JP5525835B2 - POI recommendation server and POI recommendation system - Google Patents

Description

The present invention relates to a PO recommendation server and a POI recommendation system.

  Currently, a navigation device that searches a point that matches a user's preference as a via point is known (see, for example, Patent Document 1). Moreover, while detecting user's action history information, the action history information analysis which discriminate | determines whether a user's staying place is a business relevant point or a private relevant point based on a time zone, an illumination intensity level, and a noise level. An apparatus is known (see, for example, Patent Document 2).

JP 2009-139244 A JP 2009-43057 A

  However, the technique of Patent Document 1 has a problem that the user needs to input preference information in advance in order to search for a waypoint reflecting the user's preference. In addition, the technique disclosed in Patent Document 1 has a problem that it is not possible to search for a waypoint in consideration of the user's space-time constraints. The technique of Patent Document 2 has a problem that sensing such as an illuminance level is necessary for discrimination. Further, the technique of Patent Document 2 has a problem that public POI cannot be recommended.

  The present invention has been made in view of such circumstances, and provides a technique for determining whether the current location is a private POI or a public POI from only the user's action history information, and from the current location to the destination. It is an object to provide a technique for recommending a public POI that stops at the time of every movement in consideration of a user's space-time constraints.

In order to solve the above problem, a POI recommendation system according to an aspect of the present invention includes an information collection unit that collects behavior information of one user including a current position and a current time, and a plurality of information collected by the information collection unit. The POI extraction unit that extracts the POI related to the one user based on the behavior information of the user, and the POI extraction unit extracts the POI extraction unit based on the plurality of POI information related to the single user extracted by the POI extraction unit A plurality of POIs related to a single user are private POIs related to the one user whose stay time or frequency is high for the one user, and POIs whose stay time or frequency is low for the one user. A POI classifying unit that classifies a public POI related to a single user, and each user classified by the POI classifying unit. The public POI storage unit that stores the POI information of the public POI, the spatiotemporal constraint setting unit that sets the user's spatiotemporal constraint, and the spatiotemporal constraint related to one user set by the spatiotemporal constraint setting unit. The recommended candidate public POI to be a candidate of the recommended public POI recommended as the public POI to which the one user stops when moving from the current location to the destination from among the public POIs related to a plurality of users. And a recommended candidate POI extraction unit for extracting.

  In the POI recommendation system, the recommended candidate POI extraction unit is configured to determine whether a stay time or frequency of the one user from among a plurality of POIs related to one user extracted by the POI extraction unit by the POI classification unit. When the top two POIs are classified as the private POI related to the one user and the other POIs are classified as the public POI related to the one user, the current position of the one user is the one user If the private POI is one of the private POIs, the recommended private public POI up to the other private POI is extracted with the other private POI different from the one private POI as the destination. May be.

  In the POI recommendation system, the recommendation candidate POI extraction unit further includes a destination determination unit that determines a destination when the recommendation candidate public POI is extracted, and the POI classification unit is extracted by the POI extraction unit Among the plurality of POIs related to one user, the top two or more POIs with the most frequent staying frequency or frequency of the one user are classified as the private POI related to the one user, and other POIs are classified into the one user. And the destination determination unit determines whether the POI classification unit is a top three or more POIs having a high staying time or frequency of the one user among a plurality of POIs related to the one user. Are classified as the private POI relating to the one user, and other POIs are classified as the public POI relating to the one user. When the current position of the one user is the one private POI related to the one user, the one user is based on the current position and the current time collected by the information collecting unit. You may make it determine the said private POI used as the destination out of several other said private POIs different from the said private POI.

  In the POI recommendation system, the recommendation candidate POI extraction unit further includes a destination determination unit that determines a destination when the recommendation candidate public POI is extracted, and the POI classification unit is extracted by the POI extraction unit Among the plurality of POIs related to one user, the top two or more POIs with the most frequent staying frequency or frequency of the one user are classified as the private POI related to the one user, and other POIs are classified into the one user. And the destination determination unit, when the current position of one user is the one public POI related to the one user, and the current position collected by the information collecting unit and Based on the current time and / or the distance to each private POI for the one user, the one From among two or more of the private POI according to user, it may be determined the private POI as a destination.

  In the POI recommendation system, the recommendation candidate POI extraction unit refers to the space-time constraint, and extracts the public POI that can stay as an average stay time or more than N percent of the average stay time as the recommendation candidate POI. It may be.

  The POI recommendation system is similar to the public POI related to the one user stored in the public POI storage unit among the recommended candidate public POIs related to the one user extracted by the recommendation candidate POI extraction unit. You may make it further provide the recommendation POI determination part which determines the said recommendation candidate public POI with a high degree as said recommendation public POI.

  In order to solve the above problem, a POI recommendation system according to another aspect of the present invention includes an information collecting unit that collects behavior information of one user including a current position and a current time, and the information collecting unit Based on the plurality of behavior information, a POI extraction unit that extracts the POI related to the one user, and based on the plurality of POI information related to the one user extracted by the POI extraction unit, the POI extraction unit A plurality of extracted POIs relating to one user are a POI having a long stay time or frequency for the one user and a POI having a small stay time or frequency for the one user. A POI classification unit that classifies the public POI related to the one user, and each user classified by the POI classification unit. A public POI storage unit that stores POI information of the public POI, a spatiotemporal constraint setting unit that sets a user's spatiotemporal constraint, and the public POI relating to the one user stored in the public POI storage unit Recommended candidate POI extraction that extracts a public POI having a high degree of similarity as a recommended candidate public POI that is a candidate for a recommended public POI that is recommended as the public POI that the one user moves from the current location to the destination And the spatiotemporal constraint related to one user set by the spatiotemporal constraint setting unit, and the recommended candidate public POI related to one user extracted by the recommended candidate POI extracting unit, And a recommended POI determining unit that determines a recommended public POI. .

