimated as a work area.

  On the other hand, if it is determined in step S410 that the above condition is not satisfied, the previous estimation result is taken over without estimating the home area and work area (step S430).

  FIG. 24 is a diagram showing details of the home area estimation process (step S350). When the process of FIG. 24 is started, the area attribute estimation unit 13A of the weekday / non-weekday estimation apparatus 100 first determines whether the visit frequency and average stay time T2 of each area exceed a predetermined threshold. (Step S510). In step S510 of this example, the condition is that the visit frequency and average stay time T2 of each area exceed the threshold, but not limited to this, the visit frequency and average stay times T1, T2, Alternatively, only the visit frequency may exceed the threshold.

  Then, the area attribute estimation unit 13A estimates the area having the maximum amount of arrival time information as the home area among the areas satisfying the condition of Step S210 (Step S520).

  On the other hand, if it is determined in step S310 that the above condition is not satisfied, the area attribute estimation unit 13A takes over the previous estimation result without estimating the home area (step S330).

  As described above, according to the second embodiment, the area attribute estimation unit 13A specifies the home area or the work area according to the movement history of the vehicle, and the weekday / non-weekday estimation unit 13B relates to this specific area. Since weekdays / non-weekdays are estimated based on the movement history information, it is possible to accurately estimate weekdays / non-weekdays by eliminating noise.

  In the first and second embodiments, an example in which the weekday / non-weekday estimation device is realized by an in-vehicle device is shown. However, the present invention is not limited thereto, and the weekday / non-weekday estimation device includes the position information acquisition unit 11 and the history storage unit 12. , 120 and the estimation units 13 and 130, other configurations may be used. For example, the in-vehicle device has the position information acquisition unit 11 and the history storage units 12 and 120, and the computer having the estimation units 13 and 130 acquires the movement history information from the in-vehicle device via the network and performs estimation. There may be.

<< Embodiment 3 >>
In the third embodiment, in addition to the estimation process of the first embodiment, the user terminal carried by the user detects the movement of the user as a moving body, and based on this movement history, the area to which the home belongs or the area to which the work belongs An example in which a weekday / non-weekday is estimated based on a movement history when moving from home to work, for example, is shown.

<Functional explanation>
FIG. 25 is a functional block diagram of the weekday / non-weekday estimation apparatus 110 according to the third embodiment. The weekday / non-weekday estimation device 110 according to the present embodiment includes a mobile phone, a smartphone, a tablet terminal (slate PC (Personal Computer)), a notebook PC,
A user terminal (mobile device) 1 such as a mobile phone or a portable game machine, and includes a position information acquisition unit 11, a history storage unit 120, and an estimation unit 133.

The position information acquisition unit 11 is a sensor capable of acquiring position information, such as a GPS (Global Positioning System) receiver, and is position information indicating the position of the user as the position information of the user terminal 110 that is carried by the user and moves with the user. Get as. Further, the position information acquisition unit 11 may acquire the position not only by GPS but also by autonomous navigation using an acceleration sensor.

  The history storage unit 120 divides an area where the user moves into a plurality of areas, determines an area where the user is located based on position information, and represents a transition of the area where the user is located as the user moves. Holds history information.

  The estimation unit 133 includes an area attribute estimation unit 13C and a weekday / non-weekday estimation unit 13D. The area attribute estimation unit 13C obtains at least the visit frequency for each area based on the history information, and estimates the area with the highest visit frequency as the area to which the user's home belongs (home area). Moreover, the estimation part 13 estimates the area with the largest amount of movement within an area among the areas where the visit frequency is equal to or greater than the threshold as the area to which the user's destination belongs. Here, the user's destination is a place where the user customarily visits, for example, a work place. The destination may be an employee place engaged in a business of a store or company managed by the user. Further, it may be a place where a club activity or volunteer activity is performed, a school destination, a hospital destination, or the like. In the present application, a destination where these users regularly go is also referred to as a work place for convenience, and an area to which the destination belongs is also referred to as a work area. In the third embodiment, the area mesh defined by JIS X0410 of the Japanese Industrial Standard is used as this area as in the second embodiment.

  The weekday / non-weekday estimation unit 13D extracts a movement history related to the area specified by the area attribute estimation unit 13C, and performs a weekday / non-weekday estimation based on the extracted movement history. The weekday / non-weekday estimation process by the weekday / non-weekday estimation unit 13D is the same as the estimation process by the estimation unit 13 of the first embodiment.

