JP6972667B2 - Regional characteristic prediction method, regional characteristic prediction device and regional characteristic prediction program - Google Patents

Description

The present invention relates to a regional characteristic prediction method, a regional characteristic prediction device, and a regional characteristic prediction program.

When predicting temporal changes in a region, such as the flow of people and changes in nature, the smaller the target area, the more the influence of the surrounding area cannot be ignored. Therefore, conventionally, a method using a spatial weight matrix is known in metric statistics.

Japanese Unexamined Patent Publication No. 2012-141953

Some spatial weight matrices are determined by whether or not the regions are adjacent, and some are determined by whether or not they are within a certain distance range. In either case, it depends on whether or not they are adjacent to a certain region. The effect is expressed isotropically in the spatial weight matrix.

However, even if they are adjacent to a certain area, each adjacent area is not necessarily affected to the same extent. Therefore, even if a spatial weighting matrix that isotropically expresses the influence of adjacent regions is used, it may not be possible to accurately predict temporal changes in the regions.

An object of the present invention is to provide a regional characteristic prediction method, a regional characteristic prediction device, and a regional characteristic prediction program capable of accurately predicting regional characteristics.

In one embodiment, the regional characteristic prediction method refers to a first storage unit that stores map information to generate a spatial weighting matrix that indicates the adjacency between the subdivided regions contained in the map information, and maps. The spatial weight matrix is corrected, and the corrected spatial weight matrix and the map are referred to with reference to the second storage unit that stores the information of the elements that affect the connection between the subdivided regions in association with the information. The computer executes a process of predicting the characteristics of the prediction target based on the information and the data related to the characteristics of the prediction target associated with the information.

As one aspect, it is possible to accurately predict the characteristics of a region.

It is a figure which shows schematic the hardware composition of the area characteristic prediction apparatus which concerns on 1st Embodiment. It is a functional block diagram of a regional characteristic prediction device. FIG. 3A is a diagram showing an example of mesh data included in the mesh map DB, and FIG. 3B is a diagram showing an example of the center coordinates of the mesh. FIG. 4A is a diagram showing route superimposed mesh data included in the route DB, and FIG. 4B is a diagram showing route data included in the route DB. It is a figure which shows an example of attribute data. It is a flowchart which shows the processing of a region characteristic prediction apparatus. 7 (a) is a diagram showing nine meshes used in the explanation of FIG. 6, and FIG. 7 (b) is a diagram showing a spatial weight matrix in the mesh of FIG. 7 (a). 8 (a) is a diagram showing line superposition mesh data used in the explanation of FIG. 6, and FIG. 8 (b) is a diagram showing a spatial weight matrix corrected using FIG. 8 (a). FIG. 9A is a graph showing the prediction result in the comparative example, and FIG. 9B is a graph showing the prediction result in the first embodiment. It is a graph which shows the measured value. It is a table which shows the result of having compared FIG. 10 with FIG. 9 (a) and FIG. 9 (b). It is a figure which shows the mesh data which superposed the river used in the 2nd Embodiment. FIG. 13A is a graph showing the prediction result in the comparative example, and FIG. 13B is a graph showing the prediction result in the second embodiment. It is a graph which shows the measured value in 2nd Embodiment. It is a figure which shows the modification.

<< First Embodiment >>
Hereinafter, the first embodiment of the regional characteristic prediction device will be described in detail with reference to FIGS. 1 to 11.

FIG. 1 schematically shows the hardware configuration of the regional characteristic prediction device 10. The regional characteristic prediction device 10 is, for example, a PC (Personal Computer), and as shown in FIG. 1, a CPU (Central Processing Unit) 90, a ROM (Read Only Memory) 92, a RAM (Random Access Memory) 94, and a storage unit ( Here, an HDD (Hard Disk Drive) 96, a network interface 97, a display unit 93, an input unit 95, a portable storage medium drive 99, and the like are provided. Each component of the regional characteristic prediction device 10 is connected to the bus 98. In the regional characteristic prediction device 10, a program stored in the ROM 92 or the HDD 96 (including the regional characteristic prediction program) or a program read from the portable storage medium 91 by the portable storage medium drive 99 (including the regional characteristic prediction program). ) Is executed by the CPU 90, so that the functions of each part shown in FIG. 2 are realized. Note that FIG. 2 also shows various DBs and data stored in the HDD 96 and the like.

