Disaster information providing system, disaster information providing method, and disaster information providing program

Abstract

To appropriately calculate an estimated value of a damage scale due to a disaster even when there is a missing value in a report value of the damage scale and an unexpected disaster occurs.SOLUTION: The disaster information providing system holds a report value including a missing value of a damage scale of a disaster at each of a plurality of points, adjacency information indicating an adjacency relationship between points, and a probability distribution model in which the damage scale at each point follows, and calculates an estimation value including uncertainty of a damage scale at a point having a report value that is a missing value and a parameter controlling a spatial random effect included in the probability distribution, based on the report value, a spatial random effect of each point based on the adjacency information, and the probability distribution model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a disaster information providing system, a disaster information providing method, and a disaster information providing program. [Background technology]

[0002] In order to respond appropriately to a disaster, it is desirable to understand the full extent of the damage caused by the disaster. However, during a disaster, information collection devices such as sensors may be damaged, or staff who should report damage information may be focused on responding to the scene and may not be able to collect enough information. As a result, the damage information collected by those making disaster response decisions may contain many missing values. Therefore, it is difficult to grasp the full extent of the damage by simply visualizing the collected information.

[0003] Japanese Patent Laid-Open Publication No. 2019-211931 (Patent Document 1) is a background technology in this technical field. In the technology described in Patent Document 1, a collection unit of a presentation device collects, from multiple information sources, information related to crisis response, which information is associated with each of the multiple information sources, a completion unit complements information that was not collected by the collection unit, an estimation unit estimates a risk of a crisis based on the information collected by the collection unit and the information complemented by the completion unit, a calculation unit calculates the reliability of the risk estimated by the estimation unit based on the degree of complementation by the completion unit, and a presentation unit presents the risk estimated by the estimation unit together with the reliability calculated by the calculation unit (see Abstract). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2019-211931 Summary of the Invention [Problem to be solved by the invention]

[0005] The technology described in Patent Document 1 requires simulation results for damage prediction because it supplements missing information with pre-determined estimated values, but it is difficult to perform a simulation that covers all possible situations.

[0006] Furthermore, although Patent Document 1 describes the completion of missing values ​​for observational data that directly observes the scale of a disaster, such as precipitation, it does not describe the estimation of the scale of damage caused by a disaster, such as the number of evacuees.

[0007] Therefore, one aspect of the present invention is to appropriately calculate an estimated value of the scale of damage caused by a disaster even if there are missing values ​​in the reported values ​​of the scale of damage and an unexpected disaster occurs. [Means for solving the problem]

[0008] In order to solve the above problem, one aspect of the present invention employs the following configuration: A disaster information provision system includes a processor and a memory, wherein the memory stores reported values ​​including missing values ​​of the scale of damage caused by a disaster at each of a plurality of locations, neighboring information indicating neighboring relationships between the plurality of locations, and a probability distribution model to which the damage scale at each of the plurality of locations conforms, wherein the damage scale at each of the plurality of locations according to the probability distribution model reflects a spatial random effect at the location, and the spatial random effect at each of the plurality of locations reflects the neighboring relationships related to the location in the neighboring information, and the probability distribution model includes a parameter for controlling the spatial random effect, and the processor calculates, based on the reported values, the spatial random effect based on the neighboring information, a damage scale estimate including uncertainty of the damage scale at locations having the reported values ​​that are missing, and a parameter estimate including uncertainty of the parameter, and generates data for displaying the damage scale estimate. [Effects of the Invention]

[0009] According to one aspect of the present invention, even if there are missing values ​​in the reported values ​​of the scale of damage and an unexpected disaster occurs, it is possible to properly calculate an estimated value of the scale of damage caused by the disaster.

[0010] Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a block diagram illustrating a configuration example of a disaster information providing system according to a first embodiment. [Figure 2] 2 is a block diagram showing an example of the hardware configuration of a computer constituting each of a damage scale estimation device, an information providing device, and an information using device in the first embodiment. FIG. [Figure 3] 10 is a flowchart illustrating an example of data collection processing and preprocessing in the first embodiment. [Figure 4] FIG. 2 is a diagram showing an example of the data configuration of evacuee number report data in the first embodiment. [Figure 5] FIG. 2 is a diagram illustrating an example of a data configuration of designated evacuation shelter data according to the first embodiment. [Figure 6] FIG. 10 is a diagram illustrating an example of a data configuration of updated shelter data according to the first embodiment. [Figure 7A] FIG. 2 is an explanatory diagram illustrating an example of an adjacency matrix in the first embodiment. [Figure 7B] FIG. 10 is a diagram illustrating an example of a modified adjacency matrix according to the first embodiment. [Figure 8] 10 is a flowchart illustrating an example of a damage scale estimation process according to the first embodiment. [Figure 9A] FIG. 10 is a diagram showing an example of a screen configuration of a screen displaying the number of evacuees in the first embodiment. [Figure 9B] FIG. 10 is a diagram showing an example of a screen configuration of a screen displaying the number of evacuees in the first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. In this embodiment, the same components are generally designated by the same reference numerals, and repeated explanations will be omitted. It should be noted that this embodiment is merely an example for realizing the present invention, and does not limit the technical scope of the present invention. [Example]

[0013] 1 is a block diagram showing an example configuration of a disaster information provision system. The disaster information provision system includes, for example, a damage scale estimation device 100, one or more information provision devices 200, and one or more information utilization devices 250. Each information provision device 200 is connected to the damage scale estimation device 100 via a network such as the Internet. Each information provision device 200 transmits information used to estimate the scale of damage to the damage scale estimation device 100.