  The history storage unit 120 obtains the mesh code of the mesh where the user terminal 110 is located from the position information including the latitude and longitude acquired by the position information acquisition unit 11. That is, position information including latitude and longitude is converted into a regional mesh code. Note that the process of converting the position information into the regional mesh code can be performed using an existing algorithm, and thus details are omitted. Further, the area in the present embodiment is not limited to an existing area mesh, and a mesh that is divided into a mesh shape having a position and size uniquely determined for the area may be used.

  By using the regional mesh code in this way, it is possible to reduce the cost for updating and managing the map information as compared with the case of using the map information.

  Since the weekday / non-weekday estimation device of the third embodiment is a user terminal (mobile device) 110 such as a smartphone, it is activated by turning on the ignition switch of the vehicle at the time of departure, unlike a conventional in-vehicle device, The operation is not stopped when the vehicle arrives at the destination and the ignition switch of the vehicle is turned off, and it is difficult to detect departure or arrival. Therefore, in the third embodiment, it is assumed that the position information is periodically acquired and moved when the distance is more than a predetermined distance compared to the previous position information, and moved when the distance is not more than the predetermined distance. Ask for departure time and arrival time every time you move as if you are not staying. For example, it is assumed that the user has moved when position information at a position 10 m or more away from the previous position information is detected, and the arrival time at the position detected this time and the departure time of the position detected last time are stored. Specific position information (latitude, longitude, etc.) used for this comparison may be stored as history information together with the mesh code, or may be held in the memory for a predetermined period.

The position detected in the predetermined period is not limited to the comparison with the previously detected position, and the position detected within the predetermined period may be compared with the position detected this time, and may be moved when the distance is more than a predetermined distance. In this case, the distance from the center of gravity of the position detected within the predetermined period to the position detected this time, the distance from the center of the mesh to which the position detected within the predetermined period and the representative points such as the four corners to the position detected this time, It is good also as what moved when it was away from the predetermined distance or more.

  Further, the departure time and arrival time are not limited to each detection position, but may be determined for each area or each destination. For example, it is assumed that the location information is periodically acquired and moved when the area to which the current location information belongs is different from the area to which the previous location information belongs, and is not moved when it is the same. As a thing (staying), the departure time and the arrival time are obtained every time it moves. For example, when the previous time is area A and this time is area B, it is assumed that the user has moved from area A to area B, and the arrival time to area B and the departure time of area A are stored.

  Here, when the work place exists over a plurality of areas (for example, areas A and B), it may be determined that the work place is moving despite staying at the work place. For this reason, an area that has been visited within a predetermined period may be regarded as one area (area group), and the arrival time and departure time of this area may be stored. For example, area A → Area B → Area A → Area B → Area C → Area D is first located in Area A, and after coming and going between Area A and Area B, it is adjacent to Area B When moving to area C and then moving to area D adjacent to area C, areas A and B are regarded as one area, and the first time in area A is taken as the arrival time. The time of moving to the area C that is not adjacent, or the time of moving to the area D that is not adjacent to the areas A and B may be set as the departure times of the areas A and B.

  Furthermore, according to the staying time of each area, an area or a group of areas staying for a predetermined time (for example, 1 hour) or more may be set as a destination (staying place), and departure time and arrival time to the destination may be stored. For example, when moving from area A to area F via areas B, C, D, and E, the stay time in areas A and F is longer than a predetermined time, and the stay time in areas B, C, D, and E is predetermined. If it is less than the time, the area F is the destination for the area A, the time of moving from the area A to the area B is the departure time of the area A, and the time of arrival of the area F via the area E is the arrival time of the area F As for areas B, C, D, and E, storage of departure time and arrival time may be omitted. Thus, by using the departure time and arrival time limited to the destination (stay place), it is possible to improve the accuracy of estimation described later. The departure time and arrival time to the destination, the departure time and arrival time for each area, and the departure time and arrival time for each detection position may be stored, and any departure time and arrival time can be estimated. It may be selected as appropriate for use.

  FIG. 26 is a diagram illustrating an example of a feature amount obtained by the estimation unit 133 based on history information, and FIG. 27 is an explanatory diagram of an observation date when the feature amount is obtained. As shown in FIG. 26, the estimation unit 13 of the present embodiment obtains a visit frequency, an average stay time, an average number of movements, a departure time information amount, and an arrival time information amount as feature amounts.

