JP2013153286A - Information processing device and cluster generation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a cluster which is a population summing up unit for estimating the number of terminals more accurately.SOLUTION: Based on an attention area, an adjacent sector extraction part 19a extracts the aggregate of adjacent areas adjacent to the attention area. A determination part 19b determines whether or not the adjacent areas extracted by the adjacent sector extraction part 19a exceed a predetermined geographical extent and the adjacent areas determined not exceeding such extent are extracted by a candidate sector extraction part 19c as a candidate area. Then, a sector determination part 19d determines an object area based on the difference between the number of first terminals and the number of second terminals existing in a cluster configured of the extracted candidate area and the attention area.

Description

  The present invention relates to an information processing apparatus and a cluster generation method for generating a cluster composed of predetermined areas such as sectors.

  2. Description of the Related Art Conventionally, a technique for estimating the number of terminals located in a certain area (for example, sector) (so-called number of areas) based on the number of location registration information generated is known. However, the location registration information is generated at a predetermined location registration cycle, and the terminal user moves to another location registration area across the boundary of the location registration area (Location Area) (hereinafter referred to as “LA boundary”). Also generated. For this reason, in the sector facing the LA boundary, the location registration information generation probability is relatively higher than in other sectors, and the number of sectors in the sector facing the LA boundary may be overestimated. FIG. 20 is a conceptual diagram showing the correlation between the estimated number of locations around the LA boundary (vertical axis) and the distance from the LA boundary (horizontal axis), of which estimation based on the conventional method plotted with square points As is clear from the thick solid line graph for the number of locations and the thin solid line graph for the true value (true estimated number of locations) plotted with diamond-shaped points, in the vicinity of the horizontal axis center (LA boundary) in FIG. The number of estimated locations based on the conventional method is relatively larger than the true value.

  Japanese Patent Application Laid-Open No. 2004-228561 proposes a method of correcting the number of sectors in the sector facing the LA boundary in consideration of the excessive estimation of the number of sectors in the sector facing the LA boundary.

International Publication 2010 / 116903A1

  By the way, when the terminal user crosses the location registration area, the location registration cycle timer is reset to zero with the generation of the location registration signal. For this reason, in a sector slightly apart from the LA boundary (for example, a sector adjacent to the inside of the sector with respect to the LA boundary), the location registration information generation probability is relatively low, and the sector is slightly separated from the LA boundary. The number of sectors in the sector may be underestimated. At a position slightly away from the vicinity of the center of the horizontal axis (LA boundary) in FIG. 20 described above, the estimated number of locations based on the conventional method is relatively small with respect to the true value.

  However, a specific method for solving such a problem is not clearly described in Patent Document 1, and there is room for improvement in estimating the number of located areas with higher accuracy.

  Therefore, it is considered by the applicant to remove the so-called LA straddling location registration information to reduce the influence of LA straddling (Japanese Patent Application No. 2011-018860).

  However, even the method described in this application cannot completely remove the influence of the LA boundary.

  Therefore, the present invention has been made to solve the above-mentioned problems, and is a population aggregation unit for estimating the number of located areas more accurately by reducing the influence on the estimated value due to straddling the LA. An object of the present invention is to provide an information processing apparatus and a cluster generation method capable of generating a cluster.

  In order to solve the above-described problem, an information processing apparatus according to the present invention includes a selection unit that selects a target area for cluster configuration based on a pre-specified target area, and the target area and the target area. The first estimation means for estimating the number of terminals in the cluster as the first number of terminals based on the position registration signal, and the position from which the position registration information generated due to the terminal straddling the position registration area is removed Based on the registration information, second estimation means for estimating the number of terminals in the cluster as the second terminal number, the first number of terminals estimated by the first estimation means, and the second estimation means estimated by the second estimation means Based on the number of two terminals, it is determined that the target area is a determination area that constitutes a cluster, and a cluster is configured from the determined determination area and the area of interest And the cluster generation unit that includes a.

  Further, the cluster generation method of the present invention includes a selection step of selecting a target area for cluster configuration based on a pre-designated target area, and the number of terminals in the cluster composed of the target area and the target area. A first estimation step for estimating the number of first terminals based on the location registration signal, and a second estimation step for estimating the number of terminals in the cluster as the second number of terminals excluding the LA straddle based on the location registration signal. And determining, based on the number of first terminals estimated in the first estimation step and the number of second terminals estimated in the second estimation step, that the target area is a decision area constituting a cluster. And a cluster generation step of forming a cluster from the determined area and the area of interest.

  According to the present invention, a target area for cluster configuration is selected based on a pre-designated area of interest, and the number of terminals in the cluster composed of the area of interest and the target area is determined based on the location registration signal. Estimated as the number of terminals, and based on the location registration signal, the number of terminals in the cluster, excluding LA straddling, estimated as the number of second terminals, based on the estimated number of first terminals and the number of second terminals, It is determined that the target area is a determination area constituting the cluster, and a cluster is configured from the determined determination area and the area of interest. As a result, it is possible to generate a cluster which is a population aggregation unit for estimating the number of located areas with higher accuracy.

  In this information processing apparatus, the selection unit includes an adjacent area extraction unit that extracts a set of adjacent areas adjacent to the target area based on the target area, and an adjacent area extracted by the extraction unit. A candidate area extracting means for extracting the adjacent area as a candidate area when the cluster does not exceed a predetermined geographical spread, and the selecting means is selected from the candidate areas extracted by the candidate area extracting means. Select a target area, the cluster generation means, based on the difference between the number of the first terminal and the number of the second terminal in the cluster composed of the target area and the area of interest selected by the selection means, It is configured to determine that the target area is a determination area.

  According to this invention, based on the area of interest, a set of adjacent areas adjacent to the area of interest is extracted, and when the cluster to which the extracted adjacent area is added does not exceed a predetermined geographical spread, Adjacent areas are extracted as candidate areas. A target area is selected from the extracted candidate areas, and the target area is a decision area based on the difference between the number of first terminals and the number of second terminals existing in a cluster composed of the selected target area and the target area. Decide that. As a result, it is possible to generate a cluster which is a population aggregation unit for estimating the number of located areas with higher accuracy. In particular, the accuracy can be increased by limiting the area constituting the cluster based on the geographical spread.

  In the information processing apparatus according to the present invention, the selection unit has a minimum ratio of the difference with respect to a sum of the number of the first terminals and the number of the second terminals among the candidate areas extracted by the candidate area extraction unit. A certain area is selected as the target area.

  According to the present invention, the accuracy of the number of located areas can be further improved by selecting, as the target area, the one having the smallest ratio of the difference with respect to the total number of the first terminals and the second terminals.

  In the information processing apparatus of the present invention, when the candidate area extracting unit cannot extract a candidate area, the reference value for determining the geographical spread is changed, and the adjacent area extracting unit extracts the adjacent area. It is configured to re-process.

  According to the present invention, when the candidate area cannot be extracted, the reference value for determining the geographical spread is changed, and the adjacent area extraction process is performed again, so that the candidate area can be easily extracted, and the cluster can be extracted. A reasonable area to configure can be extracted.

  In the information processing apparatus according to the present invention, the selection unit does not select the narrow area as the target area when the narrow area is included in at least one of the target area and the target area. In the case where the target area or the target area includes a narrow area smaller than the target area or the target area, the first estimation means adds the number of terminals in the narrow area to the first terminal. And the second estimation means adds the number of terminals in the narrow area when the target area or the target area includes a narrow area smaller than the target area or the target area. The number of terminals is configured.

  According to the present invention, when a narrow area is included in at least one of the target area or the target area, the narrow area is not selected as the target area, and the target area or the target area has the target area. When a narrow area smaller than the area or the target area is included, the number of terminals in the narrow area is added as the first terminal number, and the target area or the target area includes a narrow area smaller than the target area or the target area. If included, by adding the number of terminals in the narrow area to the number of second terminals, for example, even if the sector serving as the area is configured with the outdoor station and the indoor station, it is appropriate Target area can be extracted.

  In the information processing apparatus of the present invention, the area of interest or the target area is an area formed by a base station arranged outdoors, and the narrow area is an area formed by a base station arranged indoors. Suppose that

  According to this invention, the area of interest or the target area is an area formed by a base station arranged outdoors, and the narrow area is an area formed by a base station arranged indoors.

