JPWO2012169152A1 - Moving destination prediction apparatus, moving destination prediction method, and moving destination prediction program - Google Patents

Abstract

The stay point set generation unit 82 generates a stay point set that is a set of position information logs measured at a stay point where the user frequently stays. The movement end location determination means 83 uses the generated stay point set to determine the stay location that is the movement end location of each position information log. The movement end location registration unit 84 stores the determined stay location as the movement end location in the movement end location storage unit 81 in association with each position information log. When a new location information log of the user is input, the destination prediction unit 85 stores a stay location corresponding to the location information log having a predetermined close relationship with the new location information log. Extracted from the means 81, the extracted staying place is predicted as the destination to which the user moves.

Description

  The present invention relates to a movement destination prediction apparatus, a movement destination prediction method, and a movement destination prediction program for predicting a movement destination of a user from position information measured irregularly.

  In recent years, LBS (Location Based Service) such as life logs and communication tools using position information has been expanded. As an LBS, for example, an application that uses a GPS (Global Positioning System) function mounted on a mobile phone or the like to add a location information log to a photograph taken with a mobile terminal with a camera function, or “tweet” with a location information log Twitter (registered trademark) and the like that can be used are widely used. In addition, as an example of LBS, there is a recommendation service that distributes content in accordance with a user's behavior pattern that can be obtained from personal location information.

  Non-Patent Document 1 describes a behavior pattern analysis method for analyzing a behavior pattern from position information measured irregularly. In the behavior pattern analysis method described in Non-Patent Document 1, latitude / longitude and time obtained by irregular positioning are used to indicate position information indicating a frequently visited place (hereinafter referred to as a stay place) and a moving route. Extract location information clusters. By using this cluster, it is possible to separate a place where a user frequently goes from other positions.

  Patent Document 1 describes a behavior prediction device that predicts a user's future behavior based on information indicating a user's place of stay and past behavior patterns using the information indicating the place of stay. Has been. The behavior prediction apparatus described in Patent Literature 1 calculates in advance the appearance probability of information indicating a stay location included in the behavior pattern, and information indicating the current stay location appears based on the calculated appearance probability. Predict user behavior from probability.

JP 2010-250759 A

Fumitaka Nakahara, Takahiro Murakami, “Analysis method of behavior pattern based on irregular positioning information history,” Ubiquitous Computing System, 2011-03-07, 2011-UBI-29, No.4, pp.1- 8

  When providing LBS, in order to improve the accuracy of the service, it is desirable to be able to periodically position information at short intervals. However, since the location information is privacy information, there is a user's desire that the location information that has been measured regularly is not provided to the service provider. Therefore, there are many users who refrain from using the service by periodically positioning the location information.

  From such a situation, for example, the positioning interval of the location information using the terminal with the location information acquisition function may be long, or the positioning of the location information may be irregular. Even when it is difficult to periodically measure the position information at short intervals in this way, by using the method described in Non-Patent Document 1, the user's daily behavior such as where the user frequently stays. It is possible to obtain the pattern with high accuracy. However, it is difficult to predict the user's travel destination only by using the method described in Non-Patent Document 1.

  Further, in the method described in Patent Document 1 in which the destination is predicted using only the staying place, in the case of irregular positioning, before the position information of the staying place is measured, the place where the staying is not performed (for example, moving) Position information, etc.) may be measured. In such a case, it is difficult to perform behavior prediction by the method described in Patent Document 1.

  Since the staying place is a place where the user frequently goes, it is possible to specify the staying place by measuring the latitude and the longitude a plurality of times at the same place. However, the location on the moving route is not always measured at the same latitude and route as in the staying place. That is, it is difficult to locate and specify a location on the movement route. Therefore, for example, when the position of the moving position is measured after the server determines the staying place of the user, the behavior can be predicted. However, when only the information that the server is moving is measured, behavior prediction (specifically, prediction of the destination) is difficult.

  Therefore, the present invention provides a movement destination prediction apparatus, a movement destination prediction method and a movement destination prediction method capable of predicting a user's movement destination even when the position information positioning interval is long and the position information positioning is irregular. The purpose is to provide a moving destination prediction program.

  The movement destination prediction apparatus according to the present invention includes a position information log that is information including a positioning position and positioning date and time of a user, and a movement end position that is a destination to which the user moves from the position indicated by the position information log. A movement end location storage means for storing the data in association with each other, a stay point set generating means for generating a stay point set that is a set of position information logs measured at a stay point where the user frequently stays, and a stay point Using the stay point set generated by the set generation means, the movement end location determining means for determining the stay location as the movement end location of each position information log, and the stay location determined by the movement end location determining means as the movement end location When the movement end location registration means stored in the movement end location storage means in association with each position information log and the user's new position information log are input, the new position A destination prediction unit that extracts a stay location corresponding to a position information log having a predetermined close relationship with the information log from the movement end location storage unit and predicts the extracted stay location as a destination where the user moves. It is characterized by having.

  The movement destination prediction method according to the present invention is a position information log measured at a stay point where the user frequently stays from a set of position information logs that are information including the positioning position and positioning date and time of the user. A stay point set that is a set of the following is generated, the stay point set that is generated is used, the stay place that is the movement end location of each position information log is determined, and each position is determined with the determined stay location as the move end location. Corresponding to the information log and stored in the movement end location storage means, when a new position information log of the user is input, it corresponds to the position information log having a predetermined relationship with the new position information log The staying place to be extracted is extracted from the movement end place storage means, and the extracted staying place is predicted as the destination to which the user moves.

