JP5547831B1 - Dwelling area estimation device, dwelling area estimation method, and dwelling area estimation program - Google Patents

Abstract

A user's living area is estimated using geomedia.
A geomedia acquisition unit (11) for estimating a user's residence area, which obtains geomedia for each user from a geomedia storage device (2) that accumulates geomedia including position information. Using the geomedia acquired by the geomedia acquisition unit 11, the location information totaling unit 13 that calculates the frequency score based on the frequency at which the user has been in the place for each place where the user has been in the past is calculated for each location. And a living area estimation unit 14 that estimates the user's living area using the frequency score.
[Selection] Figure 1

Description

  The present invention relates to a living area estimation apparatus, a living area estimation method, and a living area estimation program for estimating a user's living area from geomedia.

  Currently, geomedia is attracting attention. Geomedia is a generic term for media that handle location information, and is sometimes called geomedia, referring to services that use them. The position information is acquired as latitude and longitude from GPS or network base station information. Further, as a mechanism for obtaining map information and address information from the obtained latitude and longitude, for example, GoogleMapsAPI is disclosed.

  In geomedia, “check-in” is known as a service to disclose and share your current location. Examples of SNS services that provide check-in include “facebook” and “foursquare”.

  In addition, research has been conducted to estimate a user's living area using a blog, Twitter, or the like and use it for marketing (see Non-Patent Documents 1 and 2).

"Estimation of the blog author's residence", Yoshihito Yasuda, Tsutomu Hirao, Jun Suzuki, Hideki Amagasaki, The 12th Annual Conference of the Language Processing Society, pp. 512-515, 2006 "Estimation of Twitter user's residence by multi-class classification using regional feature word selection", Hayato Nishimura, Yoshihiko Nuwara, Seiji Amagasaki, IEICE, 4th Symposium, pp. 23-28, 2012

  In Non-Patent Documents 1 and 2, the user's living area is estimated by analyzing text information such as a blog or Twitter. However, since the geomedia does not always include text information, it is impossible to estimate the user's living area by applying Non-Patent Documents 1 and 2 to the geomedia. Further, even if text information is included in the geomedia, it cannot be used for estimation of a residential area if information about an address such as a place name is not included.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a living area estimation apparatus, a living area estimation method, and a living area estimation program for estimating a user's living area using geomedia. There is.

In order to achieve the above object, the present invention according to claim 1 is a living area estimation device for estimating a living area of a user, and a geomedia storage device for storing geomedia including position information is provided for each user. Using an acquisition means for acquiring media and a predetermined item included in the geomedia acquired by the acquisition means , it is estimated whether it is a place for a daily event or a place for an extraordinary event. , A score calculating means for calculating a score for each place, and a living area estimating means for estimating the user's living area using the score calculated for each place, and a predetermined item included in the geomedia The summary includes at least one of the presence / absence of a comment, the presence / absence of an image, and information on the person who was together .
The present invention according to claim 2 is a living area estimation apparatus, wherein the score calculation means calculates a score based on the frequency of being in the place for each place using the geomedia acquired by the acquisition means. However, the place of the extraordinary event is weighted lower than the place of the everyday event.

The gist of the present invention described in claim 3 is a living area estimation device, wherein the geomedia storage device includes a geomedia including check-in information checked in by a user from a terminal.

The present invention according to claim 4 is a residential area estimation device, further comprising filtering means for filtering geomedia of a predetermined exclusion target place where the exclusion target place is set, wherein the score calculation means includes filtering The gist is to calculate the score using a later geomedia.

The present invention according to claim 5 is a living area estimation device for estimating a living area of a user, and an acquisition means for acquiring geomedia for each user from a geomedia storage apparatus for storing geomedia including position information; Using the predetermined items included in the geomedia acquired by the acquisition means, it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place a score calculating unit that, using the geo-media and the score obtained by the obtaining unit, the center of gravity position where the user calculates the centroid of the locations which were in the past, and estimates the calculated the center of gravity as occupied zone of the user comprising a calculation unit, a principal that the predetermined item included in the geo-media, including the presence or absence of comments, the presence or absence of the image, at least one of information about the person was with To.

The present invention according to claim 6 is a living area estimation apparatus, wherein filtering is performed for filtering geomedia in a predetermined exclusion target list in which an exclusion target location is set or in which the score is smaller than a predetermined threshold. The center of gravity position calculating means further uses the geomedia after filtering to calculate the center of gravity of the place where the user has been in the past.

The present invention according to claim 7 is a living area estimation method performed by a computer system, wherein the computer system obtains geomedia for each user from a geomedia storage device that stores geomedia including location information. Step and a predetermined item included in the geomedia acquired in the acquisition step to estimate whether it is a place of a daily event or a place of an extraordinary event, and score for each place A score calculating step for calculating the living area and a living area estimating step for estimating the living area of the user using the score calculated for each place, and a predetermined item included in the geomedia includes a comment The gist is to include at least one of presence / absence, presence / absence of an image, and information on a person who was together .