[0014] The damage scale estimation device 100 has, for example, a data collection unit 111, a preprocessing unit 112, a damage scale estimation unit 113, and a presentation unit 114, all of which are functional units. The damage scale estimation device 100 also has a regional information storage unit 121, a statistical data storage unit 122, and a calculation result storage unit 123, all of which are storage units that store data.

[0015] The data collection unit 111 collects data from the information providing device 200. For example, the data collection unit 111 may collect data from the information providing device 200 at regular intervals, or may monitor the information providing device 200 and collect information when it is confirmed that the data has been updated.

[0016] The preprocessing unit 112 executes preprocessing, including processing of the data collected by the data collection unit 111 and the data stored in the regional information storage unit 121. The damage scale estimation unit 113 executes damage scale estimation processing using the data preprocessed by the preprocessing unit 112 and the data stored in the statistical data storage unit 122. The damage scale estimation unit 113 stores data indicating the processing results of the damage scale estimation processing in the calculation result storage unit 123. The presentation unit 114 generates data for displaying a screen including the processing results of the damage scale estimation processing, and transmits the data to the information utilization device 250.

[0017] Each information utilization device 250 is connected to the damage scale estimation device 100 via a network such as the Internet. The information utilization device 250 displays a screen in accordance with data transmitted from the damage scale estimation device 100.

[0018] 2 is a block diagram showing an example of the hardware configuration of a computer 1000 constituting each of the damage scale estimation device 100, the information providing device 200, and the information utilization device 250. The computer 1000 includes, for example, a CPU (Central Processing Unit) 1001, a memory 1002, an auxiliary storage device 1003, a communication device 1004, an input device 1005, and a display device 1006.

[0019] The CPU 1001 includes a processor and executes a program stored in the memory 1002. Note that, while the CPU 1001 or a GPU (Graphics Processing Unit) can be considered as an example of a processor, other semiconductor devices may also be used as long as they are capable of executing predetermined processing.

[0020] The memory 1002 includes a ROM (Read Only Memory), which is a nonvolatile storage element, and a RAM (Random Access Memory), which is a volatile storage element. The ROM stores unchanging programs (e.g., a BIOS (Basic Input / Output System)). The RAM is a high-speed, volatile storage element such as a DRAM (Dynamic Random Access Memory), and temporarily stores programs executed by the CPU 1001 and data used when the programs are executed.

[0021] The auxiliary storage device 1003 is a large-capacity, non-volatile storage device such as a magnetic storage device (HDD (Hard Disk Drive)) or a flash memory (SSD (Solid State Drive)), and stores programs to be executed by the CPU 1001 and data to be used when the programs are executed. That is, the programs are read from the auxiliary storage device 1003, loaded into the memory 1002, and executed by the CPU 1001.

[0022] The input device 1005 is a device that receives input from the user, such as a keyboard or mouse. The display device 1006 is a device that outputs the results of program execution in a format that can be viewed by the user, such as a display or printer.

[0023] The communication device 1004 is a network interface device that controls communication with other devices in accordance with a predetermined protocol, and may also include a serial interface such as a USB (Universal Serial Bus).

[0024] A part or all of the programs executed by the CPU 1001 may be provided to the computer 1000 from a removable medium (such as a CD-ROM or flash memory) which is a non-transitory storage medium, or from an external computer equipped with a non-transitory storage device via the network 150, and may be stored in the non-volatile auxiliary storage device 1003 which is a non-transitory storage medium. For this reason, the computer 1000 may have an interface for reading data from removable media.

[0025] The damage scale estimation device 100 is a computer system configured on a single physical computer 1000 or on multiple logically or physically configured computers 1000, and may operate in separate threads on the same computer 1000, or may operate on a virtual computer constructed on multiple physical computer resources. The same applies to the information providing device 200 and the information utilizing device 250.

[0026] The CPU 1001 of the computer 1000 constituting the damage scale estimation device 100 has each functional unit of the damage scale estimation device 100 shown in Fig. 1. For example, the CPU 1001 of the computer 1000 constituting the damage scale estimation device 100 functions as a data collection unit 111 by operating in accordance with a data collection program loaded into the memory 1002 of the computer 1000 constituting the damage scale estimation device 100, and functions as a preprocessing unit 112 by operating in accordance with a preprocessing program loaded into the memory 1002. The relationships between the functional units and the programs are similar for the other functional units of the damage scale estimation device 100.

[0027] Note that some or all of the functions of the damage scale estimation device 100 may be realized by a dedicated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array).