  In the information processing apparatus according to the present invention, the area of interest and the target area are mesh regions defined in advance, and the first estimation unit and the second estimation unit determine the number of terminals estimated in units of sectors. The number of terminals apportioned by the area ratio between the sector and the mesh area is calculated in units of mesh areas, and the number of first terminals and the number of second terminals in each cluster configured in units of mesh areas are estimated. Based on the target mesh area including the target area, the target mesh area as the target area constituting the cluster is selected, and the cluster generation means determines that the target mesh area is a determined mesh area constituting the cluster. Determining and generating a cluster composed of the target mesh region and the target mesh region, That.

  According to the present invention, the target area and the target area are predefined mesh areas, and the number of terminals estimated in units of sectors is prorated according to the area ratio between the sectors and the mesh areas, and each area is determined in units of mesh areas. Estimated as the number of one terminal and the number of second terminals, based on the target mesh area including the target sector, select the target mesh area as the target area that constitutes the cluster, and the target mesh area determines the mesh that constitutes the cluster A region is determined and a cluster composed of the target mesh region and the target mesh region is generated. Thereby, a cluster can be comprised per mesh area | region and an application range can be expanded more.

  In the information processing apparatus according to the present invention, the cluster generation unit is configured to determine the target mesh area based on an adjacent mesh area adjacent to the target mesh area.

  According to this invention, the target mesh region can be determined based on the adjacent mesh region adjacent to the target mesh region.

  ADVANTAGE OF THE INVENTION According to this invention, the cluster which is a totaling unit of the population for estimating the number of located areas more accurately can be produced | generated.

It is a system configuration figure of communication system 1 of this embodiment. 2 is a block diagram showing a functional configuration of a terminal number estimation device 10. FIG. 3 is a block diagram illustrating a functional configuration of a cluster generation unit 19. FIG. It is explanatory drawing which shows the relationship between an attention sector and an adjacent sector. It is explanatory drawing at the time of judging the geographical extension in attention sector A1 and adjacent sector A2. It is explanatory drawing at the time of judging the geographical breadth in attention sector A1 and adjacent sector A3. It is explanatory drawing at the time of judging the geographical extension in attention sector A1 and adjacent sector A4. It is explanatory drawing at the time of judging the geographical expansion in attention sector A1 and adjacent sectors A2 and A3. It is explanatory drawing at the time of judging the geographical expansion in attention sector A1 and adjacent sectors A3 and A4. It is a flowchart which shows the cluster production | generation process in the terminal number estimation apparatus 10. FIG. It is explanatory drawing which shows the relationship with the sector of an indoor station, and the sector of an outdoor station. It is a block diagram which shows the function structure for producing | generating a cluster for every mesh area | region. It is explanatory drawing which shows the relationship between a mesh area | region and a sector. It is a block diagram which shows the structure of the 1st, 2nd in-zone number estimation part 15. FIG. It is a figure for demonstrating the view of terminal number estimation. It is a figure for demonstrating the calculation method concerning terminal number estimation. It is a flowchart which shows a terminal number estimation process. It is a flowchart which shows the estimation process of the 1st, 2nd in-zone number. It is a flowchart which shows the calculation process of a feature-value. It is a conceptual diagram which shows the correlation with the estimated location number around LA boundary, and the distance from LA boundary.

  Embodiments of the present invention will be described with reference to the accompanying drawings. Where possible, the same parts are denoted by the same reference numerals, and redundant description is omitted.

[Configuration of communication system]
FIG. 1 is a system configuration diagram of a communication system 1 according to the present embodiment. As shown in FIG. 1, the communication system 1 includes a mobile terminal 100, a BTS (base station) 200, an RNC (radio control device) 300, an exchange 400, various processing nodes 700, and a management center 500. Yes. The management center 500 includes a social sensor unit 501, a petamining unit 502, a mobile demography unit 503, and a visualization solution unit 504.

  The exchange 400 collects position information, which will be described later, about the mobile terminal 100 via the BTS 200 and the RNC 300. The RNC 300 can measure the position of the mobile terminal 100 using the delay value in the RRC connection request signal when communication connection is performed with the mobile terminal 100. The exchange 400 can receive the position information of the mobile terminal 100 measured in this way when the mobile terminal 100 executes communication connection. The exchange 400 stores the received position information, and outputs the collected position information to the management center 500 in response to a predetermined timing or a request from the management center 500.

  The various processing nodes 700 acquire the location information of the portable terminal 100 through the RNC 300 and the exchange 400, and perform recalculation of the location in some cases, and are collected at a predetermined timing or in response to a request from the management center 500 The position information is output to the management center 500.

  As the position information of the mobile terminal 100 in the present embodiment, a sector number indicating a serving sector obtained from the position registration information, position positioning information obtained by a position information acquisition system using PRACH PD, and the like can be employed. The position data of the portable terminal 100 includes the position information as described above, identification information for identifying the portable terminal (for example, information associated with the portable terminal such as a line number), and position acquisition time information at which the position information is acquired. including. When a line number is used as identification information, it is preferable to use a value associated with the line number (for example, a hash value of the line number) instead of using the line number as it is.

  As described above, the management center 500 includes the social sensor unit 501, the petamining unit 502, the mobile demography unit 503, and the visualization solution unit 504. In each unit, the position data of the mobile terminal 100 is used. Perform statistical processing. In addition, the terminal number estimation apparatus 10 (FIG. 2) mentioned later can be comprised by the management center 500, for example.

  The social sensor unit 501 is a server device that collects data including the position data of the portable terminal 100 from each exchange 400 and various processing nodes 700 or offline. The social sensor unit 501 receives data periodically output from the exchange 400 and the various processing nodes 700, or acquires data from the exchange 400 and the various processing nodes 700 according to a predetermined timing in the social sensor unit 501. It is configured to be able to do.

  The petamining unit 502 is a server device that converts data received from the social sensor unit 501 into a predetermined data format. For example, the petamining unit 502 performs a sorting process using a user ID as a key, or performs a sorting process for each area.

  The mobile demography unit 503 is a server device that performs aggregation processing on the data processed in the petamining unit 502, that is, count processing for each item. For example, the mobile demography unit 503 can count the number of users located in a certain area, and can total the distribution of the located areas.

  The visualization solution unit 504 is a server device that processes the data aggregated in the mobile demography unit 503 so as to be visible. For example, the visualization solution unit 504 can map the aggregated data on a map. Data processed by the visualization solution unit 504 is provided to companies, government offices or individuals, and is used for store development, road traffic surveys, disaster countermeasures, environmental countermeasures, and the like. It should be noted that the information statistically processed in this way is processed so that individuals are not specified so as not to infringe privacy.

  The social sensor unit 501, the petamining unit 502, the mobile demography unit 503, and the visualization solution unit 504 are all configured by the server device as described above, and although not shown, the basic configuration of a normal information processing device Needless to say, it includes a CPU, a RAM, a ROM, an input device such as a keyboard and a mouse, a communication device that communicates with the outside, a storage device that stores information, and an output device such as a display and a printer.

[Configuration of terminal number estimation device]
Next, the terminal number estimation apparatus according to the present embodiment will be described. FIG. 2 shows a functional block configuration of the terminal number estimation apparatus 10. As shown in FIG. 2, the terminal number estimation device 10 includes a position data acquisition unit 11, an accumulation unit 12, an observation period acquisition unit 13, an observation area acquisition unit 14, a first in-zone number estimation unit 15 (first estimation Means), a signal removal unit 16, a second in-zone number estimation unit 17 (second estimation unit), a sector selection unit 18, a cluster generation unit 19 (selection unit, cluster generation unit), and an output unit 20. Yes. In the present embodiment, the terminal number estimation device 10 functions as an information processing device for generating a cluster composed of a predetermined area such as a sector based on the estimated number of terminals.

  Hereinafter, the function of each part of the terminal number estimation apparatus 10 of FIG. 2 will be described. The position data acquisition unit 11 acquires position data from the outside (for example, the exchange 400 or various processing nodes 700) and stores it in the storage unit 12. In addition, it is not essential to provide the position data acquisition unit 11 in the terminal number estimation device 10, and the position data acquired by the position data acquisition unit outside the terminal number estimation device 10 is transmitted to the terminal via, for example, a storage medium. You may input into the number estimation apparatus 10. FIG.

