JP6825880B2 - Behavior analysis system and its program using various types of location information - Google Patents

Description

The present invention relates to a system for behavior analysis using a large number of types of position information, and in particular, a large number of types of positions capable of improving the accuracy of behavior analysis by using a plurality of types of position information in association with a time series. Regarding behavioral analysis systems that use information and their programs.

[Conventional technology]
In the conventional behavior analysis system, there is usually one that analyzes the user's behavior by using coordinate data such as GPS (Global Positioning System).

[Related technology]
As related prior art documents, JP-A-2006-221329 "Behavior Predictor, Behavior Prediction Method and Behavior Prediction System" (Toshiba Co., Ltd.) [Patent Document 1], JP-A-2015-135589 "Information Analysis". "System and Information Analysis Method" (Hitachi Co., Ltd.) [Patent Document 2], Japanese Patent Application Laid-Open No. 2016-004336 "Marketing System and Marketing Method" (Willil Mobile Co., Ltd.) [Patent Document 3].

Patent Document 1 describes a behavior prediction device that estimates a behavior selection model of a moving body based on individual attributes, field attributes, and behavior history.
Further, Patent Document 2 describes an information analysis system that analyzes the characteristics of human behavior and the characteristics of an area by using a pattern of a human behavior range obtained from human movement data.
Further, in Patent Document 3, the server calculates the flow line moved by the user and the residence time in the area of the wireless device, estimates the action information performed by the user in the area, and the action information of the user and each device. It is described that the attribute information of the user is analyzed based on the installation attribute information of.

Japanese Unexamined Patent Publication No. 2006-221329 Japanese Unexamined Patent Publication No. 2015-135589 Japanese Unexamined Patent Publication No. 2016-004336

However, in the above-mentioned conventional behavior analysis system, the position information to be analyzed is coordinate data or data of a wireless area of a wireless communication device, and each of these data is analyzed as a unit, so that various types of positions are used. There is a problem that it is not possible to perform highly accurate behavior analysis by associating information with a time series.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a behavior analysis system and a program thereof for improving the accuracy of behavior analysis by using a plurality of types of position information in relation to each other in a time series. ..

<Behavior analysis system>
The present invention for solving the problems of the above-mentioned conventional example is a behavior analysis system having an analysis server that inputs behavior information including a user's position information and performs behavior analysis, and the analysis server stores the behavior information. It has a user database to be used and a spatial information conversion database that converts location information into corresponding facility information or area information, and the analysis server inputs location information with different granularity of location information. The information includes data at a position with a small granularity and data at an area with a large granularity, and the position information is sorted in time series for each user, and based on the sorted time series position information , the user's time series 2 If there is area data between two position data, and the area does not include a position on the line connecting the two position data that corresponds to the time of the area data, the user is in a time-series situation. It is characterized in that the data at a position having a small grain size in the area is estimated based on the time of the data in the area, and the behavior analysis is performed using the data at the estimated position.

According to the present invention, in the behavior analysis system, the analysis server converts data at a position having a small granularity into data in a large area including the position, and performs behavior analysis using the data in the converted area. It is a feature.

In the present invention, in the behavior analysis system, the analysis server inputs the data of the first area having a large grain size and the data of the second area having a smaller grain size than the data of the first area. Yes, the position data and the second area data can be converted into facility information or area information by the spatial information conversion database, and the first area data and the position data or / / based on the time-series position information. And the data of the new position in the first area or the data of the second area is estimated from the data of the second area, and the behavior analysis is performed using the data of the estimated position or the data of the second area. It is characterized by that.

In the above-mentioned behavior analysis system, the analysis server converts the position data into the data of the second area including the position, and the data of the second area estimated as the converted data of the second area. It is characterized by performing behavioral analysis using.