[0028] The auxiliary storage device 1003 of the computer 1000 constituting the damage scale estimation server 130 provides a storage area for realizing each storage unit shown in Fig. 1. Note that some or all of the information stored in the area information storage unit 121, the statistical data storage unit 122, and the calculation result storage unit 123 may be stored in the memory 1002 of the computer 1000 constituting the damage scale estimation device 100, or may be stored in an external database connected to the damage scale estimation device 100.

[0029] In this embodiment, the information used by the disaster information provision system does not depend on the data structure and may be expressed in any data structure. For example, the information can be stored in a data structure appropriately selected from a table, a list, a database, or a queue.

[0030] At least some of the devices included in the disaster information provision system may be integrated. For example, the damage scale estimation device 100 and the information provision device 200 may be integrated, or the damage scale estimation device 100 and the information utilization device 250 may be integrated, or the information provision device 200 and the information utilization device 250 may be integrated. Communication between devices integrated in this way is omitted.

[0031] 3 is a flowchart showing an example of data collection processing and preprocessing. The data collection unit 111 collects various data from the information providing device 200 (S301). Specifically, for example, evacuee number report data 310, flood damage data 350, and road damage data 360 are collected.

[0032] 4 is a diagram showing an example of the data configuration of the number of evacuees report data 310. The number of evacuees report data 310 is data on the number of evacuees at evacuation shelters, transmitted from the information providing device 200, for example, when a disaster occurs. Hereinafter, an evacuation shelter for which a report has been made on the number of evacuees will also be referred to as a reported evacuation shelter.

[0033] The number of evacuees report data 310 includes, for example, a local government column 311 , a shelter name column 312 , a report date and time column 313 , an opening date and time column 314 , a closing date and time column 315 , a number of evacuees column 316 , and a location column 317 .

[0034] The local government field 311 indicates the name identifying the local government in which the reported evacuation shelter is located. The evacuation shelter name field 312 indicates the name identifying the reported evacuation shelter. The report date and time field 313 indicates the date and time of the report of the number of evacuees indicated by the number of evacuees field 316. The opening date and time field 314 indicates the date and time the reported evacuation shelter was opened. The closing date and time field 315 indicates the date and time the reported evacuation shelter was closed.

[0035] The number of evacuees column 316 indicates the number of evacuees at the reporting evacuation shelter at the reporting date and time indicated in the report date and time column 313. The number of evacuees at a evacuation shelter is an example of an indicator indicating the scale of damage caused by a disaster at the location of the evacuation shelter (a location identified by the evacuation shelter). Furthermore, since the number of evacuees report data 310 is data related to evacuation shelters where the number of evacuees was actually reported, only actual values ​​(actual reported values) that are not missing values ​​may be stored in the number of evacuees column 316. The location column 317 indicates the location of the reporting evacuation shelter, for example, by latitude and longitude.

[0036] Returning to the explanation of Figure 3, the pre-processing unit 112 extracts evacuation shelters from those included in the evacuation shelter number report data 310. Specifically, for example, the pre-processing unit 112 generates extracted reported evacuation shelter data 320 by extracting the municipality column 311, the evacuation shelter name column 312, and the location column 317 from the evacuation shelter number report data 310. The pre-processing unit 112 integrates the extracted reported evacuation shelter data 320 with the designated evacuation shelter data 410 pre-stored in the area information storage unit 121 to generate updated evacuation shelter data 330 (S302).

[0037] 5 is a diagram showing an example of the data configuration of the designated evacuation shelter data 410. The designated evacuation shelter data 410 is data indicating designated evacuation shelters, which are evacuation shelters that are determined in advance before a disaster occurs and that are available when a disaster occurs. Hereinafter, the reported evacuation shelters and designated evacuation shelters will be collectively referred to as simply evacuation shelters.

[0038] The designated evacuation shelter data 410 includes, for example, a municipality column 411, a shelter name column 412, and a location column 413. The municipality column 411 indicates the name identifying the municipality in which the designated evacuation shelter is located. The shelter name column 412 indicates the name of the designated evacuation shelter. The location column 413 indicates the location of the designated evacuation shelter, for example, by latitude and longitude.

[0039] 6 is a diagram showing an example of the data configuration of the updated shelter data 330. The updated shelter data 330 is data indicating shelters included in at least one of the extracted and reported shelter data 320 and the designated shelter data 410.

[0040] When a disaster occurs, new evacuation shelters may be opened in an emergency in addition to designated evacuation shelters. Therefore, the number of evacuees reported data 310 (extracted reported evacuation shelter data 320) may indicate reported evacuation shelters that are not shown in the designated evacuation shelter data 410. Furthermore, when a disaster occurs, due to circumstances such as delays in reporting the number of evacuees at designated evacuation shelters, information about some designated evacuation shelters may not be reflected in the number of evacuees reported data 310. Therefore, the designated evacuation shelter data 410 may indicate reported evacuation shelters that are not shown in the number of evacuees reported data 310 (extracted reported evacuation shelter data 320).

[0041] In this way, the evacuation shelter number report data 310 and the designated evacuation shelter data 410 do not necessarily cover all evacuation shelters. Therefore, in step S302, the preprocessing unit 112 generates updated evacuation shelter data 330 consisting of records included in at least one of the extracted reported evacuation shelter data 320 and the designated evacuation shelter data 410, thereby generating more comprehensive data on evacuation shelters.