The visit frequency is a value (unit: days) obtained by dividing the number of days of visiting the area (mesh) by the number of observation days. In this example, since the number of visits is counted every day, the number of days visited is counted as 1 even when visiting multiple times a day. The number of observation days is the number of days to acquire data for calculating the feature amount, and the data is acquired by going back a day having valid data from the starting day as many days as shown in FIG. . In the example of FIG. 27, when the area attribute is determined on July 25, since the data for the current day is not yet prepared, the data is acquired retroactively for a predetermined number of days (for example, six days) from July 25. At this time, holidays and long vacation periods are often excluded because they often take different actions and may not be valid data. Note that the long vacation period is New Year's holiday or Golden Week, and is arbitrarily set. Moreover, the day when the user terminal 110 was not normally used, for example, the day when it was not carried by the user or the day when the power was turned off is excluded. In the example of FIG. 27, July 21 is excluded because it is a public holiday, July 24 is excluded because it is a power-off day, and data for 6 observation days is acquired retroactively to July 17. To do. That is, in an area where the number of visit days from July 17 to July 23 is 5, the number of visit days (5 days) / the number of observation days (6 days) = visit frequency (0.833...).

  The average stay time is calculated by summing up the difference between the departure date and time and the arrival date and time (departure date and time-arrival date and time) for each mesh code as the time spent in the area (mesh) and dividing by the number of visit days.

Departure time information amount is the information amount of time distribution that departs from the area (mesh) (unit: bit)
It is also a value indicating the variation in departure time, and takes a large value if the variation in departure time is large, and takes a small value if the variation in departure time is small. For example, a day is divided into a predetermined number of time zones (in this example, 24 time zones in increments of 1 hour), and the number of departures from that area to another non-adjacent area is counted for each time zone to obtain a frequency distribution. It was. In other words, the frequency distribution was obtained by excluding cases of movement within an area or movement between adjacent areas. In calculating the frequency distribution, a pseudo count may be added to the value of each time zone. For example, in order to be able to quickly reflect in the estimation results that the home or work place has changed due to moving or moving, it is necessary to set the observation period to be short and to be able to estimate from the data within this short observation period. There is. However, if the observation period is set to be short, the appearance frequency of an event with a low appearance frequency becomes almost zero, which is not correctly reflected in the estimation result, and a so-called zero frequency problem occurs. Therefore, the pseudo count is added to the value of each time zone to correct the appearance frequency of the event having a low appearance frequency, thereby avoiding the zero frequency problem. Note that the value of the pseudo count may be changed according to the length of the observation period, the range (width) of the feature amount used for estimation, variation, and the like. It may be obtained and stored as a relational expression or a data table, and a pseudo count value corresponding to an observation period or a feature amount may be used during estimation processing.

The arrival time information amount is the information amount of time distribution arriving at the area (mesh) (unit: bit)
It is also a value indicating the variation in arrival time, and takes a large value if the arrival time variation is large, and takes a small value if the arrival time variation is small. For example, a day is divided into a predetermined number of time zones (in this example, 24 time zones in increments of 1 hour), and the number of times of arrival in this area after leaving another area not adjacent to each time zone is counted. The distribution was determined. In other words, the frequency distribution was obtained by excluding cases of movement within an area or movement between adjacent areas. In calculating the frequency distribution, a pseudo count may be added to the value of each time zone. For example, in the example of FIG. 20 described above, the number of observation days is set to 20 days, the number of arrivals is counted in 24 time zones in 1-hour increments from 0:00 to 23:00, the pseudo count is added, and the formula (1) is calculated. The calculation result when the base of the logarithm is set to 2 is obtained as the arrival time information amount.

<Device configuration>
The weekday / non-weekday estimation apparatus 110 of the third embodiment is a computer as shown in FIG. 4, for example, as in the first embodiment.

<Estimation method>
Next, details of processing executed by the user terminal 110 according to the program will be described. The user terminal 110 stores a movement history as the user moves, and estimates a home area, a work area, and the like based on the movement history.

  FIG. 28 is a diagram illustrating an example of processing for storing a movement history. The user terminal 110 starts the process of FIG. 28 at a predetermined timing. For example, the processing in FIG. 28 may be started periodically every hour or every 15 minutes, or when the movement of the user terminal 110 is detected by an acceleration sensor, the processing in FIG. 28 may be started. .