  In order to solve the above problem, a POI recommendation server according to another aspect of the present invention uses a public POI that is a POI with less staying time or frequency as a recommended public POI, and is based on a user's space-time constraint in movement between POIs. One of the public POI storage units that store the POI information of the public POI related to each user and the public POIs related to a plurality of users stored in the public POI storage unit. A recommended candidate POI extraction unit that extracts a recommended candidate public POI that is a candidate for the recommended public POI when the user moves from the current location to the destination with reference to the spatio-temporal constraint relating to the one user; , The recommended candidate publication related to one user extracted by the recommended candidate POI extraction unit. A recommended POI determining unit that determines, as the recommended public POI, the recommended candidate public POI having a high similarity to the public POI related to the one user stored in the public POI storage unit from among the POIs Features.

  In order to solve the above problem, a POI recommendation server according to another aspect of the present invention uses a public POI that is a POI with less staying time or frequency as a recommended public POI, and is based on a user's space-time constraint in movement between POIs. A POI recommendation server that outputs the public POI for each user, and a public POI storage unit that stores the POI information of the public POI for each user, and the recommended public POI candidates when one user moves from the current location to the destination As a recommended candidate public POI, a recommended candidate POI that extracts a public POI having a high similarity to the public POI related to the one user from the public POI related to a plurality of users stored in the public POI storage unit One extracted by the extraction unit and the recommended candidate POI extraction unit From among the recommended candidate public POI according to chromatography The, the time with reference to the space constraints of the user of the one, characterized in that it comprises a recommendation POI determination unit configured to determine the recommended public POI.

  According to the present invention, a public POI that stops at the time of movement to the destination without requiring the user to input preference information and does not require sensing of illuminance level, etc. Recommendations can be made in consideration of space constraints. Therefore, for example, when moving between private POIs (for example, moving from work to home), it is possible to recommend a public POI that can be visited in time. It will be possible to provide an opportunity to experience.

It is an example of the functional block diagram of the user terminal 1 by the 1st Embodiment of this invention. 4 is an explanatory diagram for explaining functions of a user terminal 10. FIG. 3 is a flowchart illustrating an example of an operation of a user terminal 10. It is an example of the functional block diagram of the POI recommendation system 1 by the 2nd Embodiment of this invention. It is explanatory drawing for demonstrating the function of the POI recommendation system. It is explanatory drawing for demonstrating the function of the POI recommendation system. 5 is a flowchart showing an example of the operation of the POI recommendation system 1. It is an example of the functional block diagram of the POI recommendation system 2 by the 3rd Embodiment of this invention.

(First embodiment)
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is an example of a functional block diagram of a user terminal 1 according to the first embodiment of the present invention. FIG. 2 is an explanatory diagram for explaining the function of the user terminal 1. As shown in FIG. 1, the user terminal 10 (for example, a mobile phone, which corresponds to a POI attribute determination device) includes an information collection unit 100, a POI extraction unit 110, a POI attribute determination unit 120, a display unit 150, and a storage unit 190. .

  The information collection unit 100 collects (acquires) the current position of the user terminal 10 as the behavior information of the user of the user terminal 10 at a predetermined timing (hereinafter, timing A), and the time when the current position is collected Collect (current time). For example, the information collecting unit 100 collects information from a GPS log (time, latitude, longitude) or a wireless LAN base station. When collecting the behavior information, the information collection unit 100 stores the collected behavior information in the storage unit 190. Note that the timing A is, for example, when a predetermined time has elapsed since the previous collection, or when moved a predetermined distance from the previous collection.

  The information collection unit 100 collects behavior information in response to a request from the POI extraction unit 110 and supplies the behavior information to the POI extraction unit 110.

  The POI extraction unit 110 has a plurality of behavior information collected by the information collection unit 100 at a predetermined timing (hereinafter referred to as timing B), that is, a plurality of behavior information (behavior history information) stored in the storage unit 190. Based on the above, a plurality of POIs related to the user are extracted.

For example, the POI extraction unit 110 performs clustering after removing action information whose moving speed is equal to or higher than a certain speed. That is, the POI extraction unit 110 performs clustering by using a clustering technique for extracting clusters of an arbitrary shape called DBscan, using history data of positions that can be acquired by GPS or positions of connected base stations of WiFi. DBscan has two parameters, MinPts and Eps. A set of points within a distance Eps from a certain point is defined as a neighborhood N _Eps (x). The POI extraction unit 110 extracts a maximal set that can be reached from a certain point by following a directly reachable density as one cluster. In addition, that the density can be directly reached from x to y means that the following formula 1 is satisfied.

  In the case of FIG. 2A which is an example of MinPts = 4, the POI extraction unit 110 performs the above clustering on the position history from 9:00 am to 18:00 pm on weekdays. The above clustering is performed on the position history from 21:00 to 6:00 in the morning, and the largest cluster is set as the home.

  When a plurality of POIs are extracted, the POI extraction unit 110 stores, in the storage unit 190, POI information (position information of each POI and information related to the stay time of each POI) indicating each extracted POI. Note that the timing B is, for example, when a predetermined time has elapsed since the previous extraction or when a predetermined number of behavior information has been accumulated (stored) since the previous extraction.

  In response to a request from the POI attribute determination unit 120, the POI extraction unit 110 supplies POI information indicating the POI that is the current position to the POI attribute determination unit 120. Specifically, when the POI attribute determination unit 120 requests the POI information, the POI extraction unit 110 requests behavior information from the information collection unit 100, and stores a plurality of POI information stored in the storage unit 190. Based on the action information acquired from the information collection unit 100, the POI that is the current position is specified, and the POI information indicating the specified POI is supplied to the POI attribute determination unit 120.

  The POI attribute determination unit 120 requests POI information indicating the POI that is the current position to the POI extraction unit 110 at a predetermined timing (hereinafter referred to as timing C). When the POI attribute determination unit 120 acquires POI information indicating the POI at the current position, the POI attribute determination unit 120 extracts a plurality of POI information related to the user extracted by the POI extraction unit 110 (a plurality of POI information stored in the storage unit 190). ) To determine the attribute of the POI that is the current position (whether it is a private POI or a public POI).