The present invention according to claim 8 is a living area estimation method performed by a computer system, wherein the computer system obtains geomedia for each user from a geomedia storage device that stores geomedia including location information. Step and a predetermined item included in the geomedia acquired in the acquisition step to estimate whether it is a place of a daily event or a place of an extraordinary event, and score for each place Calculating the center of gravity of the place where the user has been in the past using the geomedia acquired in the acquiring step and the score, and estimating the calculated center of gravity as the user's living area performs a centroid position calculation step, and the predetermined item included in the geo-media, presence or absence of comments, the presence or absence of the image, together And summarized in that at least one of information about people who were.

The present invention according to claim 9 is a residential area estimation program executed by a computer system, and acquires geomedia for each user from a geomedia storage device that stores geomedia including location information in the computer system. Using the acquisition step and a predetermined item included in the geomedia acquired in the acquisition step, it is estimated whether it is a place for a daily event or a place for an extraordinary event. A score calculating step for calculating a score and a living area estimating step for estimating the living area of the user using the score calculated for each place are executed, and the predetermined item included in the geomedia includes: The gist is to include at least one of the presence / absence of a comment, the presence / absence of an image, and information about the person who was together .

The present invention according to claim 10 is a residential area estimation program executed by a computer system, and acquires geomedia for each user from a geomedia storage device that stores geomedia including location information in the computer system. Using the acquisition step and a predetermined item included in the geomedia acquired in the acquisition step, it is estimated whether it is a place for a daily event or a place for an extraordinary event. Using the score calculating step for calculating the score, the geomedia acquired in the acquiring step, and the score, the center of gravity of the place where the user has been in the past is calculated, and the calculated center of gravity is estimated as the user's living area and the center of gravity position calculation step of, by the execution, the predetermined item included in the geo-media, whether comments, The presence or absence of the image, and summarized in that at least one of information about people who were together.

  ADVANTAGE OF THE INVENTION According to this invention, the objective of this invention can provide the living area estimation apparatus which estimates a user's living area using a geomedia, the living area estimation method, and a living area estimation program.

It is a block diagram which shows the structure of the living area estimation apparatus which concerns on the 1st Embodiment of this invention. It is a figure which shows the data structural example of a geomedia storage apparatus. It is a block diagram which shows the structure of the living area estimation apparatus which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the living area estimation apparatus which concerns on the modification of the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the living area estimation apparatus which concerns on the 3rd Embodiment of this invention. It is a block diagram which shows the structure of the living area estimation apparatus which concerns on the 4th Embodiment of this invention. It is explanatory drawing for demonstrating the modification of the 4th Embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described.

<First Embodiment>
FIG. 1 is a diagram showing a configuration of a living area estimation apparatus 1 according to the first embodiment of the present invention.

  The illustrated living area estimation apparatus 1 includes a geomedia acquisition unit 11, a geomedia storage unit 12, a position information totaling unit 13, a living area estimation unit 14, and an estimation result output unit 15.

  The geomedia acquisition unit 11 acquires the geomedia for each user from the geomedia storage device 2 that stores the geomedia including the position information, and stores it in the geomedia storage unit 12.

  In the geomedia storage device 2, geomedia generated based on the position information transmitted from each user terminal 3 is stored for each user. The information transmitted from the user terminal 3 may be check-in information that the user has checked in using the user terminal 3, or position information transmitted by the user terminal 3. As the position information transmitted by the user terminal 3, for example, latitude and longitude acquired from GPS, base station information, or the like can be considered. In addition, the information stored in the geomedia storage device 2 includes a case where check-in information and position information transmitted by the user terminal 3 are mixed, or only the check-in information is transmitted, or the user terminal 3 transmits the information. The case where only position information to be performed is considered.

  Note that the geomedia of the geomedia storage device 2 is not limited to the above information. For example, geomedia generated based on the location information of each user terminal 3 transmitted by a base station or the like may be used.

  Check-in is an action (operation) for informing a friend / acquaintance of his / her current location on SNS (Social Networking Service). The user inputs (checks in) the location where the user is located from the user terminal 3. Also, when checking in, the user can also send a comment, an image such as a photo / moving picture, the name of the person with whom the user is together, and share / publish it with the current location. In this case, the geomedia storage device 2 is provided in a system that provides an SNS service.

  The position information totalization unit 13 (frequency score calculation means) uses the geomedia acquired by the geomedia acquisition unit 11 and based on the frequency at which the user has been in the place for each place (checked-in place) in the past. To calculate the frequency score. Further, the location information totaling unit 13 estimates whether the location is a daily event location or an extraordinary event location by using a predetermined item included in the geomedia. The event location is weighted lower than the daily event location.

  The residence area estimation unit 14 estimates the residence area of the user using the frequency score calculated for each place. The estimation result output unit 15 outputs the location estimated by the residential area estimation unit 14 to the output device such as a display or a printer as the residential area of the user, or stores it in a storage unit (not shown).