  The movement destination prediction program according to the present invention is measured by a computer from a set of position information logs, which are information including a user's positioning position and positioning date and time, at a stay point where the user frequently stays. Using the stay point set generation process that generates a stay point set that is a set of location information logs, and the stay point set generated by the stay point set generation process, the stay location that is the end of movement of each location information log is determined. A movement end location determination process, a stay location determined in the movement end location determination process as a movement end location, associated with each position information log, and stored in the movement end location storage means; and When a new location information log is input, the stay location corresponding to the location information log having a predetermined close relationship with the new location information log is stored as a movement end location storage unit. Extracted from the extracted residence location the user, characterized in that to execute the destination predicting process of predicting a destination of movement.

  According to the present invention, it is possible to predict the user's travel destination even when the positioning interval of the positional information is long and the positioning of the positional information is irregular.

It is a block diagram which shows one Embodiment of the movement destination prediction apparatus by this invention. It is explanatory drawing which shows the example of a positional infomation log. It is explanatory drawing which shows the example of the action information which a positional information log memory | storage part memorize | stores. It is explanatory drawing which shows the example of the information which a destination information storage part memorize | stores. It is a flowchart which shows the operation example of the movement destination prediction apparatus of this embodiment. It is a flowchart which shows the example of the process which memorize | stores a movement end position. It is explanatory drawing which shows the example of the operation | movement which estimates the destination from the positional infomation log of a present location. It is explanatory drawing which shows the example of a positional infomation log. It is explanatory drawing which shows the example of the result by which action pattern analysis was performed. It is explanatory drawing which shows the other example of the action information memorize | stored in the positional information log memory | storage part. It is explanatory drawing which shows the example of the result of having judged the destination with respect to a positional information log. It is explanatory drawing which shows the example of the information memorize | stored in the destination information storage part. It is explanatory drawing which shows the example of the positional infomation log when a present location log is input. It is a block diagram which shows the example of the minimum structure of the movement destination prediction apparatus by this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, a life log service is assumed in which position information and acquisition time of position information are recorded using sensor information such as GPS mounted on a mobile terminal.

  FIG. 1 is a block diagram showing an embodiment of a moving destination prediction apparatus according to the present invention. The movement destination prediction apparatus in the present embodiment includes a position information log input unit 110, a position information temporary storage unit 120, a behavior pattern analysis unit 130, a behavior information storage unit 140, a behavior information registration unit 150, a destination And a prediction unit 160.

  The position information log input unit 110 stores the position information log in the position information temporary storage unit 120 when a user position information log acquired by a mobile phone or the like equipped with a GPS function is input. When predicting a destination where the user moves from the position indicated by the position information log, the position information log input unit 110 inputs the input position information log to the destination prediction unit 160.

  The position information log is information including numerical information indicating the position of the terminal that has acquired a position such as latitude, longitude, and altitude, and date / time information indicating the date and time when the numerical information (position information) is acquired. In other words, the position information log is information including the positioning position and positioning date and time of the terminal user. The location information log includes the identifier of the terminal that acquired the location information, the identifier of the terminal user (hereinafter referred to as the user identifier), and other information that can be acquired when the location information is acquired. May be.

  FIG. 2 is an explanatory diagram illustrating an example of a position information log. In the example illustrated in FIG. 2, it is indicated that the position information log includes a user identifier, date / time, latitude, and longitude.

  In the present embodiment, a case where the position information log input unit 110 first stores the position information log in the position information temporary storage unit 120 will be described. However, the location information log input unit 110 may transmit the location information log directly to the behavior pattern analysis unit 130 without storing the location information log in the location information temporary storage unit 120.

  The position information temporary storage unit 120 stores a position information log. When the position information log stored in the position information temporary storage unit 120 is accumulated to some extent, the position information log is input to the behavior pattern analysis unit 130.

  The action pattern analysis unit 130 uses the position information log acquired from the position information temporary storage unit 120 and the position information log that has already been subjected to the action pattern analysis to determine whether each position information log indicates “staying”. Is determined. Here, the location information log indicating “staying” means that the location where the location information log is located is a location where the user frequently goes or stays. The location information log indicating “move” means a location other than the location of the location information log indicating “stay”. When the position information log indicates “staying”, the location where the position information log is measured is referred to as a staying point.

  Specifically, the behavior pattern analysis unit 130 generates a stay point cluster that is a set of position information logs indicating “stay”, so that each position information log indicates “stay” or “movement”. Is determined.

  The behavior pattern analysis unit 130 may generate a stay point cluster using, for example, the method described in Non-Patent Document 1. Specifically, the behavior pattern analysis unit 130 plots the position information log on a three-dimensional space defined by latitude, longitude, and time. The behavior pattern analysis unit 130 easily determines that the Euclidean distance in the time direction is close to a plane (latitude / longitude plane) defined by latitude and longitude in the three-dimensional space (that is, the Euclidean distance becomes dense). This is done for the location information log.

  Further, the behavior pattern analysis unit 130 performs clustering of weighted position information logs. Then, the behavior pattern analysis unit 130 determines a cluster having a large number of location information logs included as a stay point cluster. Then, the behavior pattern analysis unit 130 determines that the position information log included in the stay point cluster is “stay”. For example, as described in Non-Patent Document 1, the behavior pattern analysis unit 130 refers to a cluster in which the number of clusters included in a stay point area indicating a higher range of a predetermined normal distribution is a stay point cluster. You may decide.

  When specifying the stay point, a set of position information logs having a close Euclidean distance in the space composed of three variables of latitude, longitude, and time is, for example, general-purpose clustering such as EM (Expectation Maximization) algorithm or hierarchical clustering. Generated using a technique.