  Specifically, the POI attribute determination unit 120 determines the POI attribute from the stay time and frequency of the POI. For example, the POI attribute determining unit 120 determines that the POI that is the current position is a private POI when the POI that is the current position is the POI up to the second highest POI that has a long stay time and a high frequency. For example, when the POI at the current position is POI1 or POI3 shown in FIG. 2B, it is determined that the POI is a private POI, and when it is POI3, it is a public POI.

  The POI attribute determination unit 120 determines that the POI at the current position is a private POI if the POI at the current position is the top N (N is 3 or more) POIs. Also good. The POI attribute determination unit 120 determines that the POI at the current position is a private POI when the POI at the current position is a POI with a cumulative stay frequency of M (for example, M is 80) percent or more. You may make it determine.

  The timing C is, for example, when a predetermined time has elapsed since the previous determination, when moved a predetermined distance from the previous determination, or when a predetermined operation is received from the user via an input reception unit (not shown). Etc.

  The POI attribute determination unit 120 that has determined the POI attribute that is the current position supplies the determination result to the display unit 150. The display unit 150 acquires the determination result from the POI attribute determination unit 120 and displays it.

  The POI attribute determination unit 120 requests POI information indicating the POI that is the current position from the POI extraction unit 110 at timing C, but the timing A when the information collection unit 100 collects behavior information is short. May acquire the latest POI information from the storage unit 190 as POI information indicating the POI at the current position, instead of requesting the POI information indicating the POI at the current position from the POI extraction unit 110. .

  FIG. 3 is a flowchart illustrating an example of the operation of the user terminal 10. The flowchart shown in FIG. 3 always starts when the user terminal 10 is powered on. In FIG. 3, the information collection unit 100 determines whether or not a predetermined time has elapsed since the previous collection (step S10). When the information collection unit 100 determines that a predetermined time has elapsed (step S10: Yes), the information collection unit 100 collects behavior information and stores it in the storage unit 190 (step S20). On the other hand, when the information collection unit 100 determines that the predetermined time has not elapsed (step S10: No), the information collection unit 100 skips step S20 and proceeds to step S30.

  Following step S20 or step S10 (No), the POI extraction unit 110 determines whether or not a predetermined number of behavior information has been newly accumulated since the previous extraction (step S30). When the POI extraction unit 110 determines that behavior information has been newly accumulated (step S30: Yes), the POI extraction unit 110 performs a plurality of operations related to the user based on the plurality of behavior information (behavior history information) stored in the storage unit 190. The POI is extracted (step S40), and the POI information indicating each extracted POI is stored in the storage unit 190 (step S50). On the other hand, when it is determined that the behavior information is not newly accumulated (step S30: No), the POI extraction unit 110 skips step S40 and step S50 and proceeds to step S60.

  Subsequent to step S50 or step S30 (No), the POI attribute determination unit 120 determines whether a predetermined operation has been received from the user via an input reception unit (not shown) (step S60). That is, the POI attribute determination unit 120 determines whether or not a predetermined operation has been performed by the user. When it is determined that a predetermined operation has been received from the user (step S60: Yes), the POI attribute determination unit 120 acquires POI information indicating the POI that is the current position from the POI extraction unit 110, and is stored in the storage unit 190. Based on the plurality of POI information, the attribute of the POI which is the current position is determined (step S70). The display unit 150 acquires the determination result from the POI attribute determination unit 120 and displays it (step S80). Then, this flowchart ends.

  On the other hand, when the POI attribute determination unit 120 determines that a predetermined operation has not been received from the user (step S60: No), step S70 and step S80 are skipped, and this flowchart ends.

  As described above, according to the first embodiment, it is possible to determine whether the current location is a private POI or a public POI from only the user action history information. Therefore, depending on whether the current location is a private POI or a public POI, the user needs to input in cooperation with a device that recommends the public POI to stop when moving from the current location to the destination. Therefore, the recommended POI can be presented easily.

(Second Embodiment)
Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 4 is an example of a functional block diagram of the POI recommendation system 1 according to the second embodiment of the present invention. 5 and 6 are explanatory diagrams for explaining the function of the user terminal 11. As shown in FIG. 4, the POI recommendation system 1 includes a user terminal 11 and a server 21 (corresponding to a POI recommendation server). The user terminal 11 (for example, a mobile phone) includes an information collection unit 101, a POI extraction unit 111, a POI classification unit 121, a destination determination unit 131, a spatiotemporal constraint setting unit 141, a display unit 151, and a storage unit 191. The server 21 includes a recommended candidate POI extraction unit 201, a similarity calculation unit 211, a recommended POI determination unit 221, a public POI storage unit 281, and a similarity information storage unit 291.

  The information collection unit 101 collects (acquires) the current position of the user terminal 11 as behavior information of the user of the user terminal 11 at a predetermined timing (hereinafter, referred to as timing D), and the time when the current position is collected Collect (current time). Note that the collection by the information collection unit 101 is the same as the information collection unit 100 of the first embodiment. When the behavior information is collected, the information collection unit 101 stores the collected behavior information in the storage unit 191. Note that the timing D is, for example, when a predetermined time has elapsed since the last collection or when a predetermined distance has been moved since the previous collection.

  The information collection unit 101 collects behavior information in response to a request from the POI extraction unit 111 and supplies the behavior information to the POI extraction unit 111.

  The POI extraction unit 111 has a plurality of behavior information collected by the information collection unit 101 at a predetermined timing (hereinafter referred to as timing E), that is, a plurality of behavior information (behavior history information) stored in the storage unit 191. Based on the above, a plurality of POIs related to the user are extracted. The POI extraction by the POI extraction unit 111 is the same as the POI extraction unit 110 of the first embodiment.

  When a plurality of POIs are extracted, the POI extraction unit 111 temporarily stores POI information indicating each extracted POI in the storage unit 191 and notifies the POI classification unit 121 that the POI has been extracted. The timing E is, for example, when a predetermined time has elapsed since the previous extraction or when a predetermined number of behavior information has been accumulated (stored) since the previous extraction.