[0042] The updated evacuation shelter data 330 includes, for example, a municipality column 331, a shelter name column 332, and a location column 333. The municipality column 331 indicates the name of the municipality where the evacuation shelter is located. The evacuation shelter name column 332 indicates the name of the evacuation shelter. The location column 333 indicates the location of the evacuation shelter.

[0043] Returning to the explanation of FIG. 3, the preprocessing unit 112 generates an adjacency matrix 340 indicating the adjacency relationship between the positions of the shelters indicated by the updated shelter data 330 (S303). Specifically, for example, the preprocessing unit 112 generates a Voronoi diagram by generating Voronoi regions (i.e., Voronoi regions corresponding to each shelter) in a predetermined area (e.g., a predetermined area including all of the shelters indicated by the updated shelter data 330) with each of the shelter positions indicated by the updated shelter data 330 as a generating point. Generating the Voronoi regions corresponding to each shelter makes it possible to define the adjacency relationship between each shelter. Note that since each shelter corresponds to a Voronoi region, each shelter can be considered a (Voronoi) region, and the number of evacuees at each shelter can also be considered the scale of damage caused by the disaster in that (Voronoi) region.

[0044] When the number of shelters indicated by the updated shelter data 330 is N, the preprocessing unit 112 generates an N×N adjacency matrix 340. The preprocessing unit 112 determines the value of each component of the adjacency matrix 340 as follows: The preprocessing unit 112 determines a combination of Voronoi regions (combinations of shelters) whose Voronoi boundaries are in contact as a combination of adjacent shelters, and sets the value of the component corresponding to the adjacent combination of shelters to "1." The preprocessing unit 112 determines a combination of Voronoi regions (combinations of shelters) whose Voronoi boundaries are not in contact as a combination of non-adjacent shelters, and sets the value of the component corresponding to the non-adjacent combination of shelters to "0."

[0045] The preprocessing unit 112 generates a corrected adjacency matrix 345 by correcting the adjacency matrix 340 taking into account the accessibility between the evacuation shelters (S304), and ends the preprocessing. In step S304, the preprocessing unit 112 uses, as data indicating the accessibility between the evacuation shelters, for example, at least one of the topographical data 420 and the road network data 430 stored in advance in the regional information storage unit 121, and the flood damage data 350 and the road damage data 360 collected by the data collection unit 111.

[0046] The topographical data 420 stores, for example, a plurality of polygons shown in a geographical plane space (i.e., their positions are shown), and each polygon is assigned geographical attributes such as elevation. The road network data 430 stores, for example, a plurality of lines shown in a geographical plane space, and each of the plurality of lines indicates a road on which the road is laid. The topographical data 420 and the road network data 430 are examples of geographical information of an area including an evacuation shelter.

[0047] The flood damage data 350 stores a plurality of polygons shown in a geographical plane space, and each polygon is assigned attributes related to flood damage (for example, a flag indicating whether flood damage has occurred and / or the value of the flood depth that has occurred). The road damage data 360 stores a plurality of points shown in a geographical plane space, and each point is assigned attributes related to road damage (for example, a flag indicating whether road damage has occurred). The flood damage data 350 and road damage data 360 are data related to specific damage caused by disasters that have occurred in areas that include evacuation shelters.

[0048] The preprocessing unit 112 refers to data indicating the accessibility between shelters, and identifies combinations of shelters for which the components of the adjacency matrix are "1", and identifies those combinations of shelters for which access is not possible between the Voronoi regions corresponding to the combinations of shelters.The preprocessing unit 112 generates a modified adjacency matrix 345 by changing the components of the adjacency matrix 340 corresponding to the identified combinations of shelters to "0".

[0049] For example, if the preprocessing unit 112 determines, by referring to the topographical data 420, that the elevation difference indicated by the attribute assigned to the polygon located on the Voronoi boundary between adjacent Voronoi regions is equal to or greater than a predetermined value, it determines that access between the adjacent Voronoi regions is impossible.

[0050] Furthermore, when the preprocessing unit 112 determines, for example, by referring to the road network data 430, that there are no roads that cross adjacent Voronoi regions (or when it determines that the number of roads that cross adjacent Voronoi regions is less than a predetermined number), it determines that access between the adjacent Voronoi regions is impossible.

[0051] Furthermore, if the preprocessing unit 112 determines, for example, by referring to the flood damage data 350, that the flag indicated by the attribute assigned to a polygon located on the Voronoi boundary between adjacent Voronoi regions is a value indicating that flood damage has occurred, it determines that access between the adjacent Voronoi regions is impossible.

[0052] Furthermore, the preprocessing unit 112 treats, for example, lines indicated by the road network data 430 that include points in the road damage data 360 that have been flagged with a value indicating that road damage has occurred as impassable roads (for example, roads that do not exist), and if it determines that there are no roads that cross adjacent Voronoi regions (this may also be the case if it determines that the number of roads that cross adjacent Voronoi regions is less than a predetermined number), it determines that access between the adjacent Voronoi regions is impossible.