FIG. 2 is a functional block diagram showing a function realized by the CPU 90 of the regional characteristic prediction device 10 executing a program. As shown in FIG. 2, by executing the program by the CPU 90, the functions as the spatial weight matrix generation unit 12, the correction unit 14, and the prediction unit 16 as the generation unit are realized.

The spatial weight matrix generation unit 12 generates a spatial weight matrix with reference to the mesh data stored in the mesh map DB 22 as the first storage unit. Here, the mesh map DB 22 stores mesh data as shown in FIG. 3A. The mesh data is information about a rectangular mesh arranged on a map, for example, 1 km in length and width, and includes the dimensions of each mesh and the center coordinates (X, Y) of each mesh. The area included in the map is subdivided into a plurality of areas by each mesh. The spatial weight matrix generation unit 12 acquires the center coordinates of each mesh as shown in FIG. 3B from the mesh data. Then, the adjacency relationship of each mesh is specified based on the distance between the center coordinates, and the spatial weight matrix is generated based on the specified adjacency relationship (whether or not they are adjacent).

The correction unit 14 corrects the spatial weight matrix with reference to the line DB 24 as the second storage unit. Here, the route DB 24 includes route superimposed mesh data as shown in FIG. 4A and route data as shown in FIG. 4B. As shown in FIG. 4A, the line superimposition mesh data is data in which the mesh data and the line arrangement are superposed. In FIG. 4A, the two lines are superimposed on the mesh data. The route data is data generated from the route superimposition mesh data of FIG. 4 (a), and has a field of "route name" and a field of "mesh ID" as shown in FIG. 4 (b). There is. In the route data, the name of the route and the identifier (mesh ID) of the mesh through which the route passes are associated with each other.

The correction unit 14 corrects the elements of the spatial weight matrix corresponding to these two adjacent meshes when the same line passes through the two adjacent meshes. For example, the elements of the spatial weight matrix corresponding to the two meshes are multiplied by α (α is a value larger than 1). By making such a correction, the magnitude (strength) of the connection between the meshes due to the existence of the line can be expressed in the spatial weight matrix.

The prediction unit 16 predicts data regarding the characteristics of each region (for example, the number of single households in each mesh) based on the spatial weight matrix corrected by the correction unit 14 and the attribute data 26. Here, FIG. 5 shows data on the number of single households in each mesh in each year as an example of the attribute data 26. The prediction unit 16 predicts, for example, the number of single households in each mesh of the next year by using the attribute data 26 of FIG. It can be said that the attribute data 26 is data in which information related to the characteristics (characteristics of the prediction target) predicted by the prediction unit 16 is associated with map information (each mesh).

(Regarding the processing of the regional characteristic prediction device 10)
Next, the processing of the regional characteristic prediction device 10 will be described in detail with reference to other drawings as appropriate along the flowchart of FIG. As a premise that the processing of FIG. 6 is performed, it is assumed that the mesh map DB 22 and the route DB 24 store information about the area for which the characteristics are to be predicted. Further, it is assumed that the attribute data 26 has already been input by the user via the input unit 95 or is included as a dbf file in the mesh map DB 22. In the description of this process, for the sake of simplification of the explanation, the mesh data is the data as shown in FIG. 7 (a), and the route superimposition mesh data is the data as shown in FIG. 8 (a). do.

In the process of FIG. 6, first, in step S12, the spatial weight matrix generation unit 12 acquires the center coordinates of the mesh. Specifically, the mesh ID, which is an identifier of each mesh, and the center coordinates (X, Y) are acquired from the mesh data of FIG. 7 (a) (see FIG. 3 (b)).