  In addition, the POI extraction unit 111 supplies POI information indicating the POI that is the current position to the destination determination unit 131 at a predetermined timing (hereinafter, referred to as timing F). Specifically, the POI extraction unit 111 requests behavior information from the information collection unit 101 at timing F, and includes the plurality of POI information stored in the storage unit 191 and the behavior information acquired from the information collection unit 101. The POI (private POI / public POI) that is the current position is specified based on the information, and the POI information indicating the specified POI is supplied to the destination determination unit 131. The timing F is, for example, when a predetermined time has elapsed since the previous determination, when moved a predetermined distance from the previous determination, or when a predetermined operation is received from the user via an input reception unit (not shown). Etc.

  The POI extraction unit 111 requests behavior information from the information collection unit 101 at timing F, and based on a plurality of POI information stored in the storage unit 191 and the behavior information acquired from the information collection unit 101. If the timing D at which the information collection unit 101 collects behavior information is short, the POI that is the current position is specified, but instead of the above, out of the plurality of POI information stored in the storage unit 191 The POI related to the latest POI information may be specified as the POI that is the current position.

  When the POI information indicating the POI that is the current position is supplied to the destination determining unit 131, the POI extracting unit 111 acquires POI information indicating the POI that is the destination (private POI) from the destination determining unit 131. When the POI extraction unit 111 acquires POI information indicating the POI that is the destination, the POI extraction unit 111 acquires information indicating the spatiotemporal constraint from the spatiotemporal constraint setting unit 141, and obtains the POI (private POI / public POI) that is the current position. The POI information indicating the POI information indicating the destination POI (private POI), and the information indicating the spatiotemporal constraint are supplied to the recommended candidate POI extraction unit 201 of the server 21.

  The POI classification unit 121 acquires a notification that the POI has been extracted from the POI extraction unit 111. When the POI classification unit 121 acquires the notification that the POI has been extracted, the POI classification unit 121 extracts a plurality of POI information related to the user extracted by the POI extraction unit 111, that is, a plurality of POI information temporarily stored in the storage unit 191. Based on the above, each POI is classified by attribute. That is, the POI classification unit 121 classifies each POI into a private POI and a public POI.

  Specifically, the POI classification unit 121 classifies each POI from the stay time and frequency of each POI. For example, as shown in FIG. 2 (b), the POI classification unit 121 classifies the POIs (POI1 and POI3) having the longest stay and the highest frequency into private POIs, and other POIs (POI2). Classify as public POI. In addition, the time zone of the day is divided into daytime (for example, 9: 00 to 21:00) and night (for example, 21: 00 to 9:00 of the next morning), and the top time that is long and the frequency is high in each. The POIs may be classified as private POIs.

  The POI classification unit 121 may classify the top N (N is 3 or more) POIs into private POIs. The POI classification unit 121 may classify a POI having a cumulative stay frequency of M (for example, M is 80) percent or more as a private POI.

  When the POI classification unit 121 classifies a plurality of POIs into private POIs and public POIs, the storage unit 191 stores POI information (position information of each POI and information related to the stay time of each POI) indicating the private POI. And stores (supplies) POI information (position information of each POI and information related to the stay time of each POI) in the public POI storage unit 281 of the server 21.

  The destination determination unit 131 acquires POI information indicating the POI that is the current position from the POI extraction unit 111. When acquiring the POI information indicating the POI that is the current position, the destination determining unit 131 determines a destination when the recommended candidate POI extracting unit 201 extracts a recommended candidate public POI (described later). Specifically, the destination determining unit 131 that has acquired POI information indicating the POI that is the current position determines by estimation which private POI the destination is.

  Specifically, for example, the destination determination unit 131 classifies the top three or more POIs with the highest stay time or frequency as private POIs and classifies other POIs as public POIs (that is, storage POIs). And when the POI information indicating three or more private POIs is stored in the unit 190) and the POI that is the current position is any private POI (that is, the same position as the current position) When the POI information related to the private POI having information is stored in the storage unit 191), the information is stored in the storage unit 191 based on the current position and the current time collected by the information collection unit 101. A plurality of other private Ps that are different from the private POI) From among the I, to determine the private POI to be the destination.

  In addition, for example, the destination determination unit 131 has acquired POI information indicating the POI that is the current position, and the destination determination unit 131 classifies the top two POIs with the highest stay time or frequency as private POIs. When another POI is classified as a public POI (that is, when POI information indicating two private POIs is stored in the storage unit 190), the POI at the current position is one private POI. (That is, when POI information related to a private POI having the same position information as the current position is stored in the storage unit 191), the other private POI that is different from the POI that is the current position Determine as the ground.

  Also, for example, the destination determination unit 131 that has acquired POI information indicating the POI that is the current position classifies the top two or more POIs with the longest stay time or frequency as private POIs, and makes other POIs public When it is classified as POI (that is, when POI information indicating two or more private POIs is stored in the storage unit 190) and the POI at the current position is a public POI (that is, the relevant POI information) (If POI information related to private POI having the same position information as the current position is not stored in the storage unit 191), based on the current position and current time collected by the information collecting unit 101, and / or Based on the distance to each private POI, from among the above two or more private POIs, To determine the private POI to be the earth.

  That is, the destination determining unit 131 determines the destination according to the rules as shown in FIG. According to the rules shown in FIG. 5A, for example, when the destination determination unit 131 acquires a home (a POI information indicating a home) that is a private POI as the current location at 8:00 on weekdays, A company that is a POI is determined as a destination (see I in FIG. 5A). Also, for example, when the destination determination unit 131 acquires another place (POI information indicating each place) that is a public POI as the current location at 16:00 on weekdays, the distance from the current location to the company that is a private POI ≦ When the distance is from your current location to your home, which is a private POI, the company that is the private POI is determined as the destination, and when the distance from your current location to the company that is the private POI> the distance from your current location to your home, which is the private POI The home which is POI is determined as the destination (see II in FIG. 5A). The rule is stored in a storage unit (for example, the storage unit 191) referred to by the destination determination unit 131. However, the content of the rule may naturally be different for each user terminal 11. Often, the user terminal 11 accepts a change of the rule from the user via an input acceptance unit (not shown). For example, the user may change the rules in consideration of the staying time zone of his / her home.