  FIG. 2 shows an example of geomedia stored in the geomedia storage device 2 (or geomedia storage unit 12) of the present embodiment. The illustrated geomedia has predetermined items such as a user ID, date / time, latitude, longitude, address, place (spot), comment (text information), name of a person who is together, and an image (for example, a photo, a video). .

  The user ID, date / time, latitude, and longitude are information transmitted from the user terminal 3. For the address and location, the SNS system that stores the geomedia in the geomedia storage device 2 sets the corresponding address and location using the latitude and longitude transmitted from the user terminal 3, or the user terminal 3 checks It is conceivable to send an address and location as in-information. In this embodiment, latitude and longitude are used as position information transmitted from the terminal, but information other than latitude and longitude may be used.

  In the comment, a comment (text information) input by the user when checking in is set. In addition, the name of the friend / acquaintance that the user has input at the time of check-in is set as the name. The image is set to an image such as a photo / moving picture taken by the user at the time of check-in. The comment, the name of the person who is together, and the image are transmitted from the user terminal 3 as check-in information.

  The living area estimation device 1 described above can use, for example, a general-purpose computer system including a CPU, a memory, an external storage device such as an HDD, an input device, and an output device. In this computer system, each function of the living area estimation device 1 is realized by the CPU executing a program for the living area estimation device 1 loaded on the memory. Further, the program for the residential area estimation device 1 can be stored in a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, a DVD-ROM, or can be distributed via a network. In addition, the residential area estimation apparatus 1 of this embodiment may be comprised with one server, or may be comprised with the some server.

  Next, the processing of this embodiment will be described.

  In the present embodiment, the geomedia stored in the geomedia storage device 2 includes extraordinary location information (for example, location information different from an actual living area such as a travel destination). Information is removed as much as possible so that the user's living area can be estimated. In particular, when using the check-in information, the “check-in” action is basically an extraordinary action, and therefore the extraordinary location information is removed so that the user's living area can be estimated.

  As shown in FIG. 1, first, the geomedia acquisition unit 11 acquires geomedia for each user from the geomedia storage device 2 and stores it in the geomedia storage unit 12.

  Then, the location information totaling unit 13 reads a user's geomedia record (data) from the geomedia storage unit 12 and calculates a frequency score for each location where the user has checked in or was in the past. To do.

  Specifically, the position information totaling unit 13 uses the address in the geomedia to count the frequency (number of times) of visiting each place (or region), and calculates a frequency score for each place. The location may be, for example, a local unit such as Kanto or Kansai, a prefecture unit, or a municipality or a ward unit. The position information totalization unit 13 determines the location to which the address of each geomedia record belongs using the address. Note that the coarser the place (the larger the unit), the more generally the accuracy of the living area estimation is improved.

  In this embodiment, when calculating the frequency score, not only the frequency is simply calculated, but also the time t, day of week d, category c, image presence / absence p, comment presence / absence u, the name of the person who was together It may be excluded from the weighting or frequency score calculation using the presence / absence of h. All of these pieces of information may be used, or at least one of them may be used.

The frequency score S of a place P where all information is used is calculated by Equation 1.

X _p is the number of times the location P appears in the user's geomedia (i = 1, 2,..., X _p ). When the weight W is “0”, it is equivalent to excluding from the calculation target. Further, when the weight W is “1”, this is equivalent to calculating the score only by the frequency (number of times) the place P has appeared.

  Each item will be described below.

(1) Time t
Position information compiling unit 13 acquires the time of the date and time set in the record of the geo media of the location P, and sets the weight W _t a predetermined depending on the time zone to which the time belongs. That is, using the time when the user terminal 3 transmits the check-in information or the position information, for example, it is estimated whether the user terminal 3 is around the home or the destination. For example, if time is of the time zone of morning or evening, it estimated that there is a high possibility of your neighborhood, to set the W _t to a large predetermined value than "1". On the other hand, if the time is daytime, estimates that are likely to be going out destination, sets the W _t "1" is smaller than a predetermined value.

If the time is in the morning or night time, set W _t to “1” and the corresponding geomedia record will be the target of calculation. If the time is in the day time, set W _t to “0”. The geomedia record may be excluded from the calculation target.

(2) Day of the week d
Position information compiling unit 13 obtains the day of the week of the date with the date of the location set in the geo media records P date, sets the weight W _d a predetermined depending on the day of the week. That is, using the day of the week when the user terminal 3 transmits the check-in information or the position information, it is estimated whether the user terminal 3 is near the home or the destination. For example, if the day of the week is a weekday (Monday to Friday), it is estimated that there is a high possibility of being near the home, and W _d is set to a predetermined value greater than “1” (or “1” is set to W _d) , To be calculated). On the other hand, if the date is a weekend / holiday, it is estimated that there is a high possibility of going out, and W _d is set to a predetermined value smaller than “1” (or W _d is set to “0” and calculated) Excluded).