Next, in step S14, the spatial weight matrix generation unit 12 determines the adjacency relationship between the meshes. In the present embodiment, when the distance between the center coordinates of the two meshes matches the dimensions of the meshes, it is determined that the two meshes are adjacent to each other. For example, when there are nine meshes 1 to 9 as shown in FIG. 7A, the meshes adjacent to the mesh 5 are 2, 4, 6, and 8. The number of meshes adjacent to mesh 1 is 2 and 4. In the above description, the space weight matrix generation unit 12 has described the case where the adjacency relationship between the meshes is determined based on the distance between the center coordinates of the two meshes, but the present invention is not limited to this. For example, the spatial weight matrix generation unit 12 may determine whether or not the two meshes are in contact with each other by a line, and if they are in contact with each other by a line, it may be determined that they are adjacent to each other.

Next, in step S16, the spatial weight matrix generation unit 12 generates a spatial weight matrix. Specifically, the spatial weight matrix generation unit 12 represents the adjacency relationship between the meshes determined in step S14 as a matrix. In the case of FIG. 7 (a), the spatial weight matrix generation unit 12 generates the spatial weight matrix as shown in FIG. 7 (b). In the spatial weight matrix of FIG. 7B, for example, for row "1", the values of the matrix elements in the columns "2" and "4" of the mesh numbers adjacent to mesh 1 are set to 1, and the others are set to 0. It is supposed to be. Further, for example, for row "2", the values of the matrix elements in the columns "1", "3", and "5" of the mesh numbers adjacent to mesh 2 are set to 1, and the others are set to 0. Hereinafter, the values of each matrix element are set in the same manner.

When the process of step S16 is completed, the process of correcting the spatial weight matrix is executed by the correction unit 14 in the subsequent steps S18 to S30.

When the process proceeds to step S18, the correction unit 14 identifies an unspecified mesh. For example, it is assumed that the correction unit 14 specifies the mesh 1 in FIG. 7 (a).

Next, in step S20, the correction unit 14 determines whether or not the route passes through the specified mesh. In this case, the correction unit 14 refers to the line superimposition mesh data of FIG. 8A, and determines whether or not the line passes through the mesh 1. In the example of FIG. 8A, since the route does not pass through the specified mesh 1, the determination in step S20 is denied, and the process proceeds to step S30. When the process proceeds to step S30, the correction unit 14 determines whether or not all the meshes have been specified. If the determination in step S30 is denied, the process returns to step S18. Returning to step S18, the unspecified mesh is specified. Here, for example, it is assumed that the mesh 2 is specified.

Next, when the process proceeds to step S20, it is determined whether or not the route passes through the specified mesh 2. In the example of FIG. 8A, since the line passes through the mesh 2, the determination in step S20 is affirmed, and the process proceeds to step S22.

When the process proceeds to step S22, the correction unit 14 selects an unselected adjacent mesh. Here, one of meshes 1, 3, and 5 adjacent to mesh 2 (for example, mesh 1) is selected.

Next, in step S24, the correction unit 14 determines whether or not the same line passes through the selected adjacent mesh. Here, since the same route as the mesh 2 specified in step S18 does not pass through the adjacent mesh 1 currently selected, the determination in step S24 is denied and the process proceeds to step S28. In step S28, the correction unit 14 determines whether or not all adjacent meshes have been selected, but here, since only mesh 1 of meshes 1, 3 and 5 is selected, the determination is denied. , Return to step S22.

Returning to step S22, the correction unit 14 selects, for example, mesh 3 as an unselected adjacent mesh. Next, in step S24, the correction unit 14 determines whether or not the same line as the mesh 2 specified in step S18 passes through the selected adjacent mesh 3. Here, the same line as the mesh 2 runs through the adjacent mesh 3 currently selected. Therefore, in this case, the determination in step S24 is affirmed, and the process proceeds to step S26.

When the process proceeds to step S26, the correction unit 14 integrates α (α> 1) into the corresponding matrix element. That is, the correction unit 14 sets α (= 1 × α) to the matrix element “1” in the second row from the top and the third column from the left in FIG. 7 (b) (see FIG. 8 (b)).