<Program>
The present invention is a processing program in an analysis server that inputs behavioral information including the user's location information and performs behavioral analysis. By executing the processing program on the analysis server, location information having different granularity is input. With respect to the position information including the input means to be input and the data of the position having a small grain size and the data of the area having a large grain size , the position information is rearranged in time series for each user, and based on the rearranged time series position information, the position information is sorted . If there is area data between two position data in the user's time series, and the area does not include a position that corresponds to the time of the area data on the line connecting the two position data. Assuming that the user has moved within the area according to the situation in the time series, the user acts using the estimation means for estimating the data at the position with a small grain size in the area based on the time of the data in the area and the data at the estimated position. It is characterized by functioning as an analysis means for performing analysis.

The present invention comprises a position information conversion means for converting data at a position having a small granularity into data in a large area including the position in the above program, and the analysis means analyzes behavior using the data in the converted area. It is characterized by performing.

In the above program, in the above program, the input means inputs the data of the first area having a large grain size and the data of the second area having a smaller grain size than the data of the first area. The analysis server is provided with a spatial information conversion means for converting the position data and the data of the second area into facility information or area information by the spatial information conversion database, and the estimation means is the first based on the time-series position information. The data of the new position or the data of the second area in the first area is estimated from the data of the area and the data of the position and / and the data of the second area, and the analysis means estimates the data of the estimated position. Alternatively, the behavioral analysis is performed using the data in the second area.

In the above program, in the above program, the position information conversion means converts the position data into the data of the second area including the position, and the analysis means estimates that the data is the data of the converted second area. It is characterized by performing behavioral analysis using the data of the area.

According to the present invention, the behavior analysis system has an analysis server that inputs behavior information including the user's location information and performs behavior analysis, and the analysis server corresponds to the user database that stores the behavior information and the location information. It has a spatial information conversion database that converts to facility information or regional information, and the analysis server inputs location information with different granularity for location information, and the location information is data for locations with small granularity. The position information is sorted in time series for each user, and the area data is placed between the two position data in the user's time series based on the sorted time series position information. If there is, and the area does not include a position corresponding to the time of the area data on the line connecting the two position data, it is assumed that the user has moved in the area according to the time series situation. , Since it is a behavior analysis system that estimates the data of the position with small grain size in the area based on the time of the data of the area and performs the behavior analysis using the data of the estimated position, different types of position information are time-series. Since behavioral analysis is performed by using it in association with, there is an effect that the accuracy of analysis can be improved.

It is a schematic block diagram of this system. It is a schematic diagram which shows the action history of a specific user. It is a figure which showed the example of the coordinate data provided from the providing server A in a table. It is a figure which showed the example of the mesh data provided from the providing server B in a table. It is a figure which showed the example of the administrative boundary data provided from the providing server C in a table. It is a figure which shows the management of the action log data. It is a figure which shows the action log data which was converted into the administrative world data. It is a figure which shows the action log data when a mesh area spans a plurality of administrative boundary areas. It is a figure of Example 1 which estimates the coordinate data from the administrative boundary data. It is a figure of Example 2 which estimates the coordinate data from the administrative boundary data. It is a flowchart which shows the combined use method in the control part 11 of the analysis server 1.

Embodiments of the present invention will be described with reference to the drawings.
[Outline of Embodiment]
The behavior analysis system (this system) according to the embodiment of the present invention includes user coordinate data, postal code area data, administrative boundary area data, mesh area data, communication area data of wireless devices, etc. from a plurality of action log providing servers. The analysis server acquires various types of behavior log information, organizes the data in chronological order, converts or estimates the data format of specific behavior log information, and performs behavior analysis. Since the behavioral analysis is performed by using the position information having different types of granularity in association with the time series, the accuracy of the analysis can be improved.

[This system: Fig. 1]
The behavioral analysis system (this system) according to the embodiment of the present invention will be described with reference to FIG. FIG. 1 is a schematic configuration diagram of this system.
In this system, as shown in FIG. 1, the analysis server 1, the user database (DB) 21 connected to the analysis server 1, the analysis database (DB) 22, the spatial information conversion database (DB) 23, and the analysis server 1 are connected. The network 3 and the action log providing servers 41 and 51 connected to the network 3 are provided.