  When the processing of FIG. 28 is started, the user terminal 110 first acquires the latitude and longitude of the position where the user terminal 110 exists as the position information of the user by the GPS receiver as the position information acquisition unit 11 (step S610).

  Further, the user terminal 110 converts the acquired latitude and longitude into a mesh code by the function of the history storage unit 120, and specifies the area where the user is located (step S620).

  Then, the user terminal 110 associates the mesh code with the time by the function of the history storage unit 120 to obtain history information as shown in FIG. 17 and stores it in the storage device (step S630). Here, the history storage unit 120 stores the time of arrival at the position as the arrival date and time, and the time of movement from the position as the departure date and time.

  FIG. 29 is a diagram illustrating an example of the estimation process. The estimation unit 13 of the user terminal 110 executes the process of FIG. 29 at a predetermined timing. For example, the processing of FIG. 29 may be started periodically every 6 hours or every 24 hours, or when a request is received from another application or system, or when the user selects activation, FIG. The process may be started.

  When the process of FIG. 29 is started, the user terminal 110 first reads data for a predetermined number of observation days from the storage device, and calculates the feature amount as shown in FIG. 26 (step S710).

  Further, the user terminal 110 merges the feature amounts of adjacent areas based on the calculated feature amount (step S720). For example, the feature amounts (vectors) calculated in step S110 are arranged in descending order of visit frequency. For an area (mesh code) whose visit frequency is second or lower, it is checked whether there is an adjacent area among those having a higher rank than the area. If there are adjacent areas, they are merged as one area having the same attribute as when, for example, a work place extends over a plurality of areas, and the feature values of this area are also merged. The mesh code of the merged area is the mesh code of the area having the highest visit frequency rank among the plurality of merged areas. When merging visit frequencies, a set sum is taken for each visit date (arrival date), and the sum is summed and divided by the number of observation days to obtain a visit frequency. When merging the average stay time, the values before dividing by the number of visit days are totaled and then divided by the visit days to obtain the average stay time. When merging the average number of movements, return to the history information, count the number of movements in the merged area for each observation date, add them, and divide by the number of visits to obtain the average number of movements. When merging departure time information amounts, return to the history information, find the departure time to leave the merged area, divide the day into a predetermined number of time zones, and depart from the merged area to other areas for each time zone The frequency distribution is obtained using the equation (1) in FIG. 20 as described above. When merging the amount of arrival time information, return to the history information, find the arrival time that arrived in the merged area, divide the day into a predetermined number of time zones, and arrived at the merged area from other areas for each time zone The number of times is counted, and the frequency distribution is obtained using Equation 1 as described above.

  The user terminal 110 performs home area estimation (step S730) and work area estimation (step S740) using the feature amounts calculated in step S110 and step S720.

And the movement history which concerns on a home area and a work area is extracted from the said movement history information, and a weekday / non-weekday is estimated based on this movement history (step S750). The movement history related to the home area and the work area includes arrival time and departure time of the home area, arrival time and departure time of the work area, travel time from departure from the home area to arrival at the work area, This includes the travel time from departure from the work area to arrival at the home area. For example, regarding the travel time from the home area to the work area and the travel time from the work area to the home area, the number of boarding times in the time zone required for the travel time is counted as shown in FIG. 9 to obtain the boarding rate. Used for estimation processing on weekdays / non-weekdays. In addition, as for the arrival time when arriving from the home area and arriving at the work area and the arrival time when arriving from the work area and arrives at the home area, the arrival time belongs as shown in FIG. The number of times of boarding in the time zone may be counted, and the boarding rate may be obtained and used for the weekday / non-weekday estimation process. Since the weekday / non-weekday estimation procedure is the same as that in the first embodiment, the description thereof will be omitted.

  FIG. 30 is a diagram illustrating an example of a home area estimation process. First, the user terminal 110 determines whether or not the visit frequency of each area exceeds a predetermined threshold (step S810).

  Then, the user terminal 110 estimates that the area with the highest visit frequency among the areas where the visit frequency exceeds the threshold as the home area (step S820).

  On the other hand, if it is determined in step S810 that none of the areas exceeds the threshold, the previous estimation result is taken over without estimating the home area (step S830).

  FIG. 31 is a diagram illustrating an example of work area estimation processing. First, the user terminal 110 determines whether or not the visit frequency of each area exceeds a predetermined threshold (step S910). The threshold value in step S910 may be set lower than the threshold value in step S910 described above.