(3) Category c
The position information totalization unit 13 acquires the category of the address (or place) set in the geomedia record of the place P. As the category, for example, it is conceivable to use a “tourist spot category” and a “non-tourist spot category”. In this embodiment, a sightseeing spot list in which addresses and place names are set for each sightseeing spot is created in advance and stored in a storage unit (not shown). If the address set in the geomedia record exists in the sightseeing spot list, it is determined that the address in the geomedia record is a sightseeing spot category, and if the address does not exist in the sightseeing spot list, It is determined that the address of the media record is a non-tourist destination category.

If it is determined that the non Landmark category estimates that there is a high possibility of your neighborhood, setting a predetermined value of a large predetermined non Landmark categories from "1" to W _c (or, in W _c Set “1” to calculate.) On the other hand, if it is determined that the tourist destination category, estimates that are likely to be going out destination, and sets "1" value of less than a predetermined Landmark categories to W _c (or, in W _c " “0” is set and excluded from calculation).

(4) Presence / absence of image p
The position information totalization unit 13 sets a predetermined weight W _p according to the presence or absence of an image of the geomedia record at the location P. That is, when the check-in information is transmitted, it is estimated whether the event is an extraordinary event or an ordinary event depending on whether an image such as a photo or a moving image is uploaded. If the image is not been attached estimates that there is a high possibility of home around everyday events, sets the W _p to a large predetermined value than "1" (or, "1" is set to the W _p And the calculation target). On the other hand, if an image is attached, it is estimated that there is a high possibility of going out in an extraordinary event, and W _p is set to a predetermined value smaller than “1” (or “0” is set in W _p) Is excluded from calculation).

If an image is attached in addition to the presence / absence of an image, the image itself is subjected to image analysis (image recognition), the image is classified into scenes (scenes), and a predetermined weight W _{p for} each scene. It is good also as setting. For example, in the case of a meal scene, an image is attached, but it is estimated that it is an everyday event, and W _p is set to a predetermined value larger than “1” (or “1” is set to W _p) Set and calculate.) On the other hand, if the landscape scene, estimated to be gone destination extraordinary event, it sets the W _p to "1" is smaller than a predetermined value (or, "0" is set to W _p, calculates Excluded).

  When an image is attached, the position information embedded in the image may be used instead of the position information (latitude / longitude) transmitted from the user terminal 3.

(5) Existence of comments u
The position information totalization unit 13 sets a predetermined weight W _u according to the presence or absence of a comment (text information) of the geomedia record at the location P. That is, when the check-in information is transmitted, whether the event is an extraordinary event or an ordinary event is estimated depending on whether or not a comment is uploaded together. If there is no comment, it is estimated that there is a high possibility of being near the home in a daily event, and W _u is set to a predetermined value larger than “1” (or “1” is set to W _u and the calculation target And). On the other hand, if the comment is present, it estimates that there is likely to be gone destination extraordinary event, sets the W _u "1" is smaller than a predetermined value (or, "0" is set to W _u , Excluded from calculation).

(6) Presence / absence of name
The position information totalization unit 13 sets a predetermined weight W _h according to the presence or absence of the name of the geomedia record at the place P. In other words, when you send in the check-in information, depending on whether you are uploading the name of the person who is with you or not, depending on whether you were alone or with several people, Or a daily event. For example, if you're checking in alone (if you don't have a name), assume that you are likely to be near your home for everyday events and set W _h to a predetermined value greater than "1" (or Set W _h to “1” for calculation). On the other hand, if you check in with multiple people (if you have a name), we assume that you are likely to be a destination for an extraordinary event, and set W _h to a predetermined value less than “1” ( Or, set “0” to W _h and exclude it from calculation).

  In addition, when the geomedia record of the geomedia storage device 2 is not check-in information but is location information transmitted by the user terminal 3, the frequency score is calculated using (1) time, (2) day of the week, and (3) category. Can be calculated.

  Moreover, the information of (1) to (6) described above is an example, and the frequency score may be calculated using information that can be collected by other user terminals 3. For example, the frequency score may be calculated using ambient noise, a user's heart rate, a moving speed, and the like. Specifically, for noise and heart rate, if it is large, it is estimated to be an extraordinary event, and a weight W smaller than “1” is set. Set a larger weight W. Also, if the moving speed is high, it is determined that the vehicle is on the vehicle, and it is estimated that the event is an extraordinary event, and a weight W smaller than “1” is set. Estimating that the event is an everyday event and setting a weight W greater than “1” can be considered. These pieces of information are collected at the user terminal 3 and transmitted to the geomedia storage device 2 together with the position information.

  Thus, the position information totalization unit 13 calculates a frequency score for every user for every place P that the user has checked in or has been in the past.

  The residence area estimation unit 14 estimates the residence area of each user using the frequency score of each place P calculated by the position information aggregation unit 13. Then, the estimation result output unit 15 outputs the living area estimated by the living area estimation unit 14 to a display or a printer.