[Each part of this system]
Each part of this system will be described in detail.
[Analysis server 1]
The analysis server 1 includes a control unit 11, a storage unit 12, and an interface unit 13.
The control unit 11 reads the processing program stored in the storage unit 12 and executes the processing described later while accessing each DB.
The storage unit 12 stores the processing program and the data required for the processing.
The interface unit 13 connects to the network 3 and also to each DB.

[User DB21]
The user DB 21 stores the user's action log data provided by the action log providing servers 41 and 51.
The action log data is a user ID (identification information), time information, and location information. The location information may indicate a specific point such as coordinate data, or may indicate a specific area such as a mesh code, an administrative boundary code, a zip code code, or an area information of a radio base station.

[Analysis DB22]
The analysis DB 22 stores data (result data of behavioral analysis) analyzed and processed by the control unit 11 of the analysis server 1.

[Spatial information conversion DB23]
The spatial information conversion DB 23 stores facility information such as parks, hospitals, shopping centers, and stadiums corresponding to coordinate points, mesh areas, and the like. Specifically, the range of the facility area is specified, attribute information such as the facility name and facility type is associated with the area, the coordinates of the action log data, the mesh area, and the location information of the radio base station area. Is included in the range of the facility area, and if it is included, the facility information is associated with the location information.
In this way, the facility information is converted and used in association with the coordinate data and the mesh code at the time of the action log analysis.

[Network 3]
The network 3 is usually assumed to be the Internet, but may be a dedicated network.

[Action log providing server 41]
The action log providing server 41 collects the action log data of a plurality of users, stores the action log data in the action log database (DB) 42, and transmits (provides) the action log data to the analysis server 1.
Further, the action log server 41 receives the analysis result from the analysis server 1.

[Action log providing server 51]
The action log providing server 51 collects the action log data of a plurality of users, stores the action log data in the action log database (DB) 52, and transmits (provides) the action log data to the analysis server 1.
Further, the action log server 51 receives the analysis result from the analysis server 1.

[Processing content of this system]
Next, the processing contents of this system will be specifically described.
[Outline of action log: Fig. 2]
An outline of a specific example of the process will be described with reference to FIG. FIG. 2 is a schematic diagram showing the behavior history of a specific user.
First, as a premise, there are three action log providing servers (simply referred to as "providing servers") A, B, and C. As an action history (action log), the providing server A has coordinate data and the providing server B has mesh data. (Mesh code data) is provided. The providing server C provides the administrative boundary (town chome) data (administrative boundary code data) to the analysis server 1. In FIG. 2, the user ID is common to "ID1". That is, it becomes an action log of the same person.

Here, if only the coordinate data is handled in the action log, it is recognized that the position of the point (1) → (2) → (3) has been moved, and if the action log is handled only in the mesh data, the mesh area [1] → [ 2] → It is recognized that the area of [3] has been moved, and if the action log is handled only by the administrative boundary data, it is recognized that the area of the administrative boundary area <1> → <2> has been moved.

However, since the analysis server 1 of this system manages the location information from the providing servers A, B, and C together with the time information, it can be recognized in time series regardless of the type of location information.
That is, the analysis server 1 is in chronological order: point (1) → mesh area [1] → administrative boundary area <1> → point (2) → mesh area [2] → administrative boundary area <2> → point (3) ) Move history (action history) is managed.

[Contents of action log data: Figures 3 to 5]
Next, the contents of the action log data in the example of FIG. 2 will be described with reference to FIGS. 3 to 5. FIG. 3 is a diagram showing an example of coordinate data provided by the providing server A in a table, and FIG. 4 is a diagram showing an example of mesh data provided by the providing server B in a table. FIG. Is a table showing an example of administrative boundary data provided by the providing server C.
The action log data from each providing server is output to the analysis server 1 and stored in the user DB 21 as shown in FIGS. 3 to 5.

Specifically, for example, for the user (user ID 1), the time data in which the data was acquired and the coordinate data of the acquired position information are stored in FIG. 3, and the time data and the mesh code data in FIG. 5 and FIG. Stores time data and administrative world code data.
The analysis server 1 stores the corresponding areas on the map for the mesh code and the administrative boundary code in advance in a table or the like, and the areas can be specified on the map by these codes.