  Then, the behavior pattern analysis unit 130 sets a stay point cluster including a position information log indicating “stay” as a stay point set. In addition, the behavior pattern analysis unit 130 determines whether the position information log and each position information log indicates “stay” or “move” (hereinafter referred to as a log determination result). The data is stored in the storage unit 140. In this way, the behavior pattern analysis unit 130 generates a stay point set that is a set of position information logs measured at a position where the user frequently stays (stay point).

  The behavior information storage unit 140 includes a position information log storage unit 141 and a destination information storage unit 142.

  The location information log storage unit 141 stores a location information log, an identifier for identifying the location information log (hereinafter referred to as a location information log identifier), and a log determination result by the behavior pattern analysis unit 130. Further, when the position information log indicates “staying”, the position information log storage unit 141 stores an identifier for identifying a staying point set including the position information log (hereinafter referred to as a staying point set identifier). Hereinafter, these pieces of information stored in the position information log storage unit 141 are referred to as behavior information.

  FIG. 3 is an explanatory diagram illustrating an example of behavior information stored in the position information log storage unit 141. In the example illustrated in FIG. 3, the user identifier of the user who is the positioning target of the location information log identified by the location information log identifier in the location information log storage unit 141, its positioning date, latitude, route, and log determination result , And the identifier of the staying point set including the position information log is stored in association with each other. The stay point set identifier is set when the log determination result is “stay”.

  The destination information storage unit 142 stores a position information log and a movement end location (also referred to as a movement end position) that is a destination to which the user moves from the position indicated by the position information log in association with each other. To do. In the present embodiment, when a location information log of a user is input, the behavior information registration unit 150 described later predicts the user's destination from the locations indicated by the stay point set. That is, the movement end location is selected from the locations indicated by the stay point set. Therefore, it can be said that the destination information storage unit 142 stores the position information log and the location indicated by the stay point set in association with each other. In the following description, the place indicated by the stay point set may be referred to as a stay place. Further, since the staying location stored in the destination information storage unit 142 indicates the destination, the staying point set identifier may be referred to as a destination identifier.

  The place indicated by the stay point set indicates a place derived from the stay point included in the stay point set. For example, an arbitrary place included in a certain area surrounding the entire stay point included in the stay point set may be set as the place indicated by the stay point set. However, the place indicated by the stay point set is not limited to the above contents. For example, the area formed by connecting the stay points included in the stay point set and the place included in the vicinity of the area may be set as the place indicated by the stay point set.

  The destination information storage unit 142 may store a location information log identifier instead of the location information log, and may store a stay point set identifier instead of the stay location. That is, the destination information storage unit 142 may store the location information log identifier and the stay point set identifier in association with each other.

  FIG. 4 is an explanatory diagram illustrating an example of information stored in the destination information storage unit 142. The example illustrated in FIG. 4 indicates that the destination information storage unit 142 stores a location information log identifier and a destination identifier in association with each other.

  The behavior information registration unit 150 includes a movement end position determination unit 151 and a movement end position registration unit 152.

  The movement end position determination unit 151 extracts the movement end position for each position information log using the action information stored in the action information storage unit 140 in chronological order. When extracting the movement end position, one stay point set and one or more action information are used.

  First, the movement end position determination unit 151 selects one position information log from the position information logs included in the stay point set. Then, the movement end position determining unit 151 traces the action information (specifically, the date and time when the position information log is measured) of the selected one position information log back to the nearest position information log (hereinafter, the proximity position log). Select Information Log.). Further, when the time difference between the selected one position information log and the proximity position information log is equal to or less than a predetermined threshold, the movement end position determination unit 151 determines the movement end position of the proximity position information log as one It is determined as a place (staying place) indicated by the staying point set including the position information log.

  On the other hand, when the time difference between the selected one position information log and the proximity position information log exceeds a predetermined threshold value, the movement end position determination unit 151 determines that there is no relationship between the position information logs. Specifically, the movement end position determination unit 151 determines the positioning time between the action information indicated by the position information log measured at a certain time and the action information indicated by the position information log measured at the next time. When the difference exceeds the threshold, it is determined that there is no relationship between the behavior information. In this case, the movement end position determination unit 151 does not set the movement end position in the action information of the position information log measured at the next time.

  When the proximity position information log indicates “move”, the movement end position determination unit 151 further performs the above determination between the action information of the proximity position information log and the closest position information log in time series. repeat. As described above, the movement end position determination unit 151 extracts the movement end position by using the previous action information retroactive to the time from the action information included in the stay point set.

  The movement end position registration unit 152 stores the stay point set identifier indicating the movement end position determined by the movement end position determination unit 151 in the destination information storage unit 142 in association with the position information log. The movement end position registration unit 152 may store the movement end position (that is, the staying location) in the destination information storage unit 142 in place of the staying point set identifier or together with the staying point set identifier. Further, the movement end position registration unit 152 may store the position information log identifier in the destination information storage unit 142 together with the position information log instead of the position information log. Note that the movement end position stored in the destination information storage unit 142 is used as a destination predicted by a destination prediction unit 160 described later.

  Note that the location information temporary storage unit 120 does not have to be used when the behavior pattern analysis is performed each time the location information log is input. Then, the position information log input unit 110 may transmit the position information log directly to the behavior pattern analysis unit 130. In this case, the action information registration unit 150 performs batch processing.

  When the location information log of the current location of the user is acquired, the destination prediction unit 160 performs destination prediction using the information indicated by the location information log and the information acquired from the behavior information storage unit 140. Specifically, the destination prediction unit 160 extracts a location information log that is close to the location information log of the current location (hereinafter referred to as the current location log) from the location information log storage unit 141. The destination prediction unit 160 extracts the movement end position corresponding to the extracted position information log from the destination information storage unit 142. Then, the destination prediction unit 160 predicts the destination from the extracted movement end position.