  When the destination determination unit 131 determines a private POI as a destination, the destination determination unit 131 supplies POI information indicating the private POI as a destination to the POI extraction unit 111.

  The spatiotemporal constraint setting unit 141 accepts a spatiotemporal constraint in movement between POIs from the user via an input accepting unit (not shown). When receiving the spatiotemporal constraint, the spatiotemporal constraint setting unit 141 stores (sets) information indicating the spatiotemporal constraint and supplies the information to the display unit 151.

  If there is no space-time constraint (when information indicating the space-time constraint is not set), the remaining time from the latest time to arrive at the destination-the travel time from the current location to the destination is the grace time. When the public POI that can be visited within the grace period is extracted as the recommended candidate POI and there is a spatiotemporal constraint (when information indicating the spatiotemporal constraint is set), the time set as the spatiotemporal constraint and the above grace period Public POIs that can be visited within the shorter time are extracted as recommended candidate POIs.

  The display unit 151 displays the spatiotemporal constraint from the spatiotemporal constraint setting unit 141. In addition, the display unit 151 acquires POI information indicating the recommended POI from the recommended POI determination unit 221 of the server 21 and displays it.

  The public POI storage unit 281 is supplied from the POI classification unit 121 of the plurality of user terminals 11 and indicates the POI information indicating the public POIs of a plurality of users (information having location information of each POI and information related to the stay time of each POI). Are stored in association with user information (for example, identification information for identifying the user terminal 11). That is, the POI information of public POIs of a plurality of users is stored for each user.

  The recommended candidate POI extraction unit 201 receives, from the POI extraction unit 111 of the user terminal 11, POI information indicating the POI (private POI / public POI) as the current position, POI information indicating the POI (private POI) as the destination, and , Get information indicating space-time constraints.

  When the candidate candidate POI extraction unit 201 acquires POI information indicating the POI that is the current position, POI information indicating the POI that is the destination, and information indicating the spatio-temporal constraint, a plurality of items stored in the public POI storage unit 281 are stored. Among the public POIs related to the user, the POI recommended as the public POI that stops when the user travels from the current location (the POI that is the current position) to the destination (the POI that is the destination) A POI that is a candidate for “POI” (referred to as a recommended candidate public POI) is extracted.

  For example, the recommendation candidate POI extraction unit 201 refers to information indicating the spatio-temporal constraint, and extracts a public POI that allows staying from the plurality of public POIs in the public POI storage unit 281 as the recommendation candidate POI. More specifically, the recommended candidate POI extraction unit 201 extracts a POI that satisfies the space-time constraint and can stay longer than the average stay time of each public POI (or N percent of the average stay time). In the case of the example shown in FIG. 5B, if the space-time constraint is 3 hours, the recommended candidate POI extraction unit 201 determines the required time from the departure point O to each public POI, from each public POI to the destination P. The public POIA and B are extracted as recommended candidate POIs as public POIs whose total required time and the total average stay time of each public POI are below the space-time constraint. Note that the average stay time of each public POI may be the average stay time for each time zone.

  When the recommended candidate public POI is extracted, the recommended candidate POI extracting unit 201 supplies the extracted recommended candidate public POI to the recommended POI determining unit 221.

  The similarity calculation unit 211 refers to the POI information of the public POIs of a plurality of users stored in the public POI storage unit 281 at a predetermined timing (hereinafter referred to as timing G), and calculates the similarity of each public POI. . For example, the similarity calculation unit 211 calculates the similarity between POIs in distant areas, assuming that users who have a co-reference to a specific position have a high similarity.

  For example, when there is a stay history of each POI of the user (user A, user B, user C) shown in FIG. 6A, the similarity calculation unit 211 has two POI evaluation vectors v1 and v2. Cos (v1, v2) is calculated by the following formula 2. For example, the similarity calculation unit 211 calculates the cosine similarity of POI1 = {1, 0, 1} and POI4 = {0, 0, 1} as shown in Equation 3 below.

  The similarity calculation unit 211 calculates the cosine similarity between other POIs in the same manner, and the cosine similarity matrix shown in FIG. 6B is obtained from FIG.

  When calculating the similarity of each public POI, the similarity calculation unit 211 stores the similarity information of each public POI (for example, cosine similarity information shown in FIG. 6B) in the similarity information storage unit 291. . The timing G is, for example, when a predetermined time has elapsed since the previous calculation or when a predetermined number of POI information has been accumulated (stored) since the previous calculation.

  The recommended POI determining unit 221 acquires the recommended candidate public POI from the recommended candidate POI extracting unit 201. When the recommended POI determining unit 221 acquires the recommended candidate public POI, the recommended POI determining unit 221 determines the recommended public POI from the recommended candidate public POI. Specifically, the recommended POI determining unit 221 refers to the similarity information stored in the similarity information storage unit 291 and determines a recommended candidate public POI having a high similarity to the user's public POI as the recommended public POI. .

  For example, the recommended POI determination unit 221 determines a recommended POI using item-based collaborative filtering. As an example, an example will be described in which the recommended POI of user A is determined from candidate recommended POI = {POI2, POI4}. It is assumed that cosine similarity information shown in FIG. 6B is stored in the similarity information storage unit 291.

Recommendation POI determination unit 221, POI a history of the user _{A {POI A, 1, POI} A, 2, ···} and, if the candidate POI and _{POI Candidat,} recommendation level P of the candidate _{POI Candidat} for the user A ( A, POI _candid ) is calculated by the following equation 4. For example, the recommended POI determination unit 221 calculates the recommendation level P of the candidate recommendation POI _{2 as in} the following formula 5, and calculates the recommendation level P of the candidate recommended POI _{4 as in} the following formula 6.

Then, the recommended POI determination unit 221 determines candidate recommended POI ₄ having a large recommendation level P as a recommended POI. Note that the recommended POI determining unit 221 may determine a recommended candidate POI with a recommendation degree P or subsequent points as a recommended POI depending on whether the user A has already visited POI.