  Then, the user terminal 110 estimates the area where the average number of movements is the largest among the areas where the visit frequency exceeds the threshold and is not the home area in step S730 as the work area (step S920). ).

  On the other hand, when it is determined in step S910 that none of the areas exceeds the threshold, the previous estimation result is taken over without estimating the work area (step S930).

  As described above, according to the third embodiment, the area attribute estimation unit 13C specifies the home area or the work area according to the movement history of the user, and the weekday / non-weekday estimation unit 13D relates to this specific area. Since weekdays / non-weekdays are estimated based on the movement history information, it is possible to accurately estimate weekdays / non-weekdays by eliminating noise.

<Modification 3-1>
In a general lifestyle, the visit frequency of the home area is the highest as in the third embodiment. However, in the case of a lifestyle that involves, for example, railways, hospitals, police, and fire fighting, and takes a nap at work, the frequency of home area visits may be low. Therefore, the home area may be estimated using the arrival time information amount.

  FIG. 32 is a diagram illustrating an example of a home area estimation process according to Modification 3-1. In addition, since this modified example 3-1 differs in the structure using the arrival time information amount compared with the above-mentioned Embodiment 3, and the other structure is the same, the overlapping description is abbreviate | omitted. The user terminal (weekday / non-weekday estimation device) 110 in this example determines whether or not the visit frequency of each area exceeds a predetermined threshold (step S810), and the user terminal of the area where the visit frequency exceeds the threshold is determined. Among these, an area having an arrival time information amount equal to or larger than a predetermined threshold is extracted (step S815), and the extracted area having the highest visit frequency is estimated as a home area (step S820).

  On the other hand, if it is determined in step S810 that none of the areas exceeds the threshold, the previous estimation result is taken over without estimating the home area (step S830).

  As described above, in the present modified example 3-1, the home area is estimated when the arrival time information amount is equal to or greater than the threshold and the visit frequency is maximum. Since the amount of arrival time information is a value indicating the variation in time of arrival in the area, it is small in an area where employees work at a fixed time, such as a work place, and large in an area where arrival time is not defined, such as at home. is expected. By estimating the home area using the arrival time information amount being equal to or greater than the threshold, the home area can be estimated with higher accuracy.

  In the modification 3-1, an area where the arrival time information amount is equal to or larger than the predetermined threshold is extracted in step S815. However, the present invention is not limited to this, and an area where the arrival time information amount is larger than the departure time information amount is extracted. Alternatively, an area in which the ratio of the arrival time information amount to the departure time information amount (arrival time information amount / departure time information amount) is equal to or greater than a threshold value may be extracted.

<Modification 3-2>
As described above, the frequency of visits to the home area and the work area is usually high, but the frequency of visits to the nearest station and transfer station may also be high. In addition, if a shopping mall is added to these stations, the average number of movements may be high. However, these are passing points, and the staying time is different from the home area or work area. For this reason, you may estimate a home area or a work area using average stay time.

  FIG. 33 is a diagram illustrating an example of the estimation process according to Modification 3-2. In addition, since this modified example 3-2 differs in the structure using the average stay time compared with the above-mentioned Embodiment 3, and the other structure is the same, the overlapping description is abbreviate | omitted.

  When the area attribute estimation unit 13C of the user terminal (weekday / non-weekday estimation device) 110 executes the processing of FIG. 33 at a predetermined timing, the feature amount is calculated in the same manner as described above (step S710), and the feature of the adjacent area is calculated. The amounts are merged (step S720), and an area where the average stay time is equal to or greater than a predetermined threshold is extracted (step S725).

  Then, the user terminal 110 performs home area estimation (step S730), work area estimation (step S740), and weekday / non-weekday estimation using the feature amount of the area extracted in step S125.

  As described above, in this modified example 3-2, the home area and the work area are estimated for the area where the average stay time is equal to or greater than the threshold value, so that data of simple passing points can be eliminated and the home area and the work area are accurately obtained. Can be estimated.

<Others>
The present invention includes a computer program that executes the above-described processing. Furthermore, a computer-readable recording medium on which the program is recorded also belongs to the category of the present invention. The recording medium on which the program is recorded can be estimated by causing the computer to read and execute the program on the recording medium.