Next, in step S28, the correction unit 14 determines whether or not all adjacent meshes have been selected, but here, only meshes 1 and 3 out of meshes 1, 3 and 5 have been selected. The judgment is denied, and the process returns to step S22.

Returning to step S22, the correction unit 14 selects, for example, mesh 5 as an unselected adjacent mesh. Next, in step S24, the correction unit 14 determines whether or not the same line as the mesh 2 passes through the selected adjacent mesh 5. Here, the same line as the mesh 2 runs through the adjacent mesh 5 currently selected. Therefore, in this case, the determination in step S24 is affirmed, and the process proceeds to step S26.

When the process proceeds to step S26, the correction unit 14 integrates α (α> 1) into the corresponding matrix element. That is, the correction unit 14 sets α (= 1 × α) to the matrix element “1” in the second row from the top and the fifth column from the left in FIG. 7 (b) (see FIG. 8 (b)).

After that, the process proceeds to step S28, but since all the adjacent meshes have been selected, the judgment of step S28 is affirmed, and the process proceeds to step S30. When the process proceeds to step S30, as described above, the correction unit 14 determines whether or not all the meshes have been specified, but if the determination here is denied, the process returns to step S18.

After that, the correction unit 14 sequentially identifies the meshes 3 to 9 (S18), sequentially selects the meshes adjacent to the specified meshes (S22), and the same line passes through the specified mesh and the selected adjacent meshes. It is determined whether or not it is (S24). Then, when the same line passes, the spatial weight matrix is corrected by integrating α into the corresponding matrix elements (S26).

In the case of the example of FIG. 8A, the spatial weight matrix is corrected as shown in FIG. 8B.

As described above, when all the meshes have been specified and the judgment of step S30 is affirmed, the process proceeds to step S32, and the prediction unit 16 uses the attribute data 26 and the corrected spatial weight matrix to determine the characteristics of the region. Predict. The prediction unit 16 has a function of performing Bayesian estimation using, for example, a programming language for statistical modeling (for example, Stan), and predicts the characteristics of the region based on the attribute data 26 and the corrected spatial weight matrix. To execute. For example, when the attribute data 26 is data showing the transition of the number of single households in each mesh in the past, the prediction unit 16 uses the attribute data 26 and the corrected spatial weight matrix, for example, the next year in each mesh. Predict the number of single households. In this case, when the same line passes between two adjacent meshes, the value of the matrix element of the spatial weight matrix corresponding to the two meshes is greatly corrected, so that the prediction unit 16 has corrected the corrected value. By using the spatial weight matrix, it is possible to predict the characteristics of the region in consideration of the influence between the meshes due to the existence of the line (for example, the distribution of people tends to spread along the line). ..

As described above, when the processing up to step S32 is completed, all the processing in FIG. 6 is completed.

As shown in FIG. 4A, a plurality of routes may pass through the plurality of meshes to be analyzed. In such a case, the routes of interest may be different, and the processes of steps S18 to S30 in FIG. 6 may be repeated for the routes. At this time, α may be different for each line. For example, it may be possible to increase α for conventional lines and decrease α for lines such as the Shinkansen, which are less likely to affect areas.

FIG. 9 (a) shows the data of the number of single households in each mesh set in a certain municipality as shown in FIG. 3 (a) from 1995 (1995) to 2005 (2005) (FIG. 5). Attribute data 26) and the result of predicting the number of single households in each mesh in 2010 (2010) using the spatial weight matrix (without correction) (prediction result only in 2010, others are attribute data 26) Measured values included) are shown. Further, in FIG. 9 (b), the number of single households in each mesh in 2010 (2010) is shown in the same manner as in FIG. 9 (a) by using the spatial weight matrix corrected as in the first embodiment. The predicted results are shown. Further, FIG. 10 shows an actually measured value of the number of single households in each mesh.

11 shows the difference between the predicted value of FIG. 9A and the measured value of FIG. 10 (difference: comparative example), and the difference between the predicted value of FIG. 9B and the measured value of FIG. 10 (difference:). This application) is shown. From the table of FIG. 11, the difference: the average of the absolute values of the comparative examples was 78.8, and the difference: the average of the absolute values of the present application was 27.0. As a result, it was found that the number of single households can be accurately predicted by using the spatial weighting matrix corrected as in the first embodiment.