  The residential area estimation unit 14 may output only one place having the highest frequency score as an estimation result (1-best), or may output a plurality of places having a predetermined top N frequency scores. It is good also as (N-best). Alternatively, it is also conceivable to set a threshold T1 in advance and output each location of a frequency score equal to or higher than the threshold T1 as an estimation result.

  When outputting a plurality of places as an estimation result, it is effective when there are a plurality of user residence areas. Moreover, when outputting an estimation result, you may output not only an estimated residence area but a corresponding frequency score. By outputting the frequency score, for example, when recommending recommended store information according to the estimated residential area in the subsequent processing using this estimation result, the number of recommended stores is determined according to the frequency score. It is possible to control.

  In addition to outputting the location P itself having a high frequency score as the estimated residential area, it is also possible to identify an area to which a plurality of higher ranks belong and have the frequency score output as an estimation result. . For example, when the location of the higher frequency score is Tokyo, Saitama, or Chiba, the estimated residential area may be the Kanto area.

  Even when the location P itself having a high frequency score is output in a pinpoint manner, the area to which the location P belongs may be output together as an estimation result.

  In the first embodiment described above, using geomedia, a frequency score is calculated based on the frequency at which the user has been in the place for each place where the user has been in the past, and the frequency score calculated for each place is used. To estimate the user's residence area. Thereby, in this embodiment, a user's residence area can be estimated using the positional information contained in geomedia, without analyzing text information.

  Further, in the present embodiment, by using a predetermined item included in the geomedia, it is estimated whether it is a place of a daily event or a place of an extraordinary event. By assigning a lower weight to the place than the place of the daily event, the influence of extraordinary geomedia can be reduced, and the user's living area can be estimated with higher accuracy.

  Moreover, by estimating a user's residence area from geomedia in this way, a user's profiling can be performed easily and it can utilize for marketing. For example, it is possible to appropriately recommend a store in an area corresponding to the user's residence area.

<Second Embodiment>
FIG. 3 is a diagram showing a configuration of a residential area estimation device 1A according to the second embodiment of the present invention.

  In the first embodiment, when a large amount of check-in information is included in the geomedia stored in the geomedia storage device 2 and the user does not check in in the residential area, the user's residential area cannot be estimated correctly. There is a case. Therefore, in the second embodiment, instead of estimating the user's living area directly from the check-in information or the transmitted position information, the location of the check-in or the distribution of the transmitted position information is calculated. Use to estimate the user's residential area.

  The living area estimation device 1A shown in FIG. 3 includes a geomedia acquisition unit 11, a geomedia storage unit 12, a gravity center position calculation unit 16, and an estimation result output unit 15. The geomedia acquisition unit 11, the geomedia storage unit 12, and the estimation result output unit 15 of the present embodiment are the same as the geomedia acquisition unit 11, the geomedia storage unit 12, and the estimation result output unit 15 of the first embodiment. Therefore, the description is omitted here.

  The center-of-gravity position calculation unit 16 uses the geomedia position information (such as latitude and longitude) acquired by the geomedia acquisition unit 11 to calculate the center of gravity of the place where the user has checked in or the place where the user has been in the past. And the calculated center of gravity is estimated as the user's living area. That is, the user's residence area is estimated based on the user's action area.

  Below, the process of this embodiment is demonstrated.

  The geomedia acquisition unit 11 acquires geomedia for each user from the geomedia storage device 2 and stores it in the geomedia storage unit 12.

The center-of-gravity position calculation unit 16 reads the user's geomedia record from the geomedia storage unit 12 and calculates the center of gravity of the location (position) of each read record. More specifically, the location of geo-media _{P 1, P 2, ··,} P k, and ·· P _K. K is the number of types of locations where the user has checked in or where the user has been in the past. X _pk is the frequency (number of times) of checking in at the location P _k . Further, the latitude of the place P _k is defined as Lat (P _k ), and the longitude is defined as Lon (P _k ).

In this case, the center-of-gravity position calculation unit 16 calculates the latitude E _lat and the longitude E _lon of the center of gravity of the estimated living area using the following formula 2 using the latitude and longitude of each record of the _geomedia .

Further, the center-of-gravity position calculation unit 16 may calculate the center of gravity using the frequency score S used in the first embodiment instead of calculating the center of gravity using the frequency X of Equation 2. When the frequency score S is used, the latitude E _lat and the longitude E _lon of the center of gravity of the estimated living area are as shown in the following Expression 3. Note that S _pk is a frequency score of the place P _k .

  The living area estimation apparatus 1B in this case further includes a position information totaling unit 13 as shown in FIG. And the positional information total part 13 calculates the frequency score of the place which the said latitude and longitude show for every latitude and longitude. Specifically, the position information totalization unit 13 uses a predetermined item included in the geomedia to determine whether the place is a place visited by a daily event or a place visited by an extraordinary event. For the places visited by the extraordinary event, the weighting is lower than the places visited by the daily event, and the frequency score is calculated based on the frequency and weighting for each place.