  When determining the recommended public POI, the recommended POI determining unit 221 supplies the determined recommended public POI to the display unit 151 of the user terminal 11.

  FIG. 7 is a flowchart showing an example of the operation of the POI recommendation system 1. The flowchart shown in FIG. 7A is a flowchart showing an example of the operation of the user terminal 11, and the flowchart shown in FIG. 7B is a flowchart showing an example of the operation of the server 21. Note that the flowchart shown in FIG. 7A always starts when the user terminal 11 is powered on, and the flowchart shown in FIG. 7B is the POI information indicating the POI at the current position from the user terminal 11 and the destination. The process starts by acquiring POI information indicating POI and information indicating space-time constraints. 7A and 7B, the spatiotemporal constraint setting unit 141 stores therein information indicating spatiotemporal constraints, and the public POI storage unit 281 indicates public POIs of a plurality of users. It is assumed that POI information is stored, and the similarity information storage unit 291 stores similarity information of each public POI.

  In FIG. 7A, the information collection unit 101 determines whether a predetermined time has elapsed since the previous collection (step S100). If the information collection unit 101 determines that a predetermined time has elapsed (step S100: Yes), the information collection unit 101 collects behavior information and temporarily stores it in the storage unit 191 (step S110). On the other hand, when the information collection unit 101 determines that the predetermined time has not elapsed (step S100: No), the information collection unit 101 skips step S110 and proceeds to step S120.

  Subsequent to step S110 or step S100 (No), the POI extraction unit 111 determines whether or not a predetermined number of behavior information has been newly accumulated since the previous extraction (step S120). When it is determined that the behavior information has been newly accumulated (step S120: Yes), the POI extraction unit 111 relates to the user based on a plurality of behavior information (behavior history information) temporarily stored in the storage unit 191. A plurality of POIs are extracted, and POI information indicating each POI is temporarily stored in the storage unit 191 (step S130). The POI extraction unit 111 notifies the POI classification unit 121 that the POI has been extracted.

  The POI classification unit 121 that has acquired the notification that the POI has been extracted from the POI extraction unit 111, based on the plurality of POI information temporarily stored in the storage unit 191, converts the plurality of POIs into private POI and public POI. Classify (step S140). The POI classification unit 121 that classifies a plurality of POIs into private POIs and public POIs stores POI information indicating the private POIs in the storage unit 191 (step S150), and POI information indicating the public POIs in the public POI of the server 21. The data is stored (supplied) in the storage unit 281 (step S160).

  On the other hand, when the POI extraction unit 111 determines that action information is not newly accumulated (step S120: No), the process skips step S130 to step S160 and proceeds to step S170.

  Subsequent to step S160 or step S120 (No), the POI extraction unit 111 determines whether a predetermined operation has been received from the user via an input reception unit (not shown) (step S170). That is, the POI extraction unit 111 determines whether or not there is a predetermined operation by the user.

  If the POI extraction unit 111 determines that a predetermined operation has been received from the user (step S170: Yes), the POI extraction unit 111 identifies the POI that is the current position, and supplies the POI information indicating the identified POI to the destination determination unit 131. The destination determining unit 131 that has acquired the POI information indicating the POI that is the current position determines a destination for extracting the recommended candidate public POI (step S180). The destination determination unit 131 that has determined the private POI as the destination supplies the POI information indicating the private POI as the destination to the POI extraction unit 111.

  The POI extraction unit 111 that has acquired the POI information indicating the POI that is the destination acquires the information indicating the spatiotemporal constraint from the spatiotemporal constraint setting unit 141, and the POI information that indicates the POI that is the current position, and the POI that is the destination And the information indicating the spatio-temporal constraint are supplied to the recommended candidate POI extraction unit 201 of the server 21 (step S190). Then, this flowchart ends. In addition, after execution of step S190, the user terminal 11 acquires a recommended public POI from the server 21.

  On the other hand, when it is determined that the POI extraction unit 111 has not received a predetermined operation from the user (step S170: No), step S180 and step S190 are skipped, and this flowchart ends.

  In FIG. 7B, the recommended candidate POI extraction unit 201 that has acquired the POI information indicating the POI that is the current position, the POI information that indicates the POI that is the destination, and the information indicating the spatiotemporal constraint is stored in the public POI storage unit 281. A recommendation candidate public POI is extracted from public POIs related to a plurality of stored users (step S210). The recommendation candidate POI extraction unit 201 that has extracted the recommendation candidate public POI supplies the recommendation candidate public POI to the recommendation POI determination unit 221.

  The recommended POI determining unit 221 that has acquired the recommended candidate public POI refers to the similarity information stored in the similarity information storage unit 291 and determines a recommended public POI from the recommended candidate public POIs (step S220). The recommended POI determination unit 221 that has determined the recommended public POI supplies the recommended public POI to the user terminal 11. Then, this flowchart ends. In addition, after execution of step S220, the user terminal 11 displays the recommended public POI acquired from the server 21 on the display unit 151.

  As described above, according to the second embodiment, it is determined whether the current location is a private POI or a public POI from only the user action history information, and depending on whether the current location is a private POI or a public POI, The public POI that stops when moving from the current location to the destination can be recommended in consideration of the user's space-time constraints.

  In the second embodiment, the server 21 supplies the recommended POI determined by the recommended POI determining unit 221 to the user terminal 11. However, the server 21 recommends the recommendation extracted by the recommended candidate POI extracting unit 201. The candidate POI may be supplied to the user terminal 11. When the server 21 supplies the recommended candidate POI to the user terminal 11, the recommended POI determination unit 221 may not be provided.

  In the second embodiment, the user terminal 11 includes the information collection unit 101, the POI extraction unit 111, the POI classification unit 121, the destination determination unit 131, the spatiotemporal constraint setting unit 141, the display unit 151, and the storage unit 191. The configuration in which the server 21 includes the recommended candidate POI extraction unit 201, the similarity calculation unit 211, the recommended POI determination unit 221, the public POI storage unit 281 and the similarity information storage unit 291 has been described. The configuration is not limited to this. For example, the user terminal 11 may be configured to include the recommended POI determination unit 221 instead of the server 21. Further, for example, the server 21 may include one or more of the POI extraction unit 111, the POI classification unit 121, the destination determination unit 131, the spatiotemporal constraint setting unit 141, and the storage unit 191 instead of the user terminal 11. Good.