[0053] 7A is an explanatory diagram showing an example of an adjacency matrix 340, and FIG. 7B is a diagram showing an example of a modified adjacency matrix 345. In the Voronoi diagram 700 of FIGS. 7A and 7B, each circle represents a shelter (i.e., a generating point), and the Voronoi boundary surrounding each shelter is calculated. For ease of explanation, only shelters A, B, C, and D in the Voronoi diagram 700 will be discussed below.

[0054] In the Voronoi diagram 700, shelter A and shelter B are adjacent, shelter A and shelter C are adjacent, shelter A and shelter D are adjacent, shelter B and shelter D are adjacent, and shelter C and shelter D are adjacent. Therefore, in the case of Fig. 7A, the preprocessing unit 112 generates an adjacency matrix 740 in step S303, in which the value of the component corresponding to the pair of adjacent shelters is set to "1" and the values ​​of the other components are set to "0."

[0055] As shown in the Voronoi diagram 700 in Figure 7B, when data indicating accessibility is taken into consideration, it is determined that access is not possible between the Voronoi region corresponding to shelter A and the Voronoi region corresponding to shelter C. In this case, shelter A and shelter C are treated as not adjacent, so in step S304, the preprocessing unit 112 generates a modified adjacency matrix 750 by changing each element corresponding to the combination of shelter A and shelter C in the adjacency matrix 740 to "0."

[0056] 8 is a flowchart showing an example of the damage scale estimation process. In the damage scale estimation process, the damage scale estimation unit 113 estimates the number of evacuees at each evacuation shelter indicated by the updated evacuation shelter data 330.

[0057] In the damage scale estimation process, the ZIP (Zero-Inflated Poisson model) shown in the following equation 1 is used as a model (probability distribution model) to represent the number of evacuees at each shelter.

[0058]

number

[0059] ω is the probability that the number of evacuees at a certain shelter is not zero. μ is the expected value of the number of evacuees at the shelter when the number of evacuees at the shelter is assumed to follow a Poisson distribution. y is the reported number of evacuees at the shelter.

[0060] The details of the model are shown in the following formulas 2, 3, and 4. As shown in formula 2, the number of evacuees y at each evacuation shelter i indicated by the updated evacuation shelter data 330 is i is ZIP(μ i ,ω i ) shall be in accordance with the

[0061]

number

[0062] μ i is the expected number of evacuees at shelter i when the number of evacuees is assumed to follow a Poisson distribution. i is the probability that the number of evacuees in shelter i is not 0. μ i is expressed by the following equation 3, and ω i is expressed by the following equation 4.

[0063]

number

[0064]

number

[0065] x i is a feature vector whose components indicate the features of shelter i. i The feature values ​​indicated by each component of may include, for example, statistical data (e.g., stored in advance in the statistical data storage unit 122) that are not dependent on observation data related to the disaster, such as the number of evacuees that can be accommodated at the shelter i and the number of predetermined facilities deployed, or may include observation data related to the disaster, such as the amount of precipitation, temperature, and / or estimated flooded area at the shelter i (e.g., collected by the data collection unit 111). β is x i is a parameter to be estimated, which has the same dimension as

[0066] O i is statistical data that does not depend on observational data related to disasters, for example, the population around evacuation shelter i. i is, for example, the population within a predetermined radius (for example, 2 km) of the evacuation shelter i, or the population within a range that can reach the evacuation shelter i within a predetermined time (for example, 30 minutes), and is stored in advance in the statistical data storage unit 122. i is the parameter to be estimated that indicates the spatial random effect (spatial correlation component) of shelter i. δ is x iis a parameter to be estimated, which has the same dimension as

[0067] Note that equation 3 is x i T It is not necessary to include the β term, and Equation 4 can be expressed as x i T The term δ may not be included (for example, x in Equation 3 and Equation 4). i (may be a zero vector). In the following, in this embodiment, if Equation 3 is x i T It does not include the term β, and Equation 4 is x i T The term δ is not included.

[0068] Also, the spatial random effect Ψ of shelter i i The distribution model is expressed by the following equation 5.

[0069]

number

[0070] N(μ,σ) is a normal distribution with mean μ and variance σ. ij is the value of the element corresponding to the combination of shelter i and shelter j in the modified adjacency matrix 345. Note that the processing of step S304 may be omitted. In this case, a ij is the value of the element corresponding to the combination of shelter i and shelter j in the adjacency matrix 340. ρ is the parameter to be estimated that controls the strength of the spatial correlation. τ spatial is the spatial random effect Ψ i is the parameter to be estimated that controls the magnitude of the spatial random effect Ψ i Although ρ can take different values ​​for each shelter i, it is similar between neighboring shelters i (i.e., it has spatial correlation). spatial is an example of a parameter that controls the spatial random effect.