  As described above, when a new location information log (that is, current location log) of the user is input, the destination prediction unit 160 corresponds to the location information log having a predetermined close relationship with the current location log. The staying place is predicted as the destination where the user moves. Specifically, when the current location log is input, the destination prediction unit 160 selects a location information log that is close to the current location log, such as the current location log and the time, among the location information logs stored in the location information log storage unit 141. Extract. Then, the destination prediction unit 160 extracts a staying place corresponding to the extracted position information log from the destination information storage unit 142, and predicts the extracted staying place as a destination to which the user moves.

  When the current location log is input, the destination prediction unit 160 may extract a location information log that is closest to the location location of the current location log from among the location information logs stored in the location information log storage unit 141, for example. Good. In addition, the destination prediction unit 160 may extract a position information log whose distance from the positioning position of the current position log is within a predetermined threshold from the position information logs stored in the position information log storage unit 141. In addition, the destination prediction unit 160 has a distance from the positioning position of the current location log within a predetermined threshold among the location information logs stored in the location information log storage unit 141 and the positioning time of the current location log. A position information log having a difference within a predetermined threshold may be extracted.

  Further, the destination prediction unit 160 may determine in advance whether or not the current location log indicates “move” when the current location log is input. When the current location log indicates “move”, the destination prediction unit 160 may extract a location information log close to the current location log from the location information log storage unit 141. The method for determining whether or not the current location log indicates “move” may be the same as the method in which the behavior pattern analysis unit 130 determines “stay” or “move” for each position information log. The method for predicting the destination is the same as the method described above.

  Further, when it is determined that the current location log indicates “staying”, the destination prediction unit 160 may determine whether or not there is a possibility that the user moves from the position where the current location log is located. For example, the destination prediction unit 160 extracts a position information log having a time close to the current position log from the position information log storage unit 141. When the extracted location information log indicates “staying” and the destination is set in the location information log, the destination prediction unit 160 moves from the position where the user has positioned the current location log. It may be determined that there is a possibility. Note that the method of predicting the destination using the location information log when there is a possibility of moving is the same as the method performed when it is determined that the current location log indicates “move”.

  Thus, by specifying whether the current location log that predicts the destination indicates “move” or “stay” in advance, the accuracy of predicting the destination can be increased.

  The position information temporary storage unit 120 and the behavior information storage unit 140 (more specifically, the position information log storage unit 141 and the destination information storage unit 142) are realized by, for example, a magnetic disk.

  In addition, the behavior pattern analysis unit 130, the behavior information registration unit 150 (more specifically, the movement end position determination unit 151, the movement end position registration unit 152), and the destination prediction unit 160 include a program (movement destination prediction). This is realized by a CPU of a computer that operates according to a program. For example, the program is stored in a storage unit (not shown) of the movement destination prediction apparatus, and the CPU reads the program, and according to the program, the behavior pattern analysis unit 130 and the behavior information registration unit 150 (more specifically, the movement The end position determination unit 151, the movement end position registration unit 152), and the destination prediction unit 160 may operate. Also, the behavior pattern analysis unit 130, the behavior information registration unit 150 (more specifically, the movement end position determination unit 151 and the movement end position registration unit 152), and the destination prediction unit 160 each have dedicated hardware. It may be realized with.

  Next, the operation of the movement destination prediction apparatus of this embodiment will be described. FIG. 5 is a flowchart showing an example of the entire operation of the movement destination prediction apparatus of the present embodiment.

  When the position information log is input by the position information log input unit 110, the behavior pattern analysis unit 130 performs a behavior pattern analysis on the input position information log, and whether the position information log indicates “staying” or “ It is determined whether or not “movement” is indicated (step S510). The determination result is stored in the position information log storage unit 141. After analyzing the discrimination behavior pattern, the behavior information registration unit 150 uses the location information log (stay point) indicating “stay” determined by the behavior pattern analysis unit 130 to move to the location indicated by the latitude and longitude of the location information log. An end position (that is, a destination) is extracted (step S511).

  After the movement end position is extracted and stored in the destination information storage unit 142, when the current location position information log (that is, the current location log) is input, the destination prediction unit 160 includes the current location log and the behavior information. The destination is predicted using the information stored in the storage unit 140 (step S512).

  FIG. 6 is a flowchart illustrating an example of processing for storing the movement end position in the destination information storage unit 142. Hereinafter, with reference to the flowchart illustrated in FIG. 6, the process of step S510 for performing the action pattern analysis and storing the log determination result in the position information log storage unit 141 and the process of step S511 for extracting the movement end position will be described. explain.

  When the behavior pattern analysis is completed, the behavior pattern analysis unit 130 stores the positional information log and the log determination result determined for each positional information log in the positional information log storage unit 141 (step S610).

  After all the log determination results are registered in the position information log storage unit 141, the behavior information registration unit 150 extracts the movement end position. The movement end position determination unit 151 first selects a position information log included in the stay point set stored in the position information log storage unit 141. Hereinafter, this position information log is referred to as a position information log 1. The movement end position determination unit 151 uses the position information log 1 as a base point, and extracts the position information log at the nearest time when going back in time. Hereinafter, this position information log is referred to as a position information log 2. The movement end position determination unit 151 determines whether or not the time difference between the position information log 1 and the position information log 2 is within a preset threshold (hereinafter referred to as a first threshold) (step S611). When the time difference between the two is within the first threshold (Y in step S611), the movement end position determination unit 151 determines that the movement end position of the position information log 2 is a stay point set including the position information log 1.