(Third embodiment)
Subsequently, a third embodiment of the present invention will be described with reference to the drawings. A POI recommendation system 2 according to the third embodiment of the present invention is a modification of the POI recommendation system 1 according to the second embodiment. That is, the POI recommendation system 1 according to the second embodiment is configured to first extract the recommended candidate POI based on the spatiotemporal constraint and then determine the candidate POI from the recommended candidate POI based on the similarity information. In the POI recommendation system 2 according to the third embodiment, a recommendation candidate POI is first extracted based on similarity information, and then a candidate POI is determined from the recommendation candidate POI based on a spatiotemporal constraint.

  As shown in FIG. 8, the POI recommendation system 2 includes a user terminal 12 and a server 22 (corresponding to a POI recommendation server). The user terminal 12 (for example, a mobile phone) includes an information collection unit 102, a POI extraction unit 112, a POI classification unit 122, a destination determination unit 132, a spatiotemporal constraint setting unit 142, a display unit 152, and a storage unit 192. The information collection unit 102, the POI classification unit 122, the destination determination unit 132, the spatiotemporal constraint setting unit 142, the display unit 152, and the storage unit 192 are the information collection unit 101 and the POI classification unit of the user terminal 11 in the second embodiment. 121, the destination determination unit 131, the spatiotemporal constraint setting unit 141, the display unit 151, and the storage unit 191, and thus description thereof is omitted. The server 22 includes a recommended candidate POI extraction unit 202, a similarity calculation unit 212, a recommended POI determination unit 222, a public POI storage unit 282, and a similarity information storage unit 292. The similarity calculation unit 212, the public POI storage unit 282, and the similarity information storage unit 292 are the same as the similarity calculation unit 211, the public POI storage unit 281 and the similarity information storage unit 291 of the server 21 in the second embodiment. Therefore, the description is omitted.

  The recommendation candidate POI extraction unit 202 extracts a public POI having a high similarity to the public POI related to one user stored in the public POI storage unit 282 as the recommendation candidate public POI. Specifically, the recommendation candidate POI extraction unit 202 refers to the similarity information stored in the similarity information storage unit 292, and determines a public POI having a high similarity to the user's public POI as the recommendation candidate public POI. .

  The POI extraction unit 112 is substantially the same as the POI extraction unit 111 in the second embodiment. That is, the POI extraction unit 111 in the second embodiment extracts the POI information indicating the POI that is the current position, the POI information that indicates the POI that is the destination, and the information that indicates the spatiotemporal constraint, from the recommended candidate POI of the server 21. The POI extraction unit 112 according to the third embodiment is supplied to the recommended POI determination unit 222 of the server 21.

  The recommended POI determination unit 222 refers to the spatio-temporal constraint related to one user set by the spatio-temporal constraint setting unit 141, and from the recommended candidate public POI related to the user extracted by the recommended candidate POI extraction unit 202 Determine the recommended public POI.

  For example, the recommended POI determination unit 222 refers to information indicating the spatiotemporal constraint, and recommends a public POI that can stay longer than the average stay time (or N percent of the average stay time) from the recommended candidate public POIs. Determine as POI. If the recommended public POI is determined, the recommended POI determining unit 222 supplies the determined recommended public POI to the display unit 152 of the user terminal 12.

  As mentioned above, according to 3rd Embodiment, the effect similar to 2nd Embodiment can be acquired. In the third embodiment, the user terminal 12 includes the information collection unit 102, the POI extraction unit 112, the POI classification unit 122, the destination determination unit 132, the spatiotemporal constraint setting unit 142, the display unit 152, and the storage unit 192. The configuration in which the server 22 includes the recommended candidate POI extraction unit 202, the similarity calculation unit 212, the recommended POI determination unit 222, the public POI storage unit 282, and the similarity information storage unit 292 has been described. The configuration is not limited to this. For example, the user terminal 12 may include the recommended POI determination unit 222 instead of the server 22. For example, instead of the user terminal 12, the server 22 may include one or more of the POI extraction unit 112, the POI classification unit 122, the destination determination unit 132, the spatiotemporal constraint setting unit 142, and the storage unit 192. Good.

  Note that a program for executing the processes of the user terminals 10, 11, the server 21, and the server 22 according to an embodiment of the present invention is recorded on a computer-readable recording medium, and the program recorded on the recording medium is recorded. The above-described various processes related to the processes of the user terminals 10 and 11, the server 21, and the server 22 according to the embodiment of the present invention may be performed by being read and executed by a computer system. Here, the “computer system” may include an OS and hardware such as peripheral devices. Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used. The “computer-readable recording medium” means a flexible disk, a magneto-optical disk, a ROM, a writable nonvolatile memory such as a flash memory, a portable medium such as a CD-ROM, a hard disk built in a computer system, etc. This is a storage device.

  Further, the “computer-readable recording medium” refers to a volatile memory (for example, DRAM (Dynamic) in a computer system serving as a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. Random Access Memory)) that holds a program for a certain period of time is also included. The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, what is called a difference file (difference program) may be sufficient.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes designs and the like that do not depart from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 ... POI recommendation system 2 ... POI recommendation system 10 ... User terminal (POI attribute determination apparatus)
11, 12 ... User terminals 21, 22 ... Server (POI recommendation server)
100, 101, 102 ... information collection unit 110, 111, 112 ... POI extraction unit 120 ... POI attribute determination unit 121, 122 ... POI classification unit 131, 132 ... destination determination unit 141, 142 ... spatiotemporal constraint setting unit 150, 151, 152 ... display units 190, 191, 192 ... storage units 201, 202 ... recommended candidate POI extraction units 211, 212 ... similarity calculation units 221, 222 ... recommended POI determination units 281, 282 ... public POI storage units 291, 292 ... Similarity information storage unit

Claims (9)