  The accumulation | storage part 12 preserve | saves the positional data over several time about many users (mobile terminal).

  The observation period acquisition unit 13 acquires observation period information including a set of observation start time and observation end time.

  The observation area acquisition unit 14 acquires observation area information associated with one or more pieces of position information. The observation area information here is given as, for example, a sector number, a latitude / longitude, a geographical range (for example, a municipality), etc., and the observation area acquisition unit 14 displays the expression format and position information of the acquired observation area information. It is desirable to provide a database that manages information that associates the expression format (for example, correspondence information between sector numbers and latitude and longitude).

  The first located area estimation unit 15 reads the position data from the storage unit 12, and based on the position data, calculates the number of terminals located in the observation area during the observation period as the first located number (first terminal). Number). The number-of-terminals estimation method by the first serving number estimation unit 15 is not limited to a specific method, and various methods can be adopted. An example of the terminal number estimation method and the configuration of the first in-zone number estimation unit 15 for executing the estimation method will be described later.

  In addition to the examples described later, for example, a method described in Japanese Patent Application No. 2010-212456 filed earlier by the present applicant can be used as the terminal number estimation method. In this method, the information analysis device receives position data indicating the position of the user, positioning time information from which the position information is obtained, and point data including the user ID from the outside, and determines the positioning time from the point data for each user. Extracts point data immediately before the target time and point data whose positioning time is immediately after the target time, and for each user, the position indicated by the point data immediately before the target time and the position indicated by the point data immediately after the target time Is a series of methods for estimating the user's position at the target time by calculating the population distribution of a predetermined calculation target area unit at the target time based on the estimated position of each user. Yes, this method can be applied to terminal number estimation.

  The signal removal unit 16 reads position data from the storage unit 12, and generates position registration information (hereinafter referred to as "position registration information by straddling LA") generated from the position data resulting from the terminal straddling the position registration area. ) To obtain position data after removal. The “position data after removal” obtained here includes the position registration information excluding the position registration information across the LA in the position registration information. Of course, position information other than the position registration information (for example, GPS positioning) Information) and the like. In addition, information indicating that the position registration information is based on the LA straddle is described in the position registration information, and the signal removal unit 16 can remove the position registration information based on the LA stride based on the information.

  Based on the position data after removal obtained by the signal removal unit 16, the second number-of-zones estimation unit 17 determines the number of terminals located in the observation area during the observation period as the second number of regions (second Estimated as the number of terminals). The method for estimating the number of terminals by the second serving number estimation unit 17 is not limited to a specific method, and various methods can be adopted including the method described in Japanese Patent Application No. 2010-212456 described above. However, it is desirable that the method is the same as the terminal number estimation method by the first serving number estimation unit 15. In the present embodiment, an example in which the second number-of-services estimation unit 17 employs the same terminal number estimation method as that of the first region-of-services estimation unit 15 will be described. An example of a specific terminal number estimation method and the configuration of the second in-zone number estimation unit 17 for executing the estimation method will be described later.

  The target sector selection unit 18 is a part that selects a target sector based on a sector specified by an operator of the terminal number estimation device 10 or a sector specified via an external network. For example, when the operator wants to know the population distribution in a certain sector, the operator selects a certain sector as the target sector via the target sector selection unit 18. The target sector selection unit 18 adds the selected target sector to the cluster. The cluster here is a cluster which is an initial value, and adjacent sectors are sequentially added to this cluster.

  Based on the target sector selected by the target sector selection unit 18, the cluster generation unit 19 selects a target sector that constitutes a cluster that is a population aggregation unit with little influence of straddling LA. In determining, whether the selected target sector is valid based on the first and second number of visited areas estimated by the first located number estimating unit 15 and the second located number estimating unit 17 Judgment is made, and if it is valid, the target sector is determined as a determination sector constituting the cluster.

  The cluster generation unit 19 extracts an adjacent sector adjacent to the cluster based on the cluster including the sector of interest, and selects a candidate sector in consideration of geographical spread from the extracted adjacent sector. Then, based on the candidate sector and cluster selected here (initially composed only of the sector of interest), one or a plurality of target sectors constituting the cluster are extracted from the plurality of candidate sectors, and this is clustered. Add to If the cluster to which the target sector has been added does not satisfy the predetermined condition shown in Equation (1) described later, an adjacent sector adjacent to the cluster to which the target sector has been added is extracted, and this is used as a new candidate sector. And repeat these processes. Further, when a predetermined condition represented by equation (2) described later is satisfied, the extracted target sector is determined as a determined sector. When the predetermined condition is not satisfied, a new target sector is added to the cluster. As a cluster, the above process is performed again.

  As described above, the cluster handled by the cluster generation unit 19 is initially composed of the sector of interest, and when the target sector is added, a new cluster is generated and the range becomes large. Depending on the conditions, the cluster includes many sectors, and processing is performed on sectors adjacent to such a cluster. Hereinafter, detailed processing and configuration of the cluster generation unit 19 will be described.

  The cluster generation unit 19 will be described with reference to FIG. FIG. 3 is a block diagram illustrating a functional configuration of the cluster generation unit 19. As shown in FIG. 3, the cluster generation unit 19 includes an adjacent sector extraction unit 19a (selection unit, adjacent area extraction unit), a determination unit 19b, a candidate sector extraction unit 19c (selection unit, candidate area extraction unit), and sector determination. The unit 19d (selection means, cluster generation means) is included.

  The adjacent sector extracting unit 19a extracts an adjacent sector adjacent to the target sector selected by the target sector selecting unit 18. FIG. 4 is an explanatory diagram showing the relationship between the sector of interest and adjacent sectors. As shown in FIG. 4, the adjacent sector extraction unit 19a selects and extracts adjacent sectors A2 to A7 adjacent to the sector of interest A1.

  The adjacent sector extraction unit 19a generates a Voronoi region by performing Voronoi division based on the representative points in each sector, and creates a Delaunay triangulation for the Voronoi region, thereby allowing the adjacent relationship between the sectors. Can be grasped. Note that the adjacent relationship between sectors may be grasped by other methods, and in that case, it is not necessary to create a Delaunay triangular network. For example, the adjacent relationship may be grasped by using the sector boundary estimation method described in Japanese Patent Application No. 2011-080204 filed earlier by the applicant.

  The determination unit 19b determines the geographical distance between sectors, that is, if a plurality of sectors are configured as one cluster, determines whether or not the cluster is too wide, and the sector A1 of interest. This is a part for determining whether or not the geographical spread between the adjacent sectors A2 to A7 exceeds the reference value. For example, the determination unit 19b determines whether the distance between the representative point of each sector and the center of gravity of the cluster exceeds a threshold value d. The barycentric point of the cluster is calculated based on the representative points of the sectors constituting the cluster. Note that the above method based on the distance from the center of gravity is an example, and the present invention is not limited to this.

  Further, the determination unit 19b may determine the geographical spread of each of the adjacent sectors A2 to A7 with respect to the sector of interest A1, or may determine the geographical spread of a plurality of adjacent sectors. May be. For example, the geographical spread between the sector of interest A1 and the adjacent sectors A2 and A3 may be determined.

  Here, with reference to FIG. 5 to FIG. 9, a schematic diagram showing how the determination unit 19 b makes any one or a plurality of adjacent sectors A <b> 2 to A <b> 7 with respect to the sector A <b> 1 as one cluster. Indicates. FIG. 5 schematically shows a case where the spread is determined in the case where the target sector A1 and the adjacent sector A2 are formed as one cluster, and FIG. 6 shows the target sector A1 and the adjacent sector A3. In the case of one cluster, FIG. 7 schematically shows a case where the spread is determined when the sector of interest A1 and the adjacent sector A4 are set as one cluster. As shown in FIG. 5 to FIG. 7, the determination unit 19 b extracts adjacent sectors one by one and puts them in a cluster for determination. However, the determination unit 19 b is not limited to this, and as illustrated in FIG. 8 and FIG. 9, You may make it judge the geographical extension in case the attention sector and several adjacent sectors are made into one cluster. In FIG. 8, adjacent sectors A2 and A3 are included in one cluster with respect to the sector of interest A1. In FIG. 9, adjacent sectors A3 and A4 are included in one cluster with respect to the sector of interest A1.