  In this case, the movement end position determination unit 151 checks whether or not the position information log 2 indicates “movement” (step S612). When the position information log 2 indicates “movement” (Y in step S612), the movement end position determination unit 151 uses the stay point set indicating the movement end position as a destination corresponding to the position information log 2 as a destination information storage unit. 142 (step S613). For example, information including the latitude and longitude of the stay point is registered in the destination information storage unit 142.

  Further, the movement end position determination unit 151 extracts the position information log of the closest time when the time is traced back, using the position information log 2 as a base point. Hereinafter, this position information log is referred to as a position information log 3. The movement end position determination unit 151 determines whether or not the time difference between the position information log 2 and the position information log 3 is within a set threshold (that is, the first threshold) (step S614). If the time difference between the two is within the first threshold (Y in step S614), the movement end position determination unit 151 determines that the movement end position of the position information log 3 is a stay point set including the position information log 1. Hereinafter, the processing after step S612 is repeated with the position information log 3 as a base point.

  On the other hand, when the position information log 2 does not indicate “move” (ie, indicates “stay”) in step S612 (ie, “N” in step S612), the movement end position determination unit 151 sets the stay point set of the position information log 1 It is confirmed whether or not the identifier and the stay point set identifier of the position information log 2 are different (step S615). If the two stay point set identifiers are different (Y in step S615), the movement end position determination unit 151 sets the stay point set indicating the movement end position in the destination information storage unit 142 as a destination corresponding to the position information log 2. Store (step S616).

  After the movement end position is stored in the destination information storage unit 142, the movement end position determination unit 151 changes the stay point as the base point to the next stay point (step S617). The residence point to be changed is arbitrary. For example, the movement end position determination unit 151 may select a stay point included in the same stay point set or may select a stay point included in another stay point. Thereafter, the process after step S611 for extracting the movement end position is repeated using the changed stay point as a base point.

  Next, when the time difference between the position information log 1 and the position information log 2 is larger than the first threshold value in step S611 (N in step S611), or in step S615, the stay point set identifier of the position information log 1 and When the staying point set identifiers of the location information log 2 are the same (N in step S615), processing performed thereafter will be described. In this case, the movement end position determination unit 151 changes the stay point as the base point to the next stay point without updating the information indicating the movement end position (destination information) in the position information log 2 (step S618). . Thereafter, the process after step S611 for extracting the movement end position is repeated using the changed stay point as a base point.

  In the above description, the case where the first threshold value is set to indicate a date / time difference has been described. However, the value set as the first threshold is not limited to a value indicating a date / time difference. For example, a value indicating a difference in distance may be set as the first threshold.

  Next, the operation when the location information log of the current location (that is, the current location log) is input will be described. As illustrated in FIG. 4, when the location information log of the current location of the user is input by the location information log input unit 110, the destination prediction unit 160 performs destination prediction (step S512 in FIG. 5). Hereinafter, the process of step S512 will be described in detail. FIG. 7 is an explanatory diagram illustrating an example of an operation for predicting the destination from the location information log (current location log) of the current location.

  First, the destination prediction unit 160 compares the input current location log with the information stored in the location information log storage unit 141 (step S710). Then, the destination predicting unit 160 checks whether or not the current location log indicates “move” (step S711). When the current location log indicates “move” (Y in step S711), the destination predicting unit 160 extracts a location information log close to the current location log from the location information log storage unit 141 (step S712). At this time, the destination prediction unit 160 specifies a position information log close to the current location log using a predetermined threshold (hereinafter referred to as a second threshold). Note that the second threshold value is different from the first threshold value described above.

  For example, the destination prediction unit 160 extracts a position information log whose distance from the positioning location of the current location log is a second threshold value or less from the position information log stored in the position information log storage unit 141. Then, the destination prediction unit 160 predicts the destination using the extracted position information log (step S713).

  On the other hand, in step S711, when the current location log does not indicate “move” (that is, indicates “stay”) (N in step S711), the destination prediction unit 160 determines the positioning time and the like included in the current location log. It is confirmed from the information whether or not the user is likely to move (step S714). When there is a possibility of movement (Y in step S714), the destination predicting unit 160 predicts the destination using the position information log (step S715). On the other hand, when there is no possibility of moving (N in step S714), the destination predicting unit 160 does not predict the destination (step S716).

  As described above, according to the present embodiment, the behavior pattern analysis unit 130 receives a set of position information logs (for example, the position information log stored in the position information temporary storage unit 120 or the position information log input unit 110). A stay point set is generated from a position information log directly input). The movement end position determination unit 151 uses the generated stay point set to determine the movement end position (stay position) of each position information log. Then, the movement end position determination unit 151 stores the determined stay location in the destination information storage unit 142 in association with each position information log as the movement end location.

  Further, when the user's current location log is input, the destination prediction unit 160 displays the stay location corresponding to the location information log close to the current location log (for example, a small difference in distance, time, etc.) as the destination information. Extracted from the storage unit 142. Then, the destination predicting unit 160 predicts the extracted staying place as a destination where the user moves. With such a configuration, it is possible to predict the user's travel destination even when the positioning interval of the positional information is long and the positioning of the positional information is irregular.

  Hereinafter, the present invention will be described with reference to specific examples, but the scope of the present invention is not limited to the contents described below. In this embodiment, a case where a user who owns a mobile terminal uses an LBS service will be described as an example. Here, the portable terminal is a terminal that can measure position information. In this service, the user uses this mobile terminal to position information. Then, the user uses this portable terminal to send information related to the service such as comments and photos about the location corresponding to the location information to the server that provides the LBS service, along with the location information obtained by positioning. Send. This server corresponds to the movement destination prediction apparatus in the above embodiment.