  And the estimation result output part 15 outputs the gravity center which the gravity center position calculation part 16 calculated to a display or a printer etc. as an estimated residence area of the said user. The estimation result output unit 15 may output the calculated latitude and longitude of the center of gravity as an estimated residence area, or obtains a corresponding address from the latitude and longitude and outputs the address as the estimated residence area. May be.

  In the second embodiment described above, using geomedia, the center of gravity of the place where the user has been in the past is calculated based on the frequency of being in the place, and the calculated center of gravity is used as the living area of the user. presume. Thereby, in this embodiment, a user's residence area can be estimated using the positional information contained in geomedia, without analyzing text information. That is, in this embodiment, even when the user is not checking in in the residential area, the user's residential area is estimated using the distribution of the locations where the user has checked in or where the user has been in the past. Therefore, the user's living area can be estimated correctly.

  Further, in the present embodiment, by using a predetermined item included in the geomedia, it is estimated whether it is a place of a daily event or a place of an extraordinary event. For locations, when calculating the center of gravity using a frequency score that is weighted lower than the location of everyday events, reduce the impact of extraordinary geomedia and estimate the user's living area with higher accuracy Can do.

<Third Embodiment>
In the second embodiment, the center of gravity is calculated using all the geomedia records stored in the geomedia storage device 2 (or the geomedia storage unit 12). Therefore, for example, when checking in at a travel destination or the like, it is affected by an irregular place, which causes a decrease in the estimation accuracy of the residential area. Therefore, in the third embodiment, an irregular place such as a travel destination is removed in advance to estimate the living area. In the present embodiment, the geomedia records stored in the geomedia storage device 2 are filtered using an exclusion target position information list created in advance.

  FIG. 5 is a configuration diagram showing a configuration of a residential area estimation device 1C according to the third embodiment of the present invention. The illustrated living area estimation device 1C includes a geomedia acquisition unit 11, a geomedia storage unit 12, a location information exclusion unit 17 (filtering means), an exclusion target location information list 17A, a centroid location calculation unit 16, and an estimation. And a result output unit 15. The geomedia acquisition unit 11, the geomedia storage unit 12, and the estimation result output unit 15 of the present embodiment are the same as the geomedia acquisition unit 11, the geomedia storage unit 12, and the estimation result output unit 15 of the first embodiment. Therefore, the description is omitted here.

  The exclusion target position information list 17A is a list in which position information (such as place name, address, latitude, and longitude) of exclusion objects (for example, sightseeing spots) that are not used for living area estimation is set. The exclusion target position information list 17A may be created manually, or the place name and position information appearing on the tourist information Web page on the Web is automatically collected to generate the exclusion target position information list 17A. Also good. Further, the exclusion target position information list 17A may be created using a technique described in Non-Patent Document 1 or Non-Patent Document 2.

  The position information exclusion unit 17 reads a user's geomedia record from the geomedia storage unit 12, and filters (excludes) records that match any of the exclusion target position information set in the exclusion target position information list 17A. Discard) and output to the gravity center position calculation unit 16. The center-of-gravity position calculation unit 16 uses the geomedia record after filtering, as in the second embodiment, the frequency at which the center of gravity of the place where the user has been in the past (the place where the user has checked in) is located in the place. And the calculated center of gravity is estimated as the user's living area.

  Below, the process of this embodiment is demonstrated.

  The geomedia acquisition unit 11 acquires geomedia for each user from the geomedia storage device 2 and stores it in the geomedia storage unit 12. The position information exclusion unit 17 reads a user's geomedia record from the geomedia storage unit 12, refers to the exclusion target position information list 17A, and matches any of the exclusion target position information set in the list 17A. The record to be filtered is filtered, and the filtered geomedia record is output to the gravity center position calculation unit 16. That is, when any of the latitude and longitude, the address, and the location of each geomedia record exists in the exclusion target position information list 17A, the position information exclusion unit 17 excludes and discards the record.

  Accordingly, only the filtered geomedia record (record not corresponding to the exclusion target position) is input to the gravity center position calculation unit 16. The center-of-gravity position calculation unit 16 uses the geomedia record after filtering, as in the second embodiment, to set the center of gravity of the place where the user has checked in or the place where the user has been in the past as the place. It calculates based on the frequency of staying or the frequency score, and the calculated center of gravity is estimated as the user's living area.

  Then, the estimation result output unit 15 outputs the centroid (latitude and longitude, or address) calculated by the centroid position calculation unit 16 to a display or a printer.

  In the third embodiment described above, geomedia in an irregular place such as a travel destination is filtered in advance, a center of gravity is calculated using the geomedia after filtering, and the center of gravity is estimated as a living area. Thereby, in this embodiment, the estimation precision of a residential area can be improved more.

<Fourth Embodiment>
In the fourth embodiment, similarly to the third embodiment, irregular places such as travel destinations are removed in advance to estimate the living area. In the fourth embodiment, geomedia records are filtered without creating an exclusion target position information list in advance. That is, in this embodiment, filtering is performed without requiring the cost of creating the exclusion target position information list.