[Management of action log data: Fig. 6]
When performing the behavior analysis process, the analysis server 1 rearranges the behavior log data of FIGS. 3 to 5 in chronological order and manages the data as shown in FIG. FIG. 6 is a diagram showing management of action log data.
In the example of FIG. 2, since three types of data, coordinate data, mesh data, and administrative boundary data, are included, a number or code indicating the type of data is assigned to distinguish the data.

As shown in FIG. 6, the analysis server 1 stores time data, data type number, and position data in the user DB 21 or the analysis DB 22 in association with the user ID.
As the position data, the coordinate data, mesh code, and administrative boundary code provided by each providing server are stored as they are.

[Three ways to use]
Next, three methods of using the action log data in this system will be described.
The three usage methods include, as a first usage method, a method of using each particle size, as a second usage method, a method of using according to a large particle size, and as a third usage method, using a small particle size. If it can be estimated, there is a method of using it with a small particle size.
Here, the particle size means a frequency corresponding to a wide and narrow area area, and the coordinates are the finest and smallest, and the larger the area area, the coarser and larger the particle size.

[First usage method: Use at each particle size]
The first usage method is a method of using each particle size.
The first usage method uses the particle size of the action log data obtained from the providing server as it is, and therefore lacks uniformity as the particle size, but has versatility in analyzing the action log data.

Specifically, as shown in FIG. 6, position data is managed according to a time series and used for behavior analysis.
The spatial information conversion DB 23 may be used to convert coordinate data and mesh code into facility information, and behavior analysis may be performed using the facility information.

[Second usage method: Use according to large particle size: Fig. 7]
The second usage method is a method of using according to a large particle size.
The second usage method does not pinpoint the history of the user's position information, but is useful when performing behavior analysis from the regional characteristics of the area where the user has moved. That is, it analyzes what kind of regional characteristic area the user has moved.

Specifically, assuming that the administrative boundary area is the widest area as shown in FIG. 2, as shown in FIG. 7, the coordinate data and mesh code having a small granularity are converted into the position data of the administrative boundary code. To use. FIG. 7 is a diagram showing action log data converted into administrative boundary data.
In FIG. 7, the coordinate data and mesh data in the action log data of FIG. 6 are converted into administrative boundary data.

In the case of coordinate data, the conversion method obtains the administrative boundary area including the position of the coordinates and converts it into the administrative boundary code of the administrative boundary. In the case of mesh data, the administrative boundary area including the mesh area is obtained and converted into the administrative boundary code of the administrative boundary.

[Processing when the mesh area spans multiple administrative areas: Fig. 8]
In the second usage method, when the mesh area spans a plurality of administrative boundary areas, the area occupied by the administrative boundary in the mesh area is obtained, the ratio is calculated by area proportional division, and weighting is performed according to the ratio. In that case, the user ID 1 is treated as if it existed in a plurality of administrative boundary areas at the same time according to the above ratio.

Specifically, when the mesh area straddles two administrative boundaries and the area ratios are 80% (0.8) and 20% (0.2), the weighting coefficients are "0.8" and "0.2". ". FIG. 8 shows an example in which the mesh area straddles the administrative boundary codes “G011” and “G012”, and weighting coefficients are assigned to them. The weighting coefficient when the area is not apportioned is "1". FIG. 8 is a diagram showing action log data when the mesh area spans a plurality of administrative boundary areas.

Further, the behavior analysis in the case of adjusting to a large particle size is performed by, for example, performing a cluster analysis of the administrative boundary area in advance and referring to the transition (movement history) of the cluster information of the area. The behavioral characteristics of the user are analyzed according to what kind of cluster area the user ID1 has moved.
The cluster analysis is described in detail in Japanese Patent No. 5123422, "Profiling System Utilizing Regional Characteristics" (Patent Owner: Giken Shoji International).

[Third usage: Use with small particle size]
The third usage method is a method of using with a small particle size.
The third usage method is useful when analyzing the user's behavior in a smaller area or coordinates when the behavior log data from the providing server is data in a wide area.
Since the coordinate data, mesh data, and administrative boundary data are acquired at short time intervals (high frequency), in reality, the respective data often overlap in position.