[0071] The input to the above model is y i(including the number of evacuees as reported values ​​and the number of evacuees as missing values ​​set in step S802 described later), O i , and a ij Also, if equation 3 is x i T β term and / or Equation 4 is x i T When the δ term is included, the input to the model above is x i Further includes:

[0072] The damage scale estimation unit 113 estimates each parameter of the estimation target (i.e., ρ and τ spatial ) is set to, for example, a uniform distribution (S801). The prior distribution set in step S801 may be set in advance, or may be determined by a user input via the input device 1005 of the computer 1000 constituting the damage scale estimation device 100.

[0073] The damage scale estimation unit 113 extracts shelters for which the number of evacuees is zero in the evacuee number report data 310 and shelters that are included in the updated evacuee shelter data 330 but not in the evacuee number report data 310, and sets the number of evacuees at the extracted shelters as a missing value (S802). When the number of evacuees is unknown, there are cases where the reported value of the number of evacuees is reported as zero (an example of a predetermined value). Therefore, in step S802, the damage scale estimation unit 113 sets the number of evacuees at shelters for which the number of evacuees is zero in the evacuee number report data 310 as a missing value. Note that when the number of evacuees is unknown, if the value of the number of evacuees is a predetermined value other than zero (for example, "-1"), the damage scale estimation unit 113 sets the number of evacuees at shelters for which the number of evacuees is the predetermined value as a missing value, instead of shelters for which the number of evacuees is zero. Furthermore, the damage scale estimation unit 113 may set the number of evacuees at a shelter where the number of evacuees in the evacuee number report data 310 is the same as the value set as the initial value as a missing value. Note that the initial value may be common to all evacuation shelters, or may be different for some or all of the evacuation shelters. The initial value may be stored in advance in the area information storage unit 121, or may be collected by the data collection unit 111 in step S301.

[0074] The damage scale estimation unit 113 calculates the number of evacuees y i is fixed to the value updated in the most recent step S804, and each parameter to be estimated (ρ and τ spatial ) is updated (S803). i Each value may be fixed, for example, to a predetermined initial value or set to a missing value. i The number of evacuees y that is not set as a missing value may be fixed to a predetermined statistical value (for example, the average value). i is fixed to the reported value indicated by the number of evacuees report data 310 in step S803.

[0075] The damage scale estimation unit 113 calculates the updated parameters (ρ and τ spatial ) in step S803, the updated prior distribution is fixed, and the number of evacuees y i is updated based on the parameter update conditions (S804).

[0076] The damage scale estimation unit 113 determines whether the number of times the series of processes consisting of the process of step S803 and the process of step S804 has been executed has reached a preset update count (S805). If the damage scale estimation unit 113 determines that the number of times the series of processes has been executed has not reached the update count (S805: NO), it returns to step S803. If the damage scale estimation unit 113 determines that the number of times the series of processes has been executed has not reached the update count (S805: NO), it returns to step S803.

[0077] If the damage scale estimation unit 113 determines that the number of times the series of processes has been executed has reached the number of updates (S805: YES), it stores the parameters to be estimated and the estimated number of evacuees at each shelter i in the calculation result memory unit 123, for example, for each target time of the damage scale estimation process (S806), and terminates the damage scale estimation process.

[0078] In step S806, the damage scale estimation unit 113 specifically calculates, for example, the parameters (ρ and τ spatial ) and the mean value and 95% confidence interval of the number of evacuees corresponding to each shelter i (including shelters with missing values ​​and shelters with no missing values) μ i and number of evacuees y i The 95% confidence interval is stored in the calculation result storage unit 123. Note that the above confidence interval is not limited to the 95% confidence interval, and may be any confidence interval.

[0079] The damage scale estimation unit 113 executes the loop of processing from step S803 to step S805 using, for example, the multiple imputation method. For example, the parameter update conditions in each of step S803 and step S804 are determined by the Metropolis-Hastings method. The damage scale estimation unit 113 also updates the prior distribution in step S803 using, for example, the Metropolis-Hastings method.

[0080] Also, if Equation 3 is x i T When the term β is included, β is also included in the parameters to be estimated in steps S801, S803, and S806. i T When the term δ is included, δ is also included in the parameters to be estimated in steps S801, S803, and S806.

[0081] According to the process of FIG. 8, the damage scale estimation unit 113 estimates the parameters of the above model and the missing value of the number of evacuees from inputs including data of reported values ​​including the missing value of the number of evacuees at evacuation shelters and data showing the adjacency relationship between evacuation shelters (adjacency matrix 340, or modified adjacency matrix 345 taking accessibility into consideration). By the process of FIG. 8, the damage scale estimation unit 113 can complement the missing value of the number of evacuees without requiring prior simulations, etc. Also, the above inputs include the number of evacuees y i , (corrected) adjacency matrix element a ij , Surrounding population Oi However, it is not necessary to include observation data that directly observes the magnitude of the disaster (disaster scale) such as the amount of precipitation, the height of the tsunami, or the seismic intensity, so even if a disaster that was not anticipated occurs, the damage scale estimation unit 113 can estimate the scale of the damage.