  FIG. 6 is a diagram showing a configuration of a residential area estimation device 1D according to the fourth embodiment of the present invention.

  The illustrated living area estimation device 1D includes a geomedia acquisition unit 11, a geomedia storage unit 12, a position information totaling unit 13, a filtering unit 18, a gravity center position calculation unit 16, and an estimation result output unit 15. Regarding the geomedia acquisition unit 11, the geomedia storage unit 12, the position information totalization unit 13, and the estimation result output unit 15 of the present embodiment, the geomedia acquisition unit 11, the geomedia storage unit 12, and the position information of the first embodiment Since it is the same as that of the total part 13 and the estimation result output part 15, description is abbreviate | omitted here. The filtering unit 18 performs geomedia filtering using the frequency score described in the first embodiment. The centroid position calculation unit 16 calculates the centroid of the place where the user checked in or the place where the user was in the past, using the filtered geomedia record, as in the second embodiment. The estimated center of gravity is estimated as the living area of the user.

  Below, the process of this embodiment is demonstrated.

  The geomedia acquisition unit 11 acquires geomedia for each user from the geomedia storage device 2 and stores it in the geomedia storage unit 12. The position information totalization unit 13 reads a user geomedia record from the geomedia storage unit 12 and calculates a frequency score for each location. The calculation of the frequency score is the same as in the first embodiment.

  The filtering unit 18 performs geomedia filtering using the frequency score. As a filtering method, there are two methods for filtering a place having a low frequency score or a place that is clearly separated from other check-in places.

  When filtering a place where the frequency score is small, the filtering unit 18 filters only (excludes or discards) a geomedia record in a place less than the threshold T2 with only a place having a predetermined threshold T2 or more as a calculation target. Alternatively, the filtering unit 18 may automatically detect a place where the frequency score is clearly small by using a statistical test method. As a method for detecting an obviously small frequency score (outlier), the Smirnov-Grubbs test is known.

  In addition, when filtering the locations where the check-in location or the location that was in the past is filtered, the frequency score for each location that has been checked-in or in the past is used, for example, the frequency for each prefecture Calculate the total score, and filter the prefectures other than the prefecture with the highest total frequency score. In addition to prefectural units, city, ward units, and local units such as Kanto and Kansai may be used. Note that the coarser unit generally improves the estimation accuracy.

  Thereby, only the filtered geomedia record is input to the gravity center position calculation unit 16. The center-of-gravity position calculation unit 16 uses the filtered geomedia record to locate the center of gravity of the place where the user checked in or the place in the past in the place, as in the second embodiment. Calculation is performed based on the frequency or the frequency score, and the calculated center of gravity is estimated as the living area of the user.

  Then, the estimation result output unit 15 outputs the centroid (latitude and longitude, or address) calculated by the centroid position calculation unit 16 to a display or a printer.

  As a modification of the fourth embodiment, first, the center-of-gravity position calculation unit 16 calculates the first center of gravity using all geomedia records as in the second embodiment. Then, the filtering unit 18 calculates the distance of each place checked in from the calculated center of gravity, and filters the place where the distance from the center of gravity is the longest or the distance from the center of gravity is equal to or greater than a predetermined threshold T3. The subsequent geomedia record is output to the gravity center position calculation unit 16. And the gravity center position calculation part 16 is good also as calculating the 2nd gravity center using the geomedia record after filtering, and presuming the calculated 2nd gravity center as a residence area of the said user.

  FIG. 7 is an explanatory diagram for explaining this modification. From the first calculated first center of gravity 61, the distance from each checked-in place (indicated by “◯” in FIG. 7) is calculated, and a place 62 farthest from the center of gravity or a predetermined threshold T3 from the center of gravity is calculated. The geomedia records of the locations 62 and 63 separated as described above are filtered, and the second centroid 64 is calculated again using the filtered geomedia records, and the second centroid 64 is estimated as the living area. .

  In the fourth embodiment described above, geomedia at an irregular place such as a travel destination is filtered in advance, the center of gravity is calculated using the geomedia after filtering, and the center of gravity is estimated as a living area. Thereby, in this embodiment, the estimation precision of a residential area can be improved more. Moreover, in this embodiment, since the geomedia record is filtered without creating the exclusion target position information list in advance, the cost of creating the exclusion target position information list can be reduced.

  In addition, this invention is not limited to the said embodiment, Many deformation | transformation are possible within the range of the summary. For example, it is good also as a living area estimation apparatus which combined 1st Embodiment and 3rd Embodiment. That is, the living area estimation apparatus (see FIG. 1) of the first embodiment has an exclusion target position information list 17A and a position where the places to be excluded of the third embodiment living area estimation apparatus (see FIG. 5) are set. An information exclusion unit 17 is provided, the location information exclusion unit 17 filters geomedia records in the location of the exclusion target location information list 17A, and the location information aggregation unit 13 uses the filtered geomedia records to calculate a frequency score. May be calculated. Moreover, it is good also as a living area estimation apparatus which combined said each embodiment besides this.