  The candidate sector extracting unit 19c extracts, as a candidate sector, an adjacent sector (or a combination of adjacent sectors) that does not exceed the geographical spread when configured as a cluster by the determining unit 19b, and temporarily holds this. That is, as shown in FIGS. 5 to 7, when adjacent sectors are extracted one by one, the adjacent sectors are held as candidate sectors, and as shown in FIGS. Are extracted, the combination is held as a candidate sector.

なお、Ｅ１は、第１の在圏数推計部１５により算出されたクラスタ単位の在圏数、Ｅ２は、第２の在圏数推計部１７により算出されたクラスタ単位の在圏数を示し、ｉはセクタの識別子を示す。それぞれ、候補セクタ抽出部１９ｃにより抽出された候補セクタ集合のうち、着目セクタＡ１と、地理的広がりが基準値以内にある隣接セクタＡ２〜Ａ７のいずれか一つを入れ替えてそれぞれ計算したものである。すなわち、着目セクタＡ１＋Ａ２、着目セクタＡ１＋Ａ３、着目セクタＡ１＋Ａ４・・として、式（１）が最小となるセクタの組合せを１つ決定する。 Based on the previously created cluster and the candidate sector extracted by the candidate sector extraction unit 19c, the sector determination unit 19d determines the minimum sector set (cluster and candidate sector) as the target sector according to the following equation (1). Choose as.

E1 represents the number of visited areas in cluster units calculated by the first located area number estimating unit 15, E2 represents the number of located areas in cluster units calculated by the second located area number estimating unit 17, i indicates the identifier of the sector. Each of the candidate sector sets extracted by the candidate sector extraction unit 19c is calculated by replacing any one of the sector A1 of interest and the adjacent sectors A2 to A7 whose geographical spread is within the reference value. . That is, one sector combination that minimizes Equation (1) is determined as the target sector A1 + A2, the target sector A1 + A3, and the target sector A1 + A4.

を満たすか否か、判断する。そして、セクタ決定部１９ｄは、この式（２）を満たす対象セクタを、クラスタを構成する決定セクタと決定する。 Based on the equation (1), the sector determination unit 19d generates a cluster using the smallest target sector, and further

It is determined whether or not the above is satisfied. Then, the sector determining unit 19d determines a target sector that satisfies the equation (2) as a determined sector that constitutes a cluster.

  Returning to FIG. 2, the output unit 20 outputs the number of terminals using the clusters generated by the cluster generation unit 19 as a totaling unit. The output here includes various output modes such as display output, audio output, and print output. The output unit 20 may output the cluster (a set of target sectors) generated by the cluster generation unit 19 without outputting the number of terminals.

  Furthermore, the output unit 20 may output the number of terminals according to the following processing when counting the number of terminals.

The first serving number estimation unit 15 acquires the number of terminals L _ij of the specified attribute j in the target sector i specified by an operator or the like, and the incoming number acquisition unit (not shown) in the target sector i The number of incoming calls C _i is acquired.

And the 1st in-zone number estimation part 15 acquires the in-zone number Na in the cluster produced | generated in the cluster production | generation part 19 as mentioned above. Then, the population distribution calculation unit (not shown) obtains the number of incoming calls C _i and the total number of incoming calls ΣC _i obtained for each cluster, the number of visited areas Na in the cluster, and the number of acquired terminals. Based on L _ij , the population distribution is calculated by calculating the corrected number of terminals S _ij in the sector. Thereby, it is possible to accurately calculate the population distribution for each sector and for each attribute, which eliminates the influence of straddling the location registration area.

を算出することにより人口分布を算出することができる。 In calculating the population distribution, when the target sector and the target sector are represented as sector i, the population distribution calculation unit (not shown) determines the number of mobile terminals 100 having the attribute j in the target sector and the target sector. L _ij , where N is the number of areas in the cluster composed of the target sector and the target sector, and C _i is the number of incoming calls in the target sector and the target sector,

The population distribution can be calculated by calculating.

Note that the number of terminals L _ij in the present embodiment includes the number of terminals indicating the number of mobile stations, the number of entrances, or a feature amount indicating the existence ratio in a certain observation time zone. ) Indicates the amount of terminals existing, that is, the number of terminals.

  Next, processing of the terminal number estimation device 10 will be described. FIG. 10 is a flowchart showing cluster generation processing in the terminal number estimation apparatus 10. First, in the determination unit 19b and the sector determination unit 19d, an allowable threshold k for LA error and a threshold d for cluster geographical spread are set (S101). Further, the representative point of each sector is extracted by the adjacent sector extracting unit 19a, and a Delaunay triangular network is created for the set of representative points (S102). Note that it is not essential to create a Delaunay triangulation, and when the adjacent sector can be grasped by another method, the Delaunay triangulation generation process in S102 can be omitted.

  Then, the sector of interest to be subject to population distribution is selected by the sector of interest selection unit 18 and added to the cluster set as the initial value (S103). Next, the adjacent sector extraction unit 19a extracts the adjacent sector adjacent to the sector of interest, sets this as an adjacent sector set (S104), and extracts one adjacent sector from the adjacent sector set (S105). When the extracted one adjacent sector is added to the cluster including the sector of interest, the determination unit 19b determines whether or not the geographical spread of the cluster exceeds the threshold value d (S106).

  Here, when the determining unit 19b determines that the geographical spread of the cluster is equal to or less than the threshold value d, the one adjacent sector is determined to be a candidate sector, and the candidate sector extracting unit 19c determines the candidate sector set. And is stored in the candidate sector extracting unit 19c (S107). When the above-described processing is performed on all adjacent sectors (S108: YES), the candidate sector extraction unit 19c checks whether there is a candidate sector (S109).

  Here, if it is determined that no candidate sector is held in the candidate sector extraction unit 19c, the determination unit 19b updates the threshold d by adding a predetermined step value to the threshold d. (S110). Then, the sector that has been held as a sector constituting the cluster is reset (S111), the extraction process of the set of adjacent sectors adjacent to the cluster is performed again, and the candidate sector extraction process is performed again (S104). . Although the accuracy is improved by the processing in step S104, the processing is time-consuming and is not necessarily essential.

  If there is a candidate sector in S109 (S109: YES), the sector determining unit 19d extracts the sector that minimizes the above formula (1) from the candidate sector set as the target sector (S112). The extracted target sector is added to the cluster (S113). The sector determination unit 19d verifies whether or not the above equation (2) using the number of areas in the target sector configured by the extracted clusters is less than 2k (S114). If it is determined that the target sector is satisfied, it is determined that the target sector is a determined sector, a cluster including the determined sector is output, and the cluster is used as a unit for estimating the population distribution (S115). If it is determined that the condition is not satisfied, a set of adjacent sectors adjacent to the created cluster is extracted (S104), and candidate sector extraction / determination processing is performed again.

  FIG. 5 shows a schematic diagram when the set of adjacent sectors is extracted again without satisfying Expression (2) in S114. As shown in FIG. 5, it is assumed that the target sector A1 and its adjacent adjacent sectors A2 to A7 are extracted, of which the target sector A1 and the adjacent sector A2 are combinations of sectors that minimize Equation (1). In that case, in S104, adjacent sectors A3 to A10 adjacent to the sector of interest A1 and the adjacent sector A2 are extracted.

  In this way, a cluster is generated based on adjacent sectors adjacent to the sector of interest. For the generated cluster, a sector having a small ratio with respect to the total sum of the number of in-zones considering LA straddling and the number of in-zones excluding LA straddling and the total sum is selected. Therefore, a cluster with a small error across the LA is formed, and a unit capable of calculating a more accurate population distribution can be generated.

  In the above description in FIG. 10, for convenience of description, adjacent sectors are added to the cluster one by one, and the geographical distance is determined and the expressions (1) and (2) are checked. However, the present invention is not limited to this, and a plurality of adjacent sectors may be added to the cluster to determine the appropriateness of geography.

  Next, an application example in the above-described embodiment will be described. In general, base stations (including antennas) that form a sector are mainly set outdoors, and are installed indoors, such as buildings and basements, where the sector has a large geographical area. There is a small indoor station. In some cases, the indoor station and the outdoor station are mixed in the same area, and the ranges covered by the sector overlap.