[Example of third usage method 1: FIG. 9]
For example, in the administrative boundary area of the action log data provided by the providing server C in a specific period, when the action log data of a plurality of coordinate data is provided from the providing server A, from the administrative boundary data (administrative boundary data). It is possible to estimate new coordinate data (using). The estimation method will be described with reference to FIG. FIG. 9 is a diagram of Example 1 in which coordinate data is estimated from administrative boundary data. FIG. 9 shows an example in which the coordinates of two points exist in the administrative boundary area.

As shown in FIG. 9, it is assumed that the providing server A provides the coordinate data (1) and the coordinate data (3) for the user ID 1, and the providing server C provides the administrative boundary data. The date and time of the coordinate data (1) is 11:00 on October 10, 2016, the date and time of the administrative world data is 11:30 on October 10, 2016, and the date and time of the coordinate data (3) is October 10, 2016. Suppose it is 12:00.

Since the coordinate data (1) and the coordinate data (3) are included in the relevant administrative boundary area, it is estimated that the user with user ID 1 was in the relevant administrative boundary area from 11:00 to 12:00 on October 10, 2016. Further, it is estimated that the position of the coordinate data (1) is being moved to the position of the coordinate data (3) during this time zone.

Here, since the user with user ID 1 was in the administrative boundary area at 11:30, the user with user ID 1 linearly moves from the position of the coordinate data (1) to the position of the coordinate data (3). It is presumed that they moved within the relevant administrative area instead of doing so.
Based on this, the position of the movement route at 11:30 (coordinate data [2]) is estimated with reference to the date and time of the coordinate data and the date and time of the administrative boundary data.

[Example 2: Third usage method: FIG. 10]
Even if either one or both of the coordinate data (1) and (3) is outside the range of the administrative boundary area of the action log data, new coordinate data can be estimated from the administrative boundary data by the same method as described above. .. For example, as shown in FIG. 10, the coordinate data (1) is outside the left side of the administrative boundary area, the coordinate data (3) is outside the lower side of the administrative boundary area, and the administrative boundary area is at 11:30. It is presumed that when the user is present in, the user does not move linearly from the position of the coordinate data (1) to the position of the coordinate data (3), but moves through the administrative boundary area. FIG. 10 is a diagram of Example 2 in which coordinate data is estimated from administrative boundary data. FIG. 10 shows an example in which the coordinates of two points exist outside the administrative boundary area.

In this way, new coordinate data can be estimated from the area data based on the acquisition time and the positional relationship of the area data and the coordinate data.
In the above example, the coordinate data is estimated from the administrative boundary data, but the coordinate data can be estimated from the mesh data by the same method.

It is also possible to estimate mesh data with a narrower range from wide-area administrative boundary data. In this case, the coordinate data is converted into the mesh data of the mesh area including the position, and the estimation is performed together with the original mesh data (data provided by the providing server B). The estimation method is the same as in the case of coordinate data, and the coordinate data is replaced with mesh data to estimate from the administrative boundary data to the mesh data.

When estimating the mesh data from the administrative boundary data, the cluster information of the administrative boundary area and the cluster information of the estimated mesh area may not match. In that case, for the estimated mesh area, the cluster information of the administrative boundary area may be used, or the behavior analysis may be performed using the cluster information of the estimated mesh area.
Further, the spatial information conversion DB 23 may be used to convert the coordinate data and the mesh code into the facility information, and the behavior analysis may be performed using the facility information.

[Use other areas]
In the above example, the coordinate data, mesh data, and administrative boundary data have been used for explanation, but the above three usage methods can be realized by using the action log data earned in the zip code area and the communication area of the wireless device. It may be.

[Use of statistical data]
In the behavior analysis, the behavior analysis is performed using the location of the area and the cluster information of the area, but the behavior analysis may be performed using the statistical information of the area. In particular, since various statistical data have been published for the administrative area, it is possible to perform highly accurate behavioral analysis by using these statistical data.