  Here, the computer-readable recording medium refers to a recording medium in which information such as data and programs is accumulated by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer. Examples of such a recording medium that can be removed from the computer include a flexible disk, a magneto-optical disk, an optical disk, a magnetic tape, and a memory card. In addition, examples of the recording medium fixed to the computer include a hard disk drive and a ROM.

  In said Embodiment 3, modification 3-1 and modification 3-2, although the example which implement | achieves the weekday / non-weekday estimation apparatus 110 with user terminals, such as a smart phone which a user carries, was not restricted to this, The weekday / non-weekday estimation device 110 may have another configuration as long as it includes the position information acquisition unit 11, the history storage unit 120, and the estimation unit 13. For example, a user carries a terminal having a position information acquisition unit 11 and a history storage unit 120 like a GPS logger, and a computer having an estimation unit 13 acquires history information from the terminal via a network and performs estimation. The structure to perform may be sufficient.

<Others>
The present invention includes a computer program that executes the above-described processing. Furthermore, a computer-readable recording medium on which the program is recorded also belongs to the category of the present invention. With respect to the recording medium on which the program is recorded, the above estimation process can be performed by causing the computer to read and execute the program on the recording medium.

  Here, the computer-readable recording medium refers to a recording medium in which information such as data and programs is accumulated by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer. Examples of such a recording medium that can be removed from the computer include a flexible disk, a magneto-optical disk, an optical disk, a magnetic tape, and a memory card. In addition, examples of the recording medium fixed to the computer include a hard disk drive and a ROM.

DESCRIPTION OF SYMBOLS 1 Weekday / non-weekday estimation apparatus 2 Network 11 Movement information acquisition part 12,120 History storage part 13,130,133 Estimation part

Claims (6)

A movement information acquisition unit for acquiring movement information of the moving body;
A history storage unit that associates the movement information of the mobile body with the date and time, and stores the movement information as movement history information;
Dividing one day into a plurality of time zones, obtaining the movement frequency of each time zone from the movement history information for each day of the week, and using the movement frequencies of the plurality of time zones on each day of the week as feature vectors, similarity of the feature vectors for each day of the week Each day of the week is classified into two groups according to the similarity of the feature vector of each day, and the day of the week classified into a group with a large number of elements is weekday, and the day of the week classified into a group with a small number of elements is not An estimator for estimating weekdays;
Weekday / non-weekday estimation device. The weekday / non-weekday estimation device according to claim 1, wherein the movement frequency is a frequency at which an arrival time of the moving body belongs to each time zone based on the movement history information . The history storage unit divides an area where the moving body moves into a plurality of areas, determines an area where the moving body is located based on position information of the moving body, and the movement accompanying the movement of the moving body Holding the movement history information representing the transition of the area where the body is located,
The estimation unit obtains a visit frequency for each area based on the movement history information, estimates the area having the highest visit frequency as an area to which a home belongs, and among the movement history information, an area to which the home belongs 3. The weekday / non-weekday estimation device according to claim 1, wherein the movement history information is extracted, and the weekdays or non-weekdays are estimated based on the extracted movement history information. Acquiring movement information of the moving object;
Associating movement information of the mobile object with date and time, and storing it as movement history information;
Dividing one day into a plurality of time zones, obtaining the movement frequency of each time zone from the movement history information for each day of the week, and using the movement frequencies of the plurality of time zones on each day of the week as feature vectors, similarity of the feature vectors for each day of the week Each day of the week is classified into two groups according to the similarity of the feature vector of each day, and the day of the week classified into a group with a large number of elements is weekday, and the day of the week classified into a group with a small number of elements is not Estimating a weekday;
A weekday / non-weekday estimation method in which a computer is executed. 5. The weekday / non-weekday estimation method according to claim 4, wherein, as the movement frequency, a frequency at which the arrival time of the moving body belongs to each of the time zones is obtained based on the movement history information . Acquiring movement information of the moving object;
Associating movement information of the mobile object with date and time, and storing it as movement history information;
Dividing one day into a plurality of time zones, obtaining the movement frequency of each time zone from the movement history information for each day of the week, and using the movement frequencies of the plurality of time zones on each day of the week as feature vectors, similarity of the feature vectors for each day of the week Each day of the week is classified into two groups according to the similarity of the feature vector of each day, and the day of the week classified into a group with a large number of elements is weekday, and the day of the week classified into a group with a small number of elements is not Estimating a weekday;
Weekday / non-weekday estimation program for causing a computer to execute.

Images