  FIG. 8 is an explanatory diagram illustrating an example of a location information log used in the present embodiment. As illustrated in FIG. 8, in this embodiment, it is assumed that the user identifier, the date and time (positioning date and time), and the latitude and longitude as the positioning position are included in the position information log.

  The behavior pattern analysis unit 130 of the server performs behavior pattern analysis after acquiring the sent location information log. FIG. 9 is an explanatory diagram illustrating an example of a result of the behavior pattern analysis performed on the position information log. Specifically, FIG. 9 shows an example of a result of dividing whether the position information log indicates “staying” or “movement” by analyzing the user's position information log by behavior pattern analysis. FIG. 10 is an explanatory diagram showing another example of behavior information stored in the position information log storage unit 141. The position information log storage unit 141 may store action information in the format illustrated in FIG.

  In the present embodiment, the location information log identifier of the location information log indicating “staying” is represented by using curly brackets “[]”. In addition, the position information log identifier of the position information log indicating “move” is expressed using parentheses “()”. In the example illustrated in FIG. 10, the position information logs [2] and [3] indicate the position information log identifiers of the position information logs indicating “stay”, and the position information logs (1) and (2) indicate “move”. The position information log identifier of the position information log indicating

  The movement end position determination unit 151 extracts a movement end position by using a stay point set for each position information log. First, the movement end position discriminating unit 151 uses the position information log [3] included in the stay point set as a base point, and the difference in date and time is a threshold value with respect to the position information log (2) whose time is closest. Check whether it is within or not.

  Here, when the threshold is 3 hours, the positioning time (date and time) of the location information log identified by the location information log [3] and the positioning time (date and time) of the location information log identified by the location information log (2) ) Is within 3 hours. Therefore, the requirement is satisfied. Therefore, the movement end position determination unit 151 determines the movement end position (destination) of the position information log (2) as the stay point set S1 including the position information log [3] (see FIG. 9).

  The movement information log identified by the position information log (2) indicates “movement”. Therefore, the movement end position discriminating unit 151 checks whether or not the difference in date and time is within a threshold with respect to the position information log (1) whose time is the closest by using the position information log (2) as a base point. To do.

  The difference between the positioning time (date / time) of the location information log identified by the location information log (2) and the positioning time (date / time) of the location information log identified by the location information log (1) is within 3 hours. . Therefore, the requirement is satisfied. Therefore, the movement end position determination unit 151 determines that the movement end position (destination) of the position information log (1) is also the stay point set S1 including the position information log [3].

  The movement information log identified by the position information log (1) similarly indicates “movement”. Therefore, the movement end position discriminating unit 151 confirms whether or not the difference in date and time is within a threshold with respect to the position information log [2] whose time is closest by using the position information log (1) as a base point. To do.

  The difference between the positioning time (date / time) of the location information log identified by the location information log (1) and the positioning time (date / time) of the location information log identified by the location information log [2] is within 3 hours. . Therefore, the requirement is satisfied. Therefore, the movement end position determination unit 151 determines that the movement end position (destination) of the position information log [2] is also the stay point set S1 including the position information log [3].

  Here, the movement information log identified by the position information log [2] indicates “staying”. For this reason, the movement end position determination unit 151 stops the process of determining the difference in date and time by going back in time. The movement end position determination unit 151 extracts the movement end position using the next stay point.

  For example, it is assumed that the position information log identified by the position information log [4] shown in FIG. 9 is selected as the next staying point. In the example illustrated in FIG. 9, the position information log having the closest point in time from the position information log [4] is the position information log identified by the position information log [3]. The difference between the positioning time (date and time) of the location information log identified by the location information log [4] and the positioning time (date and time) of the location information log identified by the location information log [3] is 3 hours or more. . Therefore, the movement end position determination unit 151 does not determine the movement end position (destination) of the position information log [4]. That is, the movement end position determination unit 151 does not store the movement end position (destination) of the position information log [3] in the destination information storage unit 142.

  Thereafter, the movement end position determination unit 151 sequentially selects the next staying point, and repeats the above processing for all the staying points. FIG. 11 is an explanatory diagram illustrating an example of a result of determining a destination for all the position information logs illustrated in FIG. 9. 12 is an explanatory diagram illustrating an example of information stored in the destination information storage unit 142 as a result of determining the destination with respect to the position information log illustrated in FIG.

  Next, the operation for predicting the destination will be described. Here, it is assumed that the destination information illustrated in FIG. 12 is already stored in the destination information storage unit 142. When the current location log is input in such a state, the destination predicting unit 160 extracts location information in the location information log storage unit 141 using a predetermined threshold value, and predicts the destination. This threshold corresponds to the second threshold in the above embodiment. Note that the threshold value used here is different from the threshold value used when the movement end position determination unit 151 goes back in time and compares the date and time differences.

  FIG. 13 is an explanatory diagram illustrating an example of a location information log when a current location log is input. Hereinafter, a method in which the destination prediction unit 160 predicts the destination for each content of the second threshold used when predicting the destination will be described.

  First, the case where only the difference in distance is used as the second threshold will be described. In this embodiment, it is assumed that the position information log whose distance from the current location log is within the threshold is the position information log identified by the position information log (1) and the position information log (3) (see FIG. 13). In this case, the destination prediction unit 160 uses the two position information logs to perform destination prediction from the positioning position indicated by the current location log. As described above, there are two location information logs measured at a distance within the threshold. However, the destinations indicated by these two points are different. Therefore, the destination candidates are the stay point set S1 or the stay point set S3.