  In areas where outdoor stations and indoor stations are mixed, a Voronoi diagram is created only at the outdoor station, and a Delaunay triangulation is formed as necessary. The adjacent sector extraction process can be performed on the. And in the sector formed in the indoor station with respect to the serving area number E1 and the serving area number E2 calculated based on the cluster composed of the sectors of the respective outdoor stations in the equations (1) and (2), The determined sector can be determined by adding the number of serving areas as the new number of serving areas E1 and E2.

  FIG. 11 shows a schematic diagram thereof. FIG. 11 (a) shows that there are sectors a1 to a3 of the indoor station in the sector A1 of the outdoor station. In such a case, using the sectors A1 to A3 of the outdoor station, etc., an adjacent sector grasping process is performed, and a candidate sector extracting process and a target sector extracting process are performed. When performing extraction processing of candidate sectors and target sectors, the number of locations E1 and E2 in Equations (1) and (2) include not only the number of sectors in the outdoor station, but also indoors in that sector. If there is a station sector, the sum of the number of sectors in the indoor station sector is calculated as E1 and E2.

  For example, in FIG. 11A, since the sector A1 includes the sectors a1 to a3, the numbers E1 and E2 in the sector A1 are calculated, and the numbers in the sectors a1 to a3 are calculated as the above-mentioned number E1. And E2 are added to obtain new locations E1 and E2.

  In addition, as shown in FIG. 11B, the sectors a4 and a5 of the indoor station may straddle a plurality of sectors. In FIG. 11B, the sector a4 of the indoor station is installed across both the sectors A1 and A6 of the outdoor station. In such a case, by adding the number of areas corresponding to the area apportionment to the number of areas E1 and E2 of the outdoor station sector, new area numbers E1 and E2 are calculated, and the above formula (1 ) And equation (2).

  As a modification, it is conceivable to form clusters in units of mesh areas instead of units in sectors. FIG. 12 is a block diagram showing a functional configuration for generating clusters in units of mesh areas as the cluster generation unit 19x.

  As shown in FIG. 21, the cluster generation unit 19x includes a target mesh region extraction unit 19o, an adjacent mesh region extraction unit 19p, a determination unit 19q, a candidate mesh region extraction unit 19r, and a mesh region determination unit 19s. ing.

  The target mesh area extraction unit 19o is a part that extracts the target mesh area based on the target sector selected by the target sector selection unit 18, and extracts the mesh area including the target sector as the target mesh area. Note that the mesh area is an area that is divided into rectangular shapes in advance, and is an area that is defined in advance as an aggregation unit for mainly calculating population distribution. For example, in FIG. 13, when sector A1 is selected as the target sector, the target mesh region M1 is extracted as the target mesh region.

  The adjacent mesh region extraction unit 19p is a part that extracts an adjacent mesh region adjacent to the target mesh region extracted by the target mesh region extraction unit 19o. When processing is performed in units of mesh areas, it is possible to grasp the adjacent relationship between mesh areas without using the Delaunay triangulation network described above. The adjacent mesh region extraction unit 19p performs the same process as the adjacent sector extraction unit 19a.

  The determination unit 19q is a part that determines the geographical spread of the cluster including the target mesh region and the adjacent mesh region. .

  The candidate mesh area extraction unit 19r is a part that extracts candidate mesh areas as candidates from adjacent mesh areas. When the determination unit 19q is omitted, this configuration can also be omitted. The candidate mesh area extraction unit 19r performs the same process as the candidate sector extraction unit 19c.

  The mesh area determination unit 19s is a part that determines a target mesh area that is a cluster configuration target from the candidate mesh area or the adjacent mesh area extracted by the candidate mesh area extraction unit 19r. The mesh area determination unit 19s performs the same processing as the sector determination unit 19d described above.

  That is, the mesh area determination unit 19 s performs the first and second area numbers calculated by the first area number estimation unit 15 and the second area number estimation unit 17. Then, the third area number and the fourth area number are calculated by proportionally transforming into mesh area units, and these area numbers are set as area numbers E1 and E2, and are expressed in Equations (1) and (2). By applying, a target mesh region that forms a cluster can be determined.

  FIG. 13 is a schematic diagram showing the relationship between mesh regions and sectors. As shown in FIG. 13, there are mesh regions M1 to M4 and sectors A1 to An. Here, when the focused sector A1 is selected as the focused sector, the focused mesh area extraction unit 19o extracts the mesh area including the focused sector A1 as the focused mesh area M1.

  Then, the adjacent mesh region extraction unit 19p extracts the adjacent mesh regions M2 and M4 that are adjacent to each other. Here, the mesh region M3 is not treated as an adjacent region, but may be included.

Note that there may be a sector of interest at the boundary of the mesh region. In this case, a mesh area in which the areas of the target sector overlap each other may be set as the target mesh area, or an arbitrary mesh area may be selected by the operator's hand. Further, as an error process, an arbitrary message may be notified to the operator.

[About the 1st and 2nd area number estimation processing]
Next, a method for calculating the number of locations is described. In this embodiment, as will be described later, as the first and second coverage numbers, the feature quantity indicating the temporal existence ratio of the location information is used instead of counting the number of location registration signals of the terminal. However, it is not limited to this.

[Configuration of first and second in-zone number estimation unit]
In FIG. 14, the structure of the 1st in-zone number estimation part 15 is shown. In the present embodiment, since the second number-of-zones estimation unit 17 employs the same terminal number estimation method as the first number-of-zones estimation unit 15, the configuration of the second number-of-zones estimation unit 17 is as follows. It is the same as that of the structure of the 1st in-zone number estimation part 15 to describe.

  As shown in FIG. 14, the first in-zone number estimation unit 15 includes an observation target acquisition unit 31, a front / rear position data acquisition unit 32, a feature amount calculation unit 33, and a terminal number estimation unit 34.

  The observation target acquisition unit 31 acquires the observation start time information and the observation end time information regarding the observation period to be observed from the observation period acquisition unit 13 and the observation area information regarding the observation area to be observed from the observation area acquisition unit 14, respectively. One or more pieces of position data including position acquisition time information after the observation start time and before the observation end time and including position information associated with the observation area information are stored as observation target position data from the storage unit 12. get. Note that the observation target position data may be further narrowed down according to separately given conditions (for example, the age group of the user of the mobile terminal).

  The front-rear position data acquisition unit 32 is the same as the first position data for position data (hereinafter referred to as “first position data”) for which a feature amount, which will be described later, calculated in the process of estimating the number of locations is obtained. Among the position data including the identification information, position acquisition time information of position data immediately before the first position data (hereinafter referred to as “second position data”), and position data immediately after the first position data The position acquisition time information (hereinafter referred to as “third position data”) is acquired. It is not essential for the front / rear position data acquisition unit 32 to acquire the entire second or third position data, and at least the position acquisition time information included in the position data may be acquired. The front / rear position data acquisition unit 32 may read the position acquisition time information of the second and third position data from the storage unit 12 or may pass it from the position data acquisition unit 11. Logically, either method can be used.

  The feature amount calculation unit 33 calculates a feature amount for each of the first position data. For example, the feature amount calculation unit 33 calculates the difference between the position acquisition time of the second position data and the position acquisition time of the third position data as the feature amount for the first position data. In addition, when the position acquisition time of the second position data is an abnormal value, the feature amount calculation unit 33 calculates the position acquisition time of the first position data and the position acquisition time of the second position data as an example here. When the difference is larger than a predetermined reference value (for example, 1 hour), the time of the second position data is set to a time that is retroactive from the position acquisition time of the first position data by a predetermined time (for example, 1 hour). The feature amount for the first position data is calculated as the acquisition time. Similarly, when the position acquisition time of the third position data is an abnormal value, the feature amount calculation unit 33 uses the position acquisition time of the first position data and the position acquisition time of the third position data as an example here. Is greater than a predetermined reference value (for example, 1 hour), the time advanced from the position acquisition time of the first position data to the future by a predetermined time (for example, 1 hour) is set in the third position data. The feature amount for the first position data is calculated using the position acquisition time. The process when the position acquisition times of the second and third position data are abnormal values is not an essential process, but the portable terminal 100 is located outside the service area by performing the above process. When the position data acquisition time interval becomes abnormally long due to the power of the mobile terminal 100 being turned off or the like, the influence of the abnormally long acquisition time interval is prevented from being excessively increased. be able to.