[Processing method on the analysis server]
Next, the function realization means realized by executing the processing program in the control unit 11 of the analysis server 1 will be described.
As the function realizing means, it has an input means, a position information conversion means, a spatial information conversion means, an estimation means, and an analysis means.

The input means inputs various types of action log data (behavior log data having different granularity) from the action log providing servers 41 and 51 and stores them in the user DB 21.
The position information conversion means converts data having a small granularity into data having a large granularity (data in a wide area). The conversion of the position information is performed as described in the second usage method.

The spatial information conversion means refers to the spatial information conversion DB 23 and converts the position data into the corresponding facility information or regional information.
The estimation means estimates the small-grain data in the area of the large-grain data based on the small-grain data using the large-grain data. The method of estimation is performed as described in the third usage method.

The analysis means analyzes the action log data input from the input means according to the first usage method by using the position data having different granularity as it is, and also according to the second usage method, the granularity is determined by the position information conversion means. Behavioral analysis is performed using position data obtained by converting small data into large-grain data, and behavioral analysis is performed using small-grain position data estimated from the large-grain data by an estimation means according to the third usage method. Do.

[Complex analysis processing: FIG. 11]
Next, the first to third usage methods have been described, but the combined usage methods will be described with reference to FIG. FIG. 11 is a flowchart showing a combined usage method in the control unit 11 of the analysis server 1.
Specifically, it is determined in the spatial information conversion DB 23 whether the position data of the input action log data can be converted into spatial information, and if it can be converted, the position data is used as it is for the behavior analysis and cannot be converted. If so, the position data having a small grain size until the position data can be converted into spatial information is estimated by an estimation means and used for behavioral analysis.

In the process of the control unit 11, as shown in FIG. 11, it is determined whether or not the particle size of the position information data of the action log data input by the input means is the smallest (S1), and the particle size is not the smallest. (In the case of No), the spatial information conversion means refers to the spatial information conversion DB 23 (S2), and determines whether or not the spatial information can be converted (S3).

If the spatial information conversion means cannot convert the data of the position information into the spatial information (in the case of No), the estimation means estimates the data having a small granularity (S4). The estimation process uses the third usage method.
Then, the process returns to process S1.

Further, when the particle size is the smallest in the determination process S1 (in the case of Yes) and when the determination process S3 can be converted into spatial information (in the case of Yes), the analysis means performs the behavior analysis process (S5) and performs the process. finish.

In FIG. 11, the size of the particle size is basically not unified, and the first usage method is used. However, if the size of the particle size is not the smallest and cannot be converted into spatial information, the data with a small particle size is estimated. Then, behavior analysis is performed, and conversion to spatial information is used by aligning data with a large granularity (position data in a large area) to data with a granularity that can be converted into spatial information or data with the smallest granularity. It is possible to improve the accuracy of the analysis process.

[Application example of the second usage method]
In the method of adjusting to a large grain size, the accuracy of the position data of a large area may be improved by referring to the position data having a small grain size with respect to the position data of the area.
Specifically, when the coordinate data with a small granularity exists in the administrative boundary area with a large granularity, the regional characteristics of the administrative boundary area are determined centering on the position of the coordinate data in the administrative boundary area.
For example, it is possible to analyze the behavior by considering both the regional characteristics of the mesh area to which the position of the coordinate data belongs and the regional characteristics of the administrative boundary area wider than that, and adding the regional characteristics of the mesh area to the regional characteristics of the administrative boundary area. Conceivable.

[Effect of Embodiment]
According to this system, various types of action log information such as user coordinate data, mesh area data, administrative boundary area data, postal code area data, and wireless device communication area data are analyzed from multiple action log providing servers. 1 acquires, organizes the data in time series, converts or estimates to the data format of specific action log information, and performs behavior analysis, and various types of position information with different granularity are time series. Since behavioral analysis can be performed by using it in association with, there is an effect that the accuracy of analysis can be improved.

In particular, in this system, even when the coordinate data with small granularity is used as a reference, new coordinate data can be inferred from the area data with large granularity to increase the position information and perform the behavior analysis process. It has the effect of increasing the accuracy of.