As described above in detail, according to the first embodiment, the spatial weight matrix generation unit 12 refers to the mesh map DB 22 and indicates the spatial weight matrix showing the adjacency relationship between the subdivided regions (mesh). (S16), and the correction unit 14 corrects the spatial weight matrix with reference to the line DB 24 (S18 to S30). Then, the prediction unit 16 predicts the characteristics of the region based on the corrected spatial weight matrix and the attribute data. As a result, in the first embodiment, it is possible to predict the characteristics of the region from the attribute data in consideration of not only the adjacency relationship between the meshes but also the influence (asymmetry) of the route on the connection between the adjacent meshes. can. Therefore, it is possible to accurately predict the characteristics of the region.

Further, in the first embodiment, the distribution of people is along the line in order to correct the spatial weight matrix so that the connection between the meshes through which the line passes is strong, that is, the weight between the meshes through which the line passes is large. It is possible to accurately predict the characteristics of the area in consideration of the fact that it tends to spread to the area.

In the first embodiment, the case where the matrix element is corrected based on whether or not the same route passes between the meshes has been described, but the present invention is not limited to this, and other transportation networks (for example, bus operation) are described. Matrix elements may be corrected based on whether a route (such as a route or a highway) passes between meshes. Further, the matrix element may be corrected based on a plurality of transportation networks, and in that case, the correction value (α) may be changed for each type of transportation network. Furthermore, the tendency for the distribution of people to spread along the route is higher in the city center, and the tendency for the distribution of people to spread along the trunk road is higher in the suburbs. The value (α) may be adjusted.

In addition, for example, if the length of the line passing through the mesh is specified and the accuracy of the prediction result of the mesh that the line passes only slightly is low, the length of the line is set so that the accuracy of the prediction result is improved. Α may be adjusted accordingly. Further, if the length of the line passing through the mesh is less than a predetermined length, the correction may not be performed.

<< Second Embodiment >>
Next, the second embodiment will be described with reference to FIGS. 12 to 14.

In the first embodiment described above, a case where the spatial weight matrix is corrected so that the connection between adjacent meshes passing through the same line is strengthened has been described, but in the second embodiment, the division of the area by the river has been described. An example of correcting the spatial weight matrix in consideration of the influence of (elements that hinder traffic) will be described.

In the second embodiment, the correction unit 14 causes the value of the matrix element corresponding to the two meshes to be multiplied by β (β <1) when the river flows between the two adjacent meshes. to correct. For example, in the mesh data in which rivers are superimposed as shown in FIG. 12, the number of meshes adjacent to mesh 5 is 2, 4, 6, 8, but there is a river between mesh 5 and mesh 2, 6. However, the mesh is divided. Therefore, the values of the matrix elements corresponding to meshes 5, 2 and meshes 5, 6 are multiplied by β. The same correction is performed for other meshes. Then, the prediction unit 16 predicts the characteristics of the region (for example, the number of unemployed people) using the corrected spatial weight matrix.

In FIG. 13 (a), the data of the number of unemployed people (1995 to 2005) in each mesh set in a certain municipality and the space weight matrix (without correction) are used to show the total unemployment in 2010. The result of predicting the number of people (comparative example) is shown. Further, in FIG. 13 (b), the result of predicting the number of unemployed people in 2010 using the same data as in the case of FIG. 13 (a) and the corrected spatial weight matrix (second implementation). Form) is shown. Further, FIG. 14 shows the transition (actual measurement value) of the number of unemployed people from 1995 to 2010.

From these results, as in the first embodiment, the average value of the absolute value of the difference between the predicted value and the measured value of the comparative example is 10.0, which is 10.0, which is the predicted value and the measured value of the second embodiment. The average value of the absolute values of the difference from and was 9.6. That is, by correcting the spatial weight matrix as in the second embodiment, the prediction accuracy was improved by 4%.