  The number-of-terminals estimation unit 34 estimates the number of terminals located in the observation area during the observation period based on the feature amount of the observation target position data and the observation period length that is the difference between the observation start time and the observation end time. . Although details will be described later, the terminal number estimation unit 34 estimates a numerical value obtained by dividing the total sum of the feature amounts of the observation target position data by twice the observation period length as the number of terminals.

即ち、各端末ａ_ｉの観測期間内のセクタＳの滞在時間ｔ_ｉの総和を観測期間の長さＴで除した結果を、端末数ｍとして推計する。ただし、端末ａ_ｉの観測期間内のセクタＳの滞在時間ｔ_ｉの真の値は観測不能であるが、各端末ａ_ｉは信号（例えば位置登録情報を含む位置登録信号）を発信し、それらの信号は観測可能である。 [Concept of terminal number estimation and calculation method]
Next, the concept and calculation method of terminal number estimation will be described. As model shown in Figure 15, during a certain observation period (length T), n pieces of terminal _{a 1, a 2, ...,} a n passes through the sector S, the observation period for each terminal _{a i} It is assumed that the stay time of the sector S is t _i (0 <t _i ≦ T). At this time, the number m of terminals present in the sector S (actually the average value of the number m of terminals present in the sector S within the observation period) is expressed by the following equation (3).

That is, the result of dividing the sum of the stay times t _i of the sectors S within the observation period of each terminal a _i by the length T of the observation period is estimated as the number of terminals m. However, although the true value of the stay time t _i of the sector S within the observation period of the terminal a _i is not observable, each terminal a _i transmits a signal (for example, a position registration signal including position registration information). These signals are observable.

（ｘ_ｉは、端末ａ_ｉが観測期間内にセクタＳで発信した信号の総数）とすると、端末数の推計とは、観測された信号ｑ_ｉｊ（ｊは１以上ｘ_ｉ以下の整数）からｍの値を推計することに他ならない。 The signals transmitted by the terminal a _i from the sector S during the observation period

If x _i is the total number of signals transmitted by the terminal a _i in the sector S during the observation period, the estimation of the number of terminals is based on the observed signal q _ij (j is an integer from 1 to x _i ). It is none other than estimating the value of m.

Now, a method for calculating the number of terminals will be described with reference to FIG. Let p _i be the density at which the signal q _ij is transmitted from the terminal a _i (ie, the number of signals per unit time). At this time, if the probability that the signal is transmitted is independent of the sector, the expected value E (x _i ) of the total number x _i of signals transmitted from the terminal a _i in the sector S within the observation period is E (x _i) since a = _{t i} × _{p i,} the expected value E _{(t i)} the following expression for the residence time _{t i} of the sector S in the observation period of the terminal _{a i} (2) is satisfied.
E (t _i ) = x _i / p _i (4)
Here, when the transmission time of the signal _{q ij} was _{u ij,} density _{p ij} of the signal _{q ij} is given by the following equation (5).
p _ij = 2 / (u _{i (j + 1)} −u _{i (j−1)} ) (5)
Here, when the signal q _ij is a signal related to the first position data, the signal q _{i (j−1)} is a signal related to the second position data, and the signal q _{i (j + 1)} is related to the third position data. Corresponds to the signal. In the present embodiment, the transmission time _{u i} of the signal of the second position data _{q i (j-1)} transmission time _{u i} of _(j-1) and the signal _{q i} of the third position data _{(j + 1) ( j + 1)} , that is, (u _{i (j + 1)} −u _{i (j−1)} ) in the above equation (5) is defined as the feature quantity w _ij for the first position data (feature quantity w _ij = u _{i (j + 1)} -u _{i (j-1)} ). Therefore, the above formula (5) is as follows.
p _ij = 2 / (u _{i (j + 1)} −u _{i (j−1)} ) = 2 / w _ij (6)

で与えられるため、端末数ｍの推計値Ｅ（ｍ）は以下の式（８）で計算することができる。

At this time, the density p _i is

Therefore, the estimated value E (m) of the number of terminals m can be calculated by the following equation (8).

As shown in the example of FIG. 16, the terminal a _i transmits the signals q _i1 , q _i2 , and q _i3 within the observation period and the period in which the terminal a _i stays in the sector S, and the signal q _i1 , The signal q _i0 is transmitted immediately before the signal q _i3 , and the signal q _i4 is transmitted immediately after the signal q _i3 , and the transmission times of the signals q _i0 , q _i1 , q _i2 , q _i3 , q _i4 are set to u _i0 , u _i1 , u, respectively. _{Assuming i2} , u _i3 , u _i4 , the above idea is that the stay time t _i of the sector S in the observation period of the terminal a _i is determined from (the midpoint of u _i0 and u _i1 ) (from u _i3 and u _i4 This is equivalent to estimating the period until the midpoint. The terminal a _i transmits the signal q _i4 while staying in the sector S, although it is not within the observation period. Just because, is not carried out to estimate the end time of the residence time t _i as the same as the end time of the observation period T. In this way, to maintain the impartiality of the estimated amount of time spent t _i.

[Terminal number estimation process]
Hereinafter, the terminal number estimation process according to the terminal number estimation method of the present invention will be described. Here, as an example, it is assumed that the position information included in the position data of the mobile terminal is given the sector number of the sector where the mobile terminal is located. Here, as the observation period information, a set of the observation start time T1 and the observation end time T2 is acquired in advance by the observation period acquisition unit 13, and as the observation area information, the sector number S is acquired in advance by the observation area acquisition unit 14. Shall be.

  As shown in FIG. 17, the position data acquisition unit 11 first acquires position data from the outside and stores it in the storage unit 12 (step S1 in FIG. 17). Thereby, the accumulation | storage part 12 will preserve | save the positional data over several time about many users (mobile terminal). In addition, after the process of step S1, you may perform the process after step S2 after time. That is, step S1 may be executed as a preliminary preparation for the processing after step S2.

  Next, the first location number estimation unit 15 reads the position data from the storage unit 12, and based on the position data, the number of terminals located in the observation area during the observation period is set as the first location number. Estimation is performed (step S2 in FIG. 17). Specific processing contents will be described later with reference to FIG.

  Next, the signal removal unit 16 reads the position data from the storage unit 12, removes the position registration information over the LA from the position data, and obtains position data after removal (step S3 in FIG. 17). Note that it is not essential for the signal removal unit 16 to read the position data from the storage unit 12, and the signal removal unit 16 obtains the position data read in step S2 of FIG. You may be handed over. In this way, the signal removal unit 16 may receive either the position data from the first serving number estimation unit 15 or read it from the storage unit 12, and logically either.

  Next, based on the position data after removal obtained by the signal removal unit 16, the second number-of-zones estimation unit 17 determines the number of terminals located in the observation area during the observation period as the second number of visited areas. (Step S4 in FIG. 17). Specific processing contents will be described later with reference to FIG.

[Example of first and second area estimation processing]
Hereinafter, an example of step S2 (first area number estimation process) in FIG. 1 and step S4 (second area number estimation process) in FIG. 17 will be described with reference to FIGS. 18 and 19. Needless to say, the estimation processing of the first and second number of locations is not limited to the processing content described below.

  As shown in FIG. 18, first, the observation target acquisition unit 31 in FIG. 14 uses a set of the observation start time T1 and the observation end time T2 as the observation period information from the observation period acquisition unit 13 as the observation area information. After acquiring S from the observation area acquisition unit 14, the storage unit 12 adds the position acquisition time information after the observation start time T1 and before the observation end time T2 to the sector number S that is the observation area information. One or more pieces of position data including position information to be associated (for example, position information is sector number S) are acquired as observation target position data (step S11 in FIG. 18). That is, the observation target acquisition unit 31 acquires position data that satisfies the following conditions as observation target position data. Condition 1: The position acquisition time is after the observation start time T1 and before the observation end time T2. That is, it is included in the observation period. Condition 2: The position information is sector S.

  Next, the following steps S12 and S13 are executed for each of the acquired observation target position data. In step S12, the front / rear position data acquisition unit 32 includes the same identification information as the first position data for the position data (first position data) for which the feature amount is to be obtained among the observation target position data. Of the position data, the position acquisition time information of the position data immediately before the first position data (second position data) and the position data immediately after the first position data (first 3 position data) is acquired. In addition, it is not essential for the front and rear position data acquisition unit 32 to acquire the entire second and third position data, and it is only necessary to acquire the position acquisition time information included in the second and third position data. The front / rear position data acquisition unit 32 may read the position acquisition time information of the second and third position data from the storage unit 12 or may pass it from the position data acquisition unit 11. Logically, either method can be used.