DESCRIPTION OF SYMBOLS 1, 1A, 1B, 1C, 1D: Living area estimation apparatus 11: Geomedia acquisition part 12: Geomedia storage part 13: Location information totaling part 14: Living area estimation part 15: Estimation result output part 16: Center of gravity position calculation part 17: Outside location information office 17A: Exclusion target location information list 18: Filtering unit 2: Geomedia storage device 3: User terminal

Claims (10)

A living area estimation device for estimating a user's living area,
An acquisition unit that acquires geomedia for each user from a geomedia storage device that stores geomedia including position information;
Using predetermined items included in the geomedia acquired by the acquisition means , it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place. A score calculation means;
Using the score calculated for each place, and a residence area estimation means for estimating the residence area of the user ,
The predetermined area included in the geomedia includes at least one of the presence / absence of a comment, the presence / absence of an image, and information about a person who was together, and the living area estimation apparatus.   The living area estimation apparatus according to claim 1,
  The score calculation means calculates a score based on the frequency of being at the place for each place using the geomedia acquired by the acquisition means, and for the place of the extraordinary event, Lower weight than event location
  A living area estimation device characterized by the above. The living area estimation apparatus according to claim 1 or 2 ,
In the geomedia storage device, there is a geomedia including check-in information checked in by a user from a terminal. It is the living area estimation apparatus of any one of Claim 1 to 3 ,
Further comprising a filtering means for filtering geomedia of a predetermined exclusion list location where the exclusion location is set,
It said score calculation means uses a geo medium after filtering occupied zone estimation apparatus and calculates the score. A living area estimation device for estimating a user's living area,
An acquisition unit that acquires geomedia for each user from a geomedia storage device that stores geomedia including position information;
Using predetermined items included in the geomedia acquired by the acquisition means, it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place. A score calculation means;
Using the geomedia acquired by the acquisition means and the score, the center of gravity of the place where the user has been in the past is calculated, and the center of gravity position calculating means for estimating the calculated center of gravity as the living area of the user, ,
The predetermined area included in the geomedia includes at least one of the presence / absence of a comment, the presence / absence of an image, and information about a person who was together, and the living area estimation apparatus. The living area estimation apparatus according to claim 5 ,
Filtering means for filtering the location of a predetermined exclusion target list in which the location of the exclusion target is set, or the geomedia where the score is smaller than a predetermined threshold,
The center-of-gravity position calculation means calculates the center of gravity of the place where the user has been in the past using the filtered geomedia. A method for estimating a living area performed by a computer system,
The computer system includes:
An acquisition step of acquiring geomedia for each user from a geomedia storage device that stores geomedia including position information;
Using a predetermined item included in the geomedia acquired in the acquisition step, it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place. A score calculating step;
Using the score calculated for each place, and performing a residence area estimation step for estimating the residence area of the user ,
The predetermined area included in the geomedia includes at least one of the presence / absence of a comment, the presence / absence of an image, and information about a person who was present together . A method for estimating a living area performed by a computer system,
The computer system includes:
An acquisition step of acquiring geomedia for each user from a geomedia storage device that stores geomedia including position information;
Using a predetermined item included in the geomedia acquired in the acquisition step, it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place. A score calculating step;
Performing a center-of-gravity position calculating step of calculating the center of gravity of the place where the user has been in the past using the geomedia acquired in the acquiring step and the score, and estimating the calculated center of gravity as the living area of the user; ,
The predetermined area included in the geomedia includes at least one of the presence / absence of a comment, the presence / absence of an image, and information about a person who was present together . A residential area estimation program executed by a computer system,
In the computer system,
An acquisition step of acquiring geomedia for each user from a geomedia storage device that stores geomedia including position information;
Using a predetermined item included in the geomedia acquired in the acquisition step, it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place. A score calculating step;
Using the score calculated for each place, and performing a residential area estimation step of estimating the residential area of the user ,
The predetermined item included in the geomedia includes at least one of presence / absence of a comment, presence / absence of an image, and information about a person who was together.
A residential area estimation program characterized by A residential area estimation program executed by a computer system,
In the computer system,
An acquisition step of acquiring geomedia for each user from a geomedia storage device that stores geomedia including position information;
Using a predetermined item included in the geomedia acquired in the acquisition step, it is estimated whether it is a place of a daily event or a place of an extraordinary event, and a score is calculated for each place. A score calculating step;
Using the geomedia acquired in the acquisition step and the score, the center of gravity of the place where the user has been in the past is calculated, and the center of gravity position calculating step of estimating the calculated center of gravity as the living area of the user is executed. Let
The predetermined item included in the geomedia includes at least one of presence / absence of a comment, presence / absence of an image, and information about a person who was together.
A residential area estimation program characterized by

Images