As described above, according to the second embodiment, information on an element (for example, a river) that hinders traffic between meshes is used as an element that affects the connection between meshes, and there is a river between meshes. Since the matrix elements are corrected so as to weaken the connection between the meshes, it is possible to accurately predict the characteristics of the region in consideration of the factors that hinder the traffic between the meshes.

In the above embodiment, the case of predicting the number of unemployed people as a characteristic of the region has been described, but the present invention is not limited to this, and for example, as a characteristic of the region, it is also possible to predict the habitat of vegetation and organisms. good.

In the above embodiment, the case where the element that hinders the traffic between meshes is a river has been described as an example, but the present invention is not limited to this, and the element that hinders the traffic between meshes is, for example, a dam, an airfield, a base, or the like. There may be.

In the first and second embodiments, the case where the prediction unit 16 predicts the characteristics of the region by Bayesian estimation has been described, but the present invention is not limited to this, and the characteristics of the region may be predicted by other methods. good.

The first and second embodiments may be combined. That is, when the traffic network connects the meshes, the value of the matrix element is corrected to be large, and when there is an element such as a river between the meshes that obstructs the traffic, the value of the matrix element is reduced. It may be corrected to.

In the first and second embodiments, the case where the regional characteristic prediction device 10 has each function of FIG. 2 has been described, but the present invention is not limited to this. For example, each function of FIG. 2 may be possessed by a cloud server 110 connected to a network 80 such as the Internet as shown in FIG. In this case, the cloud server 110 may receive the data input from the user terminal 70 and execute the process of FIG. 6 on the cloud server 110. Then, the processing result (prediction result) is transmitted from the cloud server 110 to the user terminal 70, and the processing result received by the user terminal 70 is used. The cloud server 110 may be installed in Japan or overseas.

The above processing function can be realized by a computer. In that case, a program that describes the processing content of the function that the processing device should have is provided. By executing the program on a computer, the above processing function is realized on the computer. The program describing the processing content can be recorded on a computer-readable recording medium (however, the carrier wave is excluded).

When a program is distributed, it is sold in the form of a portable recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) on which the program is recorded. It is also possible to store the program in the storage device of the server computer and transfer the program from the server computer to another computer via the network.

The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes the processing according to the program. The computer can also read the program directly from the portable recording medium and execute the processing according to the program. Further, the computer can also sequentially execute the processing according to the received program each time the program is transferred from the server computer.

The embodiments described above are examples of preferred embodiments of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the gist of the present invention.

Regarding the above description of the first and second embodiments, the following additional notes will be further disclosed.
(Appendix 1) With reference to the first storage unit for storing the map information, a spatial weight matrix showing the adjacency relationship between the subdivided areas included in the map information is generated.
The spatial weight matrix is corrected by referring to the second storage unit that stores the information of the elements that affect the connection between the subdivided regions in association with the map information.
The characteristics of the prediction target are predicted based on the corrected spatial weight matrix and the data related to the characteristics of the prediction target associated with the map information.
A regional characteristic prediction method characterized by a computer performing processing.
(Appendix 2) The regional characteristic prediction method according to Appendix 1, wherein the information of the elements affecting the connection between the subdivided regions includes information on the transportation network.
(Supplementary Note 3) The regional characteristic prediction method according to Supplementary note 2, wherein in the correction process, the spatial weight matrix is corrected so that the weight between the areas through which the transportation network passes becomes large.
(Appendix 4) The description in any of the appendices 1 to 3, wherein the information of the elements that influence the connection between the subdivided regions includes the information of the elements that hinder the traffic between the regions. Regional characteristic prediction method.
(Supplementary Note 5) The regional characteristic prediction method according to Supplementary note 4, wherein in the correction process, the spatial weight matrix is corrected so that the weights between regions where traffic is hindered become small.
(Appendix 6) With reference to the first storage unit for storing map information, a generation unit for generating a spatial weight matrix indicating an adjacency relationship between subdivided regions included in the map information, and a generation unit.
A correction unit that corrects the spatial weight matrix with reference to a second storage unit that stores information on elements that affect the connection between the subdivided regions in association with the map information.
A prediction unit that predicts the characteristics of the prediction target based on the corrected spatial weight matrix and data related to the characteristics of the prediction target associated with the map information.
Regional characteristic prediction device equipped with.
(Appendix 7) The regional characteristic prediction device according to Appendix 6, wherein the information of the elements affecting the connection between the subdivided regions includes information on the transportation network.
(Supplementary Note 8) The regional characteristic prediction device according to Supplementary note 7, wherein the correction unit corrects the spatial weight matrix so that the weight between areas through which the transportation network passes becomes large.
(Appendix 9) The description in any of the appendices 6 to 8, wherein the information of the elements that influence the connection between the subdivided regions includes the information of the elements that hinder the traffic between the regions. Regional characteristic predictor.
(Supplementary Note 10) The regional characteristic prediction device according to Supplementary note 9, wherein the correction unit corrects the spatial weight matrix so that the weights between regions where traffic is hindered become small.
(Appendix 11) With reference to the first storage unit for storing the map information, a spatial weight matrix showing the adjacency relationship between the subdivided areas included in the map information is generated.
The spatial weight matrix is corrected by referring to the second storage unit that stores the information of the elements that affect the connection between the subdivided regions in association with the map information.
The characteristics of the prediction target are predicted based on the corrected spatial weight matrix and the data related to the characteristics of the prediction target associated with the map information.
A regional characteristic prediction program that allows a computer to perform processing.