[0082] Furthermore, in the processing of Fig. 8, since the parameter and the number of evacuees follow a probability distribution, the damage scale estimation unit 113 generates estimated values ​​of the parameter and the missing value of the number of evacuees as values ​​including uncertainty (average value and confidence interval). Furthermore, in the processing of Fig. 8, the damage scale estimation unit 113 uses multiple imputation to complement the missing value of the number of evacuees, thereby making it possible to deal with the uncertainty in estimating the missing value.

[0083] In the process of FIG. 8, the damage scale estimation unit 113 calculates the average value μ i spatial random effect Ψ i and the spatial random effect Ψ i The data showing the adjacency relationship between the evacuation shelters (adjacency matrix 340 or modified adjacency matrix 345 that takes accessibility into account) is reflected in the damage scale estimation unit 113. i can be calculated.

[0084] 9A and 9B are diagrams showing an example of the screen configuration of the number of evacuee display screen 900. Due to drawing space limitations, the number of evacuee display screen 900 is shown in two parts, Fig. 9A and Fig. 9B. The number of evacuee display screen 900 is displayed on the display device 1006 of the information utilization device 250 based on screen data generated by the presentation unit 114 based on information stored in the calculation result storage unit 123.

[0085] The number of evacuees display screen 900 includes, for example, a reported number of evacuees map display area 910, an estimated number of evacuees map display area 920, and an evacuees number graph display area 930. Information on the maps displayed in the reported number of evacuees map display area 910 and the estimated number of evacuees map display area 920 is stored in advance in, for example, the regional information storage unit 121.

[0086] On the map displayed in the reported number of evacuees map display area 910 for each time, evacuation shelters included in the evacuation shelter number report data 310 at that time, whose number of evacuees was not set to a missing value in step S802, are shown superimposed (at the location of the evacuation shelter) with a circle, and the number of evacuees (i.e., the reported value) at that time as indicated by the evacuation shelter number report data 310 is shown by the shade of color inside the circle.

[0087] For example, since the number of shelters for which the number of evacuees has been reported tends to increase as time passes, the number of shelters shown in the map in the reported number of evacuees map display area 910 also tends to increase as time passes.

[0088] On the map at each time displayed in the estimated number of evacuees map display area 920, each evacuation shelter i included in the updated evacuation shelter data 330 is superimposed with a circle (at the position of the evacuation shelter), and the average number of evacuees at each evacuation shelter i μ i is shown as an estimated value by the shade of color inside the circle. The estimated numbers of evacuees at all evacuation shelters i included in the updated evacuation shelter data 330 are displayed within the map in estimated evacuation number map display area 920, so the user of information usage device 250 can obtain estimated results for the situation at each evacuation shelter early after the occurrence of a disaster.

[0089] 9A, the map for "2024.07.02 12:00" in the reported number of evacuees map display area 910 does not show any areas where the number of evacuees tends to be high, but the map for "2024.07.04 12:00" in the reported number of evacuees map display area 910 shows that the number of evacuees tends to be high in the area slightly southeast (bottom right) from the center of the map. In other words, it may be difficult for users to grasp the trend and scale of damage based only on the reported number of evacuees until some time has passed since the disaster occurred.

[0090] 9A, it can be seen that as of "2024.07.02 12:00," the estimated number of evacuees map display area 920 already shows a tendency for the number of evacuees to be high in the area slightly southeast (bottom right) of the center of the map. In other words, because estimated values ​​for the number of evacuees at all evacuation centers have been obtained, the user can grasp the trend and scale of the damage even if a short time has passed since the disaster occurred.

[0091] The number of evacuees graph display area 930 displays, for example, the average number of evacuees at each evacuation shelter i μ i and a graph showing the time series of the total value of the total value of the number of evacuees with a 95% confidence interval, and a graph showing the time series of the total value of the reported values ​​of the number of evacuees at evacuation shelters included in the evacuation shelter number report data 310, the number of evacuees of which was not set as a missing value in step S802. Note that the evacuation shelter number graph display area 930 displays the average number of evacuees μ i A graph showing the total value and the number of evacuees with a 95% confidence interval over time, and a graph showing the reported value of the number of evacuees may be displayed for each evacuation shelter i.

[0092] In the example of FIG. 9B, the total of the reported values ​​is low until a certain amount of time has passed since the occurrence of the disaster, showing a different trend from the total of the estimated values, and then rises after a certain amount of time has passed, showing a similar trend to the total of the estimated values. In this way, by displaying not only the total of the reported values ​​but also the total of the estimated values, the user can grasp the scale of the damage even if only a short time has passed since the occurrence of the disaster. Note that, on the evacuation number display screen 900, the parameters to be estimated at each time (for example, ρ and τ spatial ) and 95% confidence intervals may also be displayed.