  Since there are two destinations, the destination prediction unit 160 may predict the probability of going to each destination as 0.5. In addition, the destination prediction unit 160 calculates the distance between the position information log (1) and the position information log (3) from the positioning position indicated by the current position log, and sets each destination according to the calculated ratio of the distances. The probability of going to may be predicted. However, the method by which the destination prediction unit 160 calculates the probability is not limited to the above method. The destination prediction unit 160 may calculate the probability higher as the distance between the position information log corresponding to each extracted stay location and the positioning position of the current location log is closer, and predict the destination based on the calculated probability.

  Next, a case where a difference in distance and a difference in time are used as the second threshold will be described. Also here, it is assumed that the position information log whose distance from the current location log is within the threshold is the position information log identified by the position information log (1) and the position information log (3) (see FIG. 13). In this case, the destination prediction unit 160 uses the two position information logs to perform destination prediction from the positioning position indicated by the current location log. As described above, there are two location information logs measured at a distance within the threshold. However, the destinations indicated by these two points are different. Therefore, the destination candidates are the stay point set S1 or the stay point set S3.

  Here, it is assumed that the time difference between the positioning date / time of the position information log (3) and the positioning date / time of the current location log is within the threshold. In this case, in this case, the destination prediction unit 160 sets the stay point set S3 of the position information log (3) as the destination.

  Heretofore, the method for predicting the destination by the destination prediction unit 160 has been described. However, the method by which the destination prediction unit 160 predicts the destination is not limited to the above method. For example, when there are two destination candidates, the stay point set S1 and the stay point set S3, the destination predicting unit 160 determines the respective stay information according to the time difference and distance between each position information log and the current location log. There may be a difference in the probability of going to the point set. That is, the destination predicting unit 160 calculates a higher probability as the distance between the position information log corresponding to each extracted stay location and the current position log is closer, and further, the position information corresponding to each extracted stay location. The probability may be calculated higher as the difference between the positioning date and time of the log and the current location log is smaller. Then, the destination prediction unit 160 may predict the destination based on the calculated probability.

  Further, the destination prediction unit 160 may predict the destination indicated by the position information log measured at the position closest to the position where the current location log is positioned as the destination from the current location. Further, the destination prediction unit 160 performs weighting according to the time difference and distance between each location information log and the current location log, and determines the destination indicated by the closest location information log determined according to the weight as the current location. It may be a destination from

  By using the above method, it is possible to predict the user's moving destination without periodically positioning the position information. For example, by using the above method, when a location information log is notified while the user is moving on a train, it is possible to predict where to go from the location indicated by the location information. In this case, information regarding the destination such as a restaurant or an event can be distributed to the user in advance.

  Next, an example of the minimum configuration of the present invention will be described. FIG. 14 is a block diagram showing an example of the minimum configuration of the moving destination prediction apparatus according to the present invention. The movement destination prediction apparatus according to the present invention includes a position information log that is information including a positioning position and positioning date and time of a user, and a movement end position that is a destination to which the user moves from the position indicated by the position information log. The movement end location storage means 81 (for example, the destination information storage unit 142) that stores the information in association with each other (for example, the location information log identifier and the destination identifier in association with each other) and the location where the user frequently stays. The stay point set generation means 82 (for example, the behavior pattern analysis unit 130) that generates a stay point set that is a set of position information logs measured at the stay points, and the stay point set generated by the stay point set generation means 82 are used. Then, a movement end location determination unit 83 (for example, a movement end location determination unit 151) that determines a stay location that is a movement end location of each position information log and a movement end location determination unit 83 A movement end location registration unit 84 (for example, a movement end location registration unit 152) that stores the cut-off residence location as a movement end location in association with each position information log and is stored in the movement end location storage unit 81; When a location information log is inputted, a stay location corresponding to the location information log having a predetermined close relationship with the new location information log is extracted from the movement end location storage means 81, and the extracted stay location is Destination prediction means 85 (for example, destination prediction unit 160) for predicting the destination to which the user moves is provided.

  With such a configuration, it is possible to predict the user's travel destination even when the positioning interval of the positional information is long and the positioning of the positional information is irregular.

  Further, the destination predicting means 85 extracts the staying place corresponding to the position information log indicating the positioning position closest to the positioning position of the new position information log from the movement end place storage means 81, and uses the extracted staying place as the user. May be predicted as the destination to move.

  Further, the destination predicting means 85 extracts the staying place corresponding to the position information log whose distance from the positioning position of the new position information log is within a predetermined threshold from the movement end place storage means 81, and extracts each staying The destination to which the user moves may be predicted based on the probability calculated according to the positioning position of the position information log corresponding to the place and the distance between the positioning positions indicated by the new position information log.

  Further, the destination predicting means 85 extracts a staying place corresponding to the position information log whose distance and time difference from the positioning position of the new position information log are within a predetermined threshold from the movement end place storage means 81, Predicting the destination to which the user will move based on the distance calculated between the location information log corresponding to each extracted stay location and the new location information log, and the probability calculated according to the difference in positioning date and time Also good.

  Further, the movement end location determination means 83 selects one location information log from the location information logs included in the stay point set, and the date and time when the location information log was measured is traced back in time series to the nearest location. If the time difference from the proximity position information log, which is an information log, is equal to or less than a predetermined threshold, the staying location of the proximity location information log is determined as the location indicated by the stay point set including one location information log. May be. Further, the movement end location registration unit 84 associates the staying location that is the movement end location of the proximity position information log determined by the movement end location determination unit 83 with the proximity position information log in the movement end location storage unit 81. It may be memorized.