An information collection unit that collects behavior information of one user including the current position and current time;
A POI extraction unit that extracts a POI related to the one user based on the plurality of behavior information collected by the information collection unit;
Based on a plurality of POI information related to one user extracted by the POI extraction unit, a plurality of POIs related to one user extracted by the POI extraction unit have a longer staying time or frequency for the one user. A POI classification unit for classifying a private POI related to the one user who is a POI and a public POI related to the one user which is a POI whose stay time or frequency is low for the one user;
A public POI storage unit for storing POI information of the public POI related to each user classified by the POI classification unit;
A spatio-temporal constraint setting unit for setting a user's spatio-temporal constraint in movement between POIs;
Referring to the spatio-temporal constraint relating to one user set by the spatio-temporal constraint setting unit, the movement of the one user from the current location to the destination from among the public POIs relating to a plurality of users. A POI recommendation system comprising: a recommendation candidate POI extraction unit that extracts a recommendation candidate public POI that is a candidate of a recommended public POI recommended as the public POI that stops at the time. The recommended candidate POI extraction unit
Among the plurality of POIs related to one user extracted by the POI extraction unit, the POI classifying unit selects the top two POIs having the highest stay time or frequency of the one user as the private user If the POI is classified as a POI and the other POI is classified as the public POI related to the one user, the current position of the one user is the one private POI related to the one user. 2. The POI recommendation system according to claim 1 , wherein the recommendation candidate public POI up to the other private POI is extracted with the other private POI different from the private POI as a destination. The recommended candidate POI extracting unit further comprises a destination determining unit for determining a destination when the recommended candidate public POI is extracted,
The POI classification unit
Among the plurality of POIs related to one user extracted by the POI extraction unit, classify the top two or more POIs with the most frequent stay time or frequency of the one user as the private POI related to the one user, Classify other POIs as the public POI for the one user,
The destination determination unit
The POI classifying unit classifies the top three or more POIs with the most frequent staying frequency or frequency of the one user from the plurality of POIs relating to the one user as the private POI relating to the one user, When the POI is classified as the public POI related to the one user, and the current position of the one user is the one private POI related to the one user, the current collected by the information collecting unit 2. The POI recommendation according to claim 1 , wherein the private POI as a destination is determined from a plurality of other private POIs different from the one private POI based on the position and the current time. system. The recommended candidate POI extracting unit further comprises a destination determining unit for determining a destination when the recommended candidate public POI is extracted,
The POI classification unit
Among the plurality of POIs related to one user extracted by the POI extraction unit, classify the top two or more POIs with the most frequent stay time or frequency of the one user as the private POI related to the one user, Classify other POIs as the public POI for the one user,
The destination determination unit
When the current position of one user is the one public POI related to the one user, the current position and the current time collected by the information collecting unit, and / or the respective one related to the one user 2. The POI recommendation system according to claim 1 , wherein the private POI as a destination is determined from among the two or more private POIs related to the one user based on a distance to the private POI. The recommended candidate POI extraction unit
Referring to the space-time constraints, the public POI capable stay more than N% of the average residence time or average residence time of claims 1, wherein the extracting as the recommended candidate POI according to claim 4 POI recommendation system. Among the recommendation candidate public POIs related to one user extracted by the recommendation candidate POI extraction unit, the recommendation candidates having high similarity to the public POI related to the one user stored in the public POI storage unit The POI recommendation system according to any one of claims 1 to 5 , further comprising a recommended POI determination unit that determines a public POI as the recommended public POI. An information collection unit that collects behavior information of one user including the current position and current time;
A POI extraction unit that extracts a POI related to the one user based on the plurality of behavior information collected by the information collection unit;
Based on a plurality of POI information related to one user extracted by the POI extraction unit, a plurality of POIs related to one user extracted by the POI extraction unit have a longer staying time or frequency for the one user. A POI classification unit for classifying a private POI related to the one user who is a POI and a public POI related to the one user which is a POI whose stay time or frequency is low for the one user;
A public POI storage unit for storing POI information of the public POI related to each user classified by the POI classification unit;
A spatio-temporal constraint setting unit for setting a user's spatio-temporal constraint in movement between POIs;
A public POI having a high similarity to the public POI related to the one user stored in the public POI storage unit is recommended as the public POI to which the one user stops when moving from the current location to the destination. A recommended candidate POI extraction unit that extracts a recommended candidate public POI to be a candidate for the recommended public POI
The recommended public POI is referred to from among the recommended candidate public POIs related to one user extracted by the recommended candidate POI extracting unit with reference to the spatiotemporal constraints related to one user set by the spatiotemporal constraint setting unit. A POI recommendation system comprising a recommended POI determination unit for determining A POI recommendation server that outputs a public POI that is a POI with less staying time or frequency as a recommended public POI, based on a user's space-time constraints in movement between POIs,
A public POI storage unit for storing POI information of the public POI relating to each user;
Among the public POIs related to a plurality of users stored in the public POI storage unit, a recommended candidate public POI that is a candidate for the recommended public POI when one user moves from the current location to the destination is A recommended candidate POI extraction unit that refers to and extracts the space-time constraint relating to one user;
Among the recommendation candidate public POIs related to one user extracted by the recommendation candidate POI extraction unit, the recommendation candidates having high similarity to the public POI related to the one user stored in the public POI storage unit A POI recommendation server comprising: a recommended POI determination unit that determines a public POI as the recommended public POI. A POI recommendation server that outputs a public POI that is a POI with less staying time or frequency as a recommended public POI, based on a user's space-time constraints in movement between POIs,
A public POI storage unit for storing POI information of the public POI relating to each user;
As a recommended candidate public POI that is a candidate for the recommended public POI when moving from the current location to the destination, a public POI having a high similarity to the public POI related to the one user is stored in the public POI memory. A recommended candidate POI extraction unit that extracts from among the public POIs related to a plurality of users stored in the unit;
A recommended POI determining unit that determines the recommended public POI by referring to the space-time constraint related to the one user from the recommended candidate public POIs related to the one user extracted by the recommended candidate POI extracting unit; A POI recommendation server comprising:

Images