  In step S13, the feature amount calculator 33 calculates a feature amount for the first position data. The processing contents will be described with reference to FIG. Here, the position acquisition times of the first, second, and third position data are t1, t2, and t3, respectively. Further, a predetermined reference value (reference value relating to the difference between the position acquisition times of the first and second position data) serving as a reference for determining that the position acquisition time t2 of the second position data is an abnormal value is used as a reference. A value A (for example, 1 hour) is set, and a predetermined reference value (a position acquisition time of the first and third position data) is used as a reference for determining that the position acquisition time t3 of the third position data is an abnormal value. A reference value regarding the difference) is set as a reference value B (for example, 1 hour).

  The feature amount calculation unit 33 calculates the difference between the position acquisition times of the first and second position data (ie, the difference between the times t1 and t2) Da and the difference between the position acquisition times of the first and third position data (ie, The difference between the times t1 and t3) Db is calculated (step S31 in FIG. 19). Then, the feature amount calculation unit 33 determines whether or not the difference Da between the position acquisition times of the first and second position data is larger than a predetermined reference value A (for example, 1 hour) (step S32), If the difference Da is larger than the reference value A, the time acquired from the position acquisition time t1 of the first position data in the past by a predetermined time (for example, one hour) is set as the position acquisition time t2 of the second position data. (Step S33). Next, the feature amount calculation unit 33 determines whether or not the difference Db between the position acquisition times of the first and third position data is larger than a predetermined reference value B (for example, 1 hour) (step S34). If the difference Db is larger than the reference value B, the time that has been advanced in the future by a predetermined time (for example, 1 hour) from the position acquisition time t1 of the first position data is set to the position acquisition time t3 of the third position data. (Step S35). Then, the feature quantity calculation unit 33 calculates and stores the difference between the position acquisition time t2 of the second position data and the position acquisition time t3 of the third position data as the feature quantity for the first position data. (Step S36).

  Thus, the processes of steps S12 and S13 in FIG. 18 for one piece of observation target position data (first position data) are completed.

  Thereafter, the processes of steps S12 and S13 described above are executed for each of the observation target position data, and when the execution is completed for all the observation target position data (positive determination in step S14), the process proceeds to step S15.

In step S15, the number-of-terminals estimation unit 34 calculates the sum of the feature values w _ij for the observation target position data as shown in the above-described equation (8), and calculates the sum of the obtained feature values w _ij for the observation period. The numerical value obtained by dividing by the length T is estimated as the number of terminals. In this way, the first or second number in the area can be estimated.

  According to the example of the estimation process described above, when the number of terminals is estimated using position data, the effect of fluctuations in the reception interval is calibrated by performing correction using the acquisition time information of previous and subsequent position data. In addition, the number of terminals can be estimated with high accuracy. In addition, by performing the processing when the position acquisition times of the second and third position data described above are abnormal values in the feature amount calculation processing, the mobile terminal 100 is located outside the service area. When the position data acquisition time interval becomes abnormally long due to the power of the mobile terminal 100 being turned off or the like, the influence of the abnormally long acquisition time interval is prevented from being excessively increased. Can do.

As is clear from equation (8), the terminal number estimation unit 34 _divides each feature quantity w _ij for the observation target position data by 2 to obtain a sum of (feature quantity w _ij / 2), A numerical value obtained by dividing the obtained sum by the observation period length T may be estimated as the number of terminals. However, since the calculation method in which the sum of the feature quantities w _ij for the observation target position data is divided by twice the observation period length T as in the present embodiment, the number of divisions can be overwhelmingly smaller. There is an advantage that the load can be reduced.

Next, a modified example related to the feature amount will be described. In the embodiment described above, the time difference (time difference between the second position data and the third position data) before and after the position data (first position data) for which the feature amount is to be obtained is calculated using the first An example of calculating the feature value of the position data has been shown. When this is expressed by an equation, the feature amount can be expressed by the following equation (9). In addition, the following formula | equation (9) is only the deformation | transformation of the formula (6) mentioned above, and is equivalent to a formula (6) (namely, it is not what changed the way of thinking of a formula (6)).
w _ij = u _{i (j + 1)} −u _{i (j−1)} (9)
This modification shows another variation of the feature quantity calculation method calculated by the feature quantity calculator 33.

In the present modification, the feature amount calculation unit 33 determines the type information (for example, position data generation factors described later) about the second position data and the third position data when obtaining the feature amount of the first position data. (Generation timing)). Specifically, the feature amount calculation unit 33 calculates a correction coefficient α corresponding to the type information (generation factor here) of the third position data with respect to the time difference between the third position data and the first position data. A value obtained by multiplying the time difference between the first position data and the second position data by a correction coefficient β corresponding to the type information (generation factor in this case) of the second position data is calculated. calculate. However, in addition to the above, the feature amount calculation unit 33 may determine the correction coefficient α or β according to the type information of the first position data, and may also use the type information of the first and second position data. Accordingly, the correction coefficient β may be determined or the correction coefficient α may be determined according to the type information of the first and third position data. And the feature-value calculation part 33 makes the value which added the value obtained by these multiplications the feature-value of 1st position data. When the feature amount calculation processing in the feature amount calculation unit 33 is expressed by an equation, it is expressed by the following equation (10).
w _ij = α (u _{i (j + 1)} −u _ij ) + β (u _ij −u _{i (j−1)} ) (10)

  As the type information about the second position data and the third position data, for example, when the position data is position registration information, information on the generation factor of the position registration information can be cited. It is included in the generated location registration information. Factors for generating location registration information include that the terminal has crossed a location registration area (Location Area) boundary, that it has been generated based on location registration that is performed periodically, that the attachment process is performed by turning on the terminal, etc. For example, the detachment process is executed when the power is turned off. The set values of the correction coefficients α and β are determined in advance corresponding to these generation factors. Then, the feature amount calculation unit 33 sets the correction coefficient α for the third position data according to the information about the generation factor of the third position data, and sets the correction coefficient α for the third position data according to the information about the generation factor of the second position data. The correction coefficient β for the position data 2 may be set. Both the correction coefficients α and β may be set in advance to a value of 0 or more and 2 or less. However, this numerical range is not essential.

  For example, in the case of location registration information in which the location of the terminal and the generation opportunity of location registration information are irrelevant, such as location registration information based on location registration performed periodically, the expectation of the time spent in the current sector The value is considered to be the same before and after the location registration information is generated. On the other hand, in the case of location registration information generated by a terminal straddling a location registration area boundary, it can be determined that the terminal has not stayed in the current sector at least before the location registration information is generated. Therefore, the time that the terminal stayed in the current sector before the location registration information is generated is considered as 0, and the type information (generation factor) of the first location data is “location registration area boundary straddle” For example, the correction coefficient β in the above equation (10) (that is, the correction coefficient β related to the time difference from the previous position data) can be set to zero. As a result, it is possible to calculate a feature amount that is more realistic.

  As described above, when the feature amount calculation unit 33 calculates the feature amount for the target position data (first position data), the second and third positions which are position data before and after the first position data. The time difference between the second position data and the third position data is corrected according to the type information about the data (for example, the generation factor of the position data), and the feature amount is calculated using the corrected time difference. Thereby, the feature amount can be calculated with higher accuracy based on the type information of the position data.

[Operational effects of terminal estimation device 10]
Next, operational effects of the terminal number estimation device 10 in the present embodiment, application examples, and modifications will be described.

  In the terminal number estimation apparatus 10 of the present embodiment, the cluster generation unit 19 selects a target sector for cluster configuration based on the target sector selected by the target sector selection unit 18. Then, the first visited area estimation unit 15 estimates the number of terminals in the cluster composed of the target sector and the target sector as the first visited area based on the location registration signal, and the second visited area. The number estimator 17 estimates the number of terminals in the cluster as the second number of existing areas, excluding the LA straddling, based on the position registration signal. The cluster generation unit 19 determines a determined sector that constitutes the cluster based on the estimated first and second serving numbers, and forms a cluster from the determined sector and the sector of interest. . As a result, it is possible to generate a cluster which is a population aggregation unit for estimating the number of located areas with higher accuracy.