  In addition, the movement end location determination unit 83 measures the proximity position information log and the proximity position information log when the proximity position information log is not a staying point (for example, when the position information log indicates “movement”). Based on whether the time difference from the nearest location information log is back in time and is smaller than a predetermined threshold, the location information log includes the location where the proximity location information log ends. It may be determined that this is the place indicated by the stay point set.

  Although the present invention has been described with reference to the embodiments and examples, the present invention is not limited to the above embodiments and examples. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

  This application claims the priority on the basis of the JP Patent application 2011-127351 for which it applied on June 7, 2011, and takes in those the indications of all here.

  The present invention is preferably applied to a movement destination prediction apparatus that predicts a user's movement destination from position information measured irregularly.

DESCRIPTION OF SYMBOLS 110 Position information log input part 120 Position information temporary storage part 130 Action pattern analysis part 140 Action information storage part 141 Position information log storage part 142 Destination information storage part 150 Action information registration part 151 Movement end position discrimination | determination part 152 Movement end position registration Part 160 Destination Prediction Department

Claims (10)

A movement end location storage unit that associates and stores a position information log that is information including the positioning position and positioning date and time of the user, and a movement end location that is a destination to which the user moves from the position indicated by the position information log. When,
A stay point set generating means for generating a stay point set that is a set of position information logs measured at a stay point that is a position where a user frequently stays;
Using the stay point set generated by the stay point set generating means, a movement end location judging means for judging a stay location that is a movement end location of each position information log;
A movement end location registering unit that stores the stay location determined by the movement end location determination unit as a movement end location in association with each position information log and stored in the movement end location storage unit;
When a new location information log of the user is input, a stay location corresponding to the location information log having a predetermined close relationship with the new location information log is extracted from the movement end location storage means and extracted. A destination predicting apparatus comprising: a destination where a user moves and a destination predicting means for predicting the stayed place. The destination predicting means extracts the staying location corresponding to the position information log indicating the positioning position closest to the positioning position of the new position information log from the movement end location storage means, and the user moves the extracted staying location. The movement destination prediction apparatus according to claim 1, which predicts a place. The destination predicting means extracts a staying place corresponding to the position information log whose distance from the positioning position of the new position information log is within a predetermined threshold from the movement end place storage means, and corresponds to each extracted staying place. The movement destination prediction apparatus according to claim 1, wherein a destination to which a user moves is predicted based on a probability calculated according to a positioning position of a position information log and a distance between positioning positions indicated by the new position information log. . The destination predicting means extracts a stay location corresponding to the position information log whose distance and time difference from the positioning position of the new position information log are within a predetermined threshold from the movement end location storage means, and extracts each stay The destination to which a user moves is predicted based on the distance calculated between the position information log corresponding to the place and the new position information log, and the probability calculated according to the difference between the positioning dates and times. Mobile destination prediction device. The movement end location judging means selects one location information log from the location information logs included in the stay point set, and the date and time when the one location information log was measured is traced back in time series to the nearest location information log. When the time difference with a certain proximity position information log is equal to or less than a predetermined threshold, the staying location of the proximity location information log is determined as the location indicated by the stay point set including the one location information log,
The movement end location registering means stores the stay location, which is the movement end location of the proximity position information log determined by the movement end location determination means, in the movement end location storage means in association with the proximity position information log. The movement destination prediction apparatus according to any one of claims 1 to 4. When the proximity position information log is not the staying point, the movement end location determination means determines the time difference between the proximity position information log and the closest position information log retroactively to the time when the proximity position information log was measured. 6. The movement purpose according to claim 5, wherein, based on whether or not is smaller than a predetermined threshold, the movement end location of the proximity position information log is determined as a place indicated by a stay point set including one position information log. Land prediction device. From a set of location information logs that are information including the positioning position and positioning date and time of the user, a stay point set that is a set of the location information logs measured at the stay point where the user frequently stays is generated. And
Using the generated stay point set, determine the stay location that is the end location of each location information log,
The determined stay location is stored as a movement end location in the movement end location storage means in association with each position information log,
When a new location information log of the user is input, a staying location corresponding to the location information log having a predetermined close relationship with the new location information log is extracted from the movement end location storage means,
A moving destination prediction method characterized by predicting an extracted staying place as a destination to which a user moves. When a new location information log of the user is input, the staying location corresponding to the location information log indicating the positioning position closest to the positioning location of the new location information log is extracted from the movement end location storage means and extracted. The movement destination prediction method according to claim 7, wherein the staying place is predicted as a destination where the user moves. On the computer,
From a set of location information logs that are information including the positioning position and positioning date and time of the user, a stay point set that is a set of the location information logs measured at the stay point where the user frequently stays is generated. A dwell point set generation process to
Using the stay point set generated in the stay point set generation process, a movement end location determination process for determining a stay location that is a movement end location of each position information log;
A movement end location registration process in which the stay location determined in the movement end location determination process is stored as a movement end location in the movement end location storage unit in association with each position information log; and
When a new location information log of the user is input, a stay location corresponding to the location information log having a predetermined close relationship with the new location information log is extracted from the movement end location storage means and extracted. A destination prediction program for executing a destination prediction process that predicts a destination where a user stays as a destination. On the computer,
In the destination prediction process, the stay location corresponding to the position information log indicating the position closest to the position of the new position information log is extracted from the movement end location storage means, and the user moves the extracted stay location The movement destination prediction program according to claim 9, wherein the program predicts the location.

Images