10 Regional characteristic prediction device 12 Spatial weight matrix generation unit (generation unit)
14 Correction unit 16 Prediction unit 22 Mesh map DB (1st storage unit)
24 Line DB (2nd storage)
26 Attribute data (data related to the characteristics of the forecast target)

Claims (8)

With reference to the first storage unit for storing the map information, a spatial weight matrix showing the adjacency relationship between the subdivided regions included in the map information is generated.
The spatial weight matrix is corrected by referring to the second storage unit that stores the information of the elements that affect the connection between the subdivided regions in association with the map information.
The characteristics of the prediction target are predicted based on the corrected spatial weight matrix and the data related to the characteristics of the prediction target associated with the map information.
A regional characteristic prediction method characterized by a computer performing processing. The regional characteristic prediction method according to claim 1, wherein the information on the factors that influence the connection between the subdivided regions includes information on the transportation network. The regional characteristic prediction method according to claim 2, wherein in the correction process, the spatial weight matrix is corrected so that the weight between the areas through which the transportation network passes becomes large. The region according to any one of claims 1 to 3, wherein the information of the element that influences the connection between the subdivided regions includes the information of the element that hinders the traffic between the regions. Characteristic prediction method. The regional characteristic prediction method according to claim 4, wherein in the correction process, the spatial weight matrix is corrected so that the weight between regions where traffic is hindered becomes small. In the prediction process, the data regarding the predetermined characteristics of the region to be predicted is predicted based on the corrected spatial weight matrix and the data regarding the predetermined characteristics of each region associated with the map information. The regional characteristic prediction method according to any one of claims 1 to 5, wherein the method is characterized by. With reference to the first storage unit that stores the map information, a generation unit that generates a spatial weight matrix indicating the adjacency relationship between the subdivided regions included in the map information, and a generation unit.
A correction unit that corrects the spatial weight matrix with reference to a second storage unit that stores information on elements that affect the connection between the subdivided regions in association with the map information.
A prediction unit that predicts the characteristics of the prediction target based on the corrected spatial weight matrix and data related to the characteristics of the prediction target associated with the map information.
Regional characteristic prediction device equipped with. With reference to the first storage unit for storing the map information, a spatial weight matrix showing the adjacency relationship between the subdivided regions included in the map information is generated.
The spatial weight matrix is corrected by referring to the second storage unit that stores the information of the elements that affect the connection between the subdivided regions in association with the map information.
The characteristics of the prediction target are predicted based on the corrected spatial weight matrix and the data related to the characteristics of the prediction target associated with the map information.
A regional characteristic prediction program that allows a computer to perform processing.

Images