  Further, in the terminal number estimation device 10 of the present embodiment, the adjacent sector extraction unit 19a extracts a set of adjacent sectors adjacent to the target sector based on the target sector. The candidate sector extractor 19c extracts the candidate sector 19b that has been determined by the adjacent sector extractor 19a not to exceed the predetermined geographical extent as a candidate sector. Then, the sector determination unit 19d selects a target sector from the extracted candidate sectors. In this selection, an optimum target sector is selected based on the first number of visited areas and the second number of visited areas.

  Then, the determined sector is determined based on the difference between the first number of areas and the second number of areas existing in the cluster composed of the target sector and the target sector. As a result, it is possible to generate a cluster which is a population aggregation unit for estimating the number of located areas with higher accuracy. In particular, the accuracy can be improved by limiting the sectors constituting the cluster based on geographical spread.

  Further, in the terminal number estimation apparatus 10 of the present embodiment, the sector determination unit 19d determines, as a target sector, a sector having a minimum ratio of the difference with respect to the sum of the first and second serving numbers. By constructing a cluster from the target sector and the target sector, the accuracy of the number of serving areas can be further increased.

  Further, in the terminal number estimation apparatus 10 of the present embodiment, when the candidate sector extraction unit 19c cannot extract a candidate sector, the reference value for determining the geographical spread in the determination unit 19b is changed, for example, the reference value is gradually reduced. Thus, by adding a predetermined step value to the threshold value d and updating it, the adjacent sector extraction unit 19a re-extracts the adjacent sector, thereby making it easier to extract candidate candidate sectors and forming a cluster. The valid sectors can be extracted.

  Further, in the terminal number estimation apparatus 10 in the application example of the present embodiment, when the sector includes a narrow sector, the adjacent sector extraction unit 19a does not select the narrow sector as the target sector, and determines the sector. When the target sector or target sector includes a narrow sector smaller than the target sector or target sector, the unit 19d adds the number of terminals in the narrow sector as the first serving number, When the target sector includes the sector of interest or a narrow sector smaller than the target sector, the sum of the number of terminals in the narrow sector is used as the second terminal number, for example, an outdoor station and an indoor station Even if the sector is configured as described above, an appropriate target sector can be extracted.

  For example, the sector of interest or the target sector is a sector formed by a base station arranged outdoors, and the narrow sector is a sector formed by a base station arranged indoors.

  Further, in the terminal number estimation device 10 in the modification of the present embodiment, the sector of interest and the target sector are mesh regions that are defined in advance, and the mesh region determination unit 19s includes the first number-of-services estimation unit 15 and the The number of terminals estimated in units of sectors by the number-of-zones estimation unit 17 of 2 is apportioned by the area ratio between the sector and the mesh region, and the number of terminals in each mesh region is set as the first number of regions and the number of second regions. Based on the target mesh region including the target sector, the target mesh region as the target sector constituting the cluster is determined, and a cluster including the target mesh region and the target mesh region is generated. Thereby, a cluster can be comprised per mesh area | region and an application range can be expanded more.

  In this modification, more specifically, the adjacent mesh region extraction unit 19p extracts an adjacent mesh region adjacent to the target mesh region extracted by the target mesh region extraction unit 19o. Then, the candidate mesh region extraction unit 19r extracts a candidate mesh region from the adjacent mesh region adjacent to the target mesh region, and the mesh region determination unit 19s determines the target mesh region from the candidate mesh region.

DESCRIPTION OF SYMBOLS 10 ... Terminal number estimation apparatus, 11 ... Position data acquisition part, 12 ... Accumulation part, 13 ... Observation period acquisition part, 14 ... Observation area acquisition part, 15 ... 1st in-zone number estimation part, 16 ... Signal removal part, Reference numeral 17: Second number-of-services estimation unit, 18: Target sector selection unit, 19 ... Cluster generation unit, 19a ... Neighboring sector extraction unit, 19b ... Judgment unit, 19c ... Candidate sector extraction unit, 19d ... Sector determination unit, 19o The target mesh region extraction unit, 19p ... Neighboring mesh region extraction unit, 19q ... Determination unit, 19r ... Candidate mesh region extraction unit, 19s ... Mesh region determination unit, 19x ... Cluster generation unit, 20 ... Output unit, 31 ... Observation target Acquiring unit, 32 ... Front and rear position data acquiring unit, 33 ... Feature amount calculating unit, 34 ... Terminal number estimating unit, 100 ... Mobile terminal, 400 ... Exchange, 400 ... Each exchange, 400 ... Exchange, 500 ... Management center 501 ... social sensor unit, 502 ... peta-mining unit, 503 ... mobile demography unit, 504 ... visualization solution unit, 700 ... various processing nodes.

Claims (9)

Selection means for selecting a target area for cluster configuration based on a pre-designated area of interest;
First estimation means for estimating the number of terminals in a cluster composed of the target area and the target area as the number of first terminals based on a position registration signal;
Second estimation means for estimating the number of terminals in the cluster as the number of second terminals based on the position registration information obtained by removing the position registration information generated due to the terminal straddling the position registration area;
Based on the number of first terminals estimated by the first estimation means and the number of second terminals estimated by the second estimation means, it is determined that the target area is a decision area that constitutes a cluster. A cluster generating means for forming a cluster from the determined area and the area of interest;
An information processing apparatus comprising: The selection means includes
An adjacent area extracting means for extracting a set of adjacent areas adjacent to the target area based on the target area;
If the cluster to which the adjacent area extracted by the extracting means is added does not exceed a predetermined geographical spread, candidate area extracting means for extracting the adjacent area as a candidate area;
Have
The selection means includes
A target area is selected from the candidate areas extracted by the candidate area extracting means;
The cluster generation means includes
Based on the difference between the number of the first terminals and the number of the second terminals existing in the cluster composed of the target area selected by the selection means and the target area, the target area is determined to be a determination area. The information processing apparatus according to claim 1. The selection means includes
The candidate area selected from the candidate areas extracted by the candidate area extracting unit is selected as a target area having a minimum ratio of the difference with respect to the sum of the number of the first terminals and the number of the second terminals. 2. The information processing apparatus according to 2. When the candidate area extraction means cannot extract a candidate area, the reference value for judging the geographical spread is changed,
The information processing apparatus according to claim 2, wherein the adjacent area extracting unit re-extracts the adjacent area. When the selection means includes a narrow area in at least one of the area of interest or the target area, the selection means does not select the narrow area as the target area,
In the case where the target area or the target area includes a narrow area smaller than the target area or the target area, the first estimating means adds the number of terminals in the narrow area as the first terminal number. ,
When the target area or the target area includes a narrow area smaller than the target area or the target area, the second estimation means adds the number of terminals in the narrow area to the second terminal number. The information processing apparatus according to any one of claims 1 to 4, wherein the information processing apparatus is configured as follows. The area of interest or the target area is an area formed by base stations arranged outdoors,
The information processing apparatus according to claim 5, wherein the narrow area is an area formed by a base station arranged indoors. The area of interest and the target area are predefined mesh areas,
The first estimation means and the second estimation means calculate the number of terminals obtained by apportioning the number of terminals estimated for each sector by the area ratio between the sector and the mesh area for each mesh area, and are configured for each mesh area. Estimate the number of first and second terminals in the cluster,
The selection means selects a target mesh region as a target area constituting a cluster based on the target mesh region including the target area,
The cluster generation means determines that the target mesh region is a determined mesh region constituting a cluster, and generates a cluster including the target mesh region and the target mesh region. The information processing apparatus according to any one of claims 1 to 6.   The information processing apparatus according to claim 7, wherein the cluster generation unit determines the target mesh region based on an adjacent mesh region adjacent to the target mesh region. A selection step of selecting a target area for cluster configuration based on a pre-designated area of interest;
A first estimation step of estimating the number of terminals in the cluster composed of the target area and the target area as the number of first terminals based on the position registration signal;
A second estimation step of estimating the number of terminals in the cluster as a second number of terminals, excluding LA straddling, based on the location registration signal;
Based on the number of first terminals estimated in the first estimation step and the number of second terminals estimated in the second estimation step, it is determined that the target area is a determination area that constitutes a cluster. A cluster generation step of forming a cluster from the determined area and the area of interest;
A cluster generation method comprising:

Images