[0093] The present invention is not limited to the above-described embodiments, but includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0094] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as the programs, tables, and files that implement each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

[0095] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and do not necessarily show all the control lines and information lines in the product. In reality, it can be assumed that almost all components are interconnected. [Explanation of symbols]

[0096] 100 damage scale estimation device, 111 data collection unit, 112 preprocessing unit, 113 damage scale estimation unit, 114 presentation unit, 121 regional information storage unit, 122 statistical data storage unit, 123 calculation result storage unit, 200 information provision device, 250 information utilization device, 310 evacuee number report data, 330 updated evacuation shelter data, 340 adjacency matrix, 350 corrected adjacency matrix, 410 designated evacuation shelter data, 1000 computer, 1001 CPU, 1002 memory, 1003 auxiliary storage device, 1004 communication device, 1005 input device, 1006 display device

Claims

1. A disaster information provision system, a processor and a memory, The memory includes: Reported values ​​including missing values ​​of the scale of damage caused by disasters at each of multiple locations, Adjacency information indicating an adjacency relationship between the plurality of points; a probability distribution model that the damage scale at each of the plurality of points follows; The damage scale at each of the plurality of points according to the probability distribution model reflects a spatial random effect at the point; the spatial random effect at each of the plurality of points reflects the adjacent relationship relating to the point in the adjacent information; the probability distribution model includes parameters that control the spatial random effect; The processor: calculating a damage scale estimate including uncertainty of the damage scale at a point having the missing reported value and a parameter estimate including uncertainty of the parameter based on the reported value, the spatial random effect based on the neighboring information, and the probability distribution model; A disaster information provision system that generates data for displaying the estimated damage scale.

2. 2. The disaster information providing system according to claim 1, The memory includes: Report information indicating actual reported values ​​of the damage scale at each of one or more locations that are not the missing values; and storing designated point information indicating a designated point that has been designated in advance; The processor: determining points included in at least one of the one or more points and the specified point as the plurality of points; storing the missing value in the memory as the reported value for at least one of a point among the plurality of points that is not included in the one or more points, a point where the actual reported value indicated by the report information is a predetermined value, and a point where the actual reported value indicated by the report information is an initial value; A disaster information provision system that stores in the memory the actual report value at a point among the plurality of points where the actual report value indicated by the report information is not the specified value as the reported value at that point.

3. 2. The disaster information providing system according to claim 1, The memory includes: an adjacency matrix indicating adjacent points among the plurality of points; access information indicating accessibility between points included in the plurality of points; The processor: By referring to the access information, among the adjacent points indicated by the adjacency matrix, identify adjacent points that are inaccessible between the adjacent points; A disaster information providing system that generates the adjacent information by modifying the adjacent matrix based on the identified adjacent points.

4. The disaster information providing system according to claim 3, A disaster information providing system, wherein the access information includes at least one of geographic information of an area including the plurality of locations and information regarding specific damage caused to the area by the disaster.

5. 2. The disaster information providing system according to claim 1, The processor calculates an estimated value that includes uncertainty between the missing values ​​and the parameters based on multiple imputation and Metropolis-Hastings algorithms.

6. 2. The disaster information providing system according to claim 1, A disaster information provision system in which the processor calculates an estimated value including uncertainty in the scale of damage at points having reported values ​​that are not missing values ​​based on the reported values, the spatial random effect based on the neighboring information, and the probability distribution model, and includes the estimated value in the damage scale estimate.

7. 7. The disaster information providing system according to claim 6, the memory holds information indicating a map including the plurality of points; A disaster information provision system in which the processor generates data for superimposing and displaying the estimated damage scale values ​​for each of the locations on the map.

8. 2. The disaster information providing system according to claim 1, The processor calculates an average value of the damage scale and a predetermined confidence interval as the damage scale estimated value including uncertainty of the damage scale.

9. A disaster information providing method by a disaster information providing system, The disaster information providing system includes a processor and a memory, The memory includes: Reported values ​​including missing values ​​of the scale of damage caused by disasters at each of multiple locations, Adjacency information indicating an adjacency relationship between the plurality of points; a probability distribution model that the damage scale at each of the plurality of points follows; The damage scale at each of the plurality of points according to the probability distribution model reflects a spatial random effect at the point; the spatial random effect at each of the plurality of points reflects the adjacent relationship relating to the point in the adjacent information; the probability distribution model includes parameters that control the spatial random effect; The disaster information providing method includes: the processor calculates a damage scale estimate including uncertainty of the damage scale at a point having the missing reported value and a parameter estimate including uncertainty of the parameter based on the reported value, the spatial random effect based on the neighboring information, and the probability distribution model; The disaster information providing method, wherein the processor generates data for displaying the damage scale estimation value.

10. A disaster information provision program that causes a disaster information provision system to execute disaster information provision processing, the disaster information providing system includes a processor and a memory, The memory includes: Reported values ​​including missing values ​​of the scale of damage caused by disasters at each of multiple locations, Adjacency information indicating an adjacency relationship between the plurality of points; a probability distribution model that the damage scale at each of the plurality of points follows; The damage scale at each of the plurality of points according to the probability distribution model reflects a spatial random effect at the point; the spatial random effect at each of the plurality of points reflects the adjacent relationship relating to the point in the adjacent information; the probability distribution model includes parameters that control the spatial random effect; The disaster information provision program a process of calculating a damage scale estimation value including uncertainty of the damage scale at a point having the reported value that is the missing value and a parameter estimation value including uncertainty of the parameter based on the reported value, the spatial random effect based on the neighboring information, and the probability distribution model; and generating data for displaying the damage scale estimation value.

Images