In addition, in this system, even when the area data with a large granularity is used as a reference, the behavioral analysis process can be performed from the coordinate data with a small granularity in consideration of the regional characteristics based on a specific position in the area. It has the effect of improving the accuracy of analysis.

The present invention is suitable for a behavior analysis system and a program thereof that use a plurality of types of position information in relation to each other in a time series to improve the accuracy of behavior analysis.

1 ... Analysis server, 3 ... Network, 11 ... Control unit, 12 ... Storage unit, 13 ... Interface unit, 21 ... User database (DB), 22 ... Analysis database (DB), 23 ... Spatial information conversion database (DB), 41, 51 ... Action log providing server, 42, 52 ... Action log database (DB)

Claims (8)

It is a behavior analysis system having an analysis server that inputs behavior information including the user's location information and performs behavior analysis.
The analysis server has a user database that stores the behavior information and a spatial information conversion database that converts the location information into corresponding facility information or area information.
The analysis server inputs position information having different granularity with respect to the position information, and the position information includes data of a position having a small granularity and data of an area having a large granularity, and is time-series for each user. When the position information is rearranged and the area data is between the two position data of the user's time series based on the rearranged time series position information , the line connecting the two position data. If the area does not include a position corresponding to the time of the data in the area, it is assumed that the user has moved in the area according to the situation in the time series, and the user is assumed to have moved in the area based on the time of the data in the area. A behavior analysis system characterized in that data at a position with a small granularity in an area is estimated and behavior analysis is performed using the data at the estimated position. The action according to claim 1, wherein the analysis server converts the data at the position having a small granularity into the data in the area with a large granularity including the position, and performs the behavior analysis using the data in the converted area. Analysis system. The analysis server inputs the data of the first area having a large grain size and the data of the second area having a smaller grain size than the data of the first area with respect to the data of the area, and the position data and the second area data. The data in the area can be converted into facility information or area information by the spatial information conversion database, and the data in the first area and the data in the position and / and the second area can be converted based on the time-series position information. It is characterized in that the data of a new position or the data of the second area in the first area is estimated from the data of the above, and the behavior analysis is performed using the data of the estimated position or the data of the second area. The behavior analysis system according to claim 1 or 2. The analysis server converts the position data into the data of the second area including the position, and performs the behavior analysis using the data of the converted second area and the estimated data of the second area. The behavior analysis system according to claim 3, which is characterized. A processing program in an analysis server that inputs behavioral information including the user's location information and performs behavioral analysis.
An input means for inputting position information having different granularity by executing a processing program on the analysis server, and
Regarding the position information including the data of the position with the small granularity and the data of the area with the large granularity , the position information is rearranged in time series for each user, and based on the rearranged time series position information, the time of the user. When there is data in the area between two position data in the series, and the area does not include a position corresponding to the time of the data in the area on the line connecting the two position data. Assuming that the user has moved in the area according to a time-series situation, an estimation means for estimating data at a position having a small granularity in the area based on the time of the data in the area, and an estimation means.
A program characterized in that it functions as an analysis means for performing behavioral analysis using the data of the estimated position. The processing program in the analysis server is provided with a position information conversion means for converting data at a position having a small granularity into data in a large area including the position.
The program according to claim 5, wherein the analysis means performs a behavioral analysis using the data of the converted area. The input means inputs the data of the first area having a large granularity with respect to the data of the area and the data of the second area having a smaller granularity than the data of the first area.
The analysis server includes a spatial information conversion means that converts the position data and the data of the second area into facility information or regional information by the spatial information conversion database.
The estimation means is based on the time-series position information, from the data of the first area and the data of the position and / and the data of the second area to the data of a new position in the first area or the second. Estimate area data
The program according to claim 5 or 6, wherein the analysis means performs a behavioral analysis using the data of the estimated position or the data of the second area. The position information conversion means converts the position data into the data of the second area including the position, and then converts the position data into the data of the second area.
The program according to claim 7, wherein the analysis means performs a behavioral analysis using the converted data of the second area and the estimated data of the second area.

Images