Snow storm disaster loss prediction model training method, prediction method and system

Abstract

The embodiment of the invention provides a violent snow disaster loss prediction model training method, prediction method and system, and the method comprises the steps: obtaining the basic parameter data of the violent snow disaster loss of a target region, the basic parameter data comprises snowstorm disaster event data, meteorological station observation data, land surface reanalysis data, snowstorm disaster situation data, disaster-bearing body data and disaster-pregnant environment data; determining characteristic quantity data of the target region based on the basic parameter data; and inputting the characteristic quantity data of the target area into a violent snow disaster loss prediction model to obtain a violent snow disaster loss grade of the target area, thereby accurately predicting a future violent snow disaster condition, and realizing universality.

Description

Technical Field

[0001] The present invention relates to the technical field of snowstorm disaster loss estimation, and in particular to a snowstorm disaster loss estimation model training method, an estimation method and a system. Background Art

[0002] Existing loss assessment techniques for blizzard disasters usually use loss assessment methods based on indicator systems, such as constructing indicators such as hazard, vulnerability, and disaster-prone environment, and then using weighting methods to determine the importance weights of each indicator, and then using mathematical statistics, hierarchical analysis, factor analysis, fuzzy mathematics and other methods to perform comprehensive scoring to measure the level and pattern of disaster losses. This method has the advantages of simple operation and easy implementation. However, this method does not fully consider the complex relationship between disaster-causing factors and disaster-bearing bodies, especially the nonlinear interaction, resulting in inaccurate assessment results in some cases. Summary of the invention

[0003] In view of this, the object of the present invention is to provide a snowstorm disaster loss estimation model training method, estimation method and system to improve the efficiency and accuracy of loss assessment.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] In a first aspect, the present invention provides a method for training a snowstorm disaster loss prediction model, comprising:

[0006] Construct sample sets and snowstorm disaster loss estimation models;

[0007] The construction of the sample set includes:

[0008] A sample set is constructed based on historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, historical blizzard disaster data, disaster-bearing bodies and disaster-pregnant environment data;

[0009] The sample set is processed by eliminating voids and dimensionless processing to obtain the snowstorm disaster and loss sample set;

[0010] The construction of a snowstorm disaster loss estimation model includes:

[0011] A neural network model is created with a snowstorm disaster and loss sample set as input and snowstorm disaster loss level as output, and the parameter values ​​of the optimizer type, loss function, evaluation index, number of iterations, activation function, and batch number of the neural network model are set.

[0012] Optionally, after building the snowstorm disaster loss estimation model, the method further includes: optimizing the weights and thresholds of the neural network model by the following optimization method:

[0013] Set parameters, including population size, search range, search dimension, scaling factor range, maximum number of iterations, and crossover probability;

[0014] Initialize through mutation selection operation based on initialization parameters, calculate the objective function value of the individual, and select the best individual;

[0015] Calculate the distance between other individuals and the optimal individual and update the position;

[0016] Update the parameters, initialize the parameters, cross-check the individual positions, select new individuals and calculate the objective function value;

[0017] Update the position of the optimal individual until the maximum number of iterations is reached and output the optimal solution.

[0018] Optionally, after optimizing the weights and thresholds of the snowstorm disaster loss estimation model, the following steps are also included:

[0019] Construct training set data based on historical blizzard disaster data;

[0020] Based on the training set data, iterative training is performed until the number of iterations is reached. The iterative training is to input the training set data into the neural network model and optimize the neural network model through optimization methods.

[0021] Optionally, the weights and thresholds of the snowstorm disaster loss estimation model are optimized, and further include:

[0022] Construct test data sets, including historical snowstorm disaster event data, meteorological station observation data, land surface reanalysis data, disaster-bearing body and disaster-pregnant environment data;

[0023] Input the test set data into the blizzard disaster loss model to obtain the predicted data;

[0024] Compare the predicted data with the real data to get the accuracy of the neural network model;

[0025] Based on the accuracy of the neural network model, it is determined whether the quality of the neural network model reaches the preset target.

[0026] In a second aspect, the present invention provides a method for estimating losses caused by a snowstorm disaster, comprising:

[0027] Obtain basic parameter data of blizzard disaster losses in the target area, including blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster data, disaster-bearing body data and disaster-prone environment data;

[0028] Based on the basic parameter data, determine the characteristic quantity data of the target area;

[0029] The characteristic quantity data of the target area is input into the blizzard disaster loss estimation model to obtain the blizzard disaster loss level of the target area, wherein the blizzard disaster loss estimation model is trained using the blizzard disaster loss estimation model training method as described above.

[0030] Optionally, determining the characteristic quantity data of the target area based on the basic parameter data further includes:

[0031] Based on the blizzard disaster data, the administrative district corresponding to the blizzard disaster is obtained;

[0032] Based on the administrative districts, meteorological station observation data and blizzard disaster event data, the direct economic losses corresponding to the administrative districts were obtained.

[0033] Optionally, determining the characteristic quantity data of the target area based on the basic parameter data further includes:

[0034] Performing time-space matching on the disaster-affected volume data to obtain the grid data of the disaster-affected volume data;

[0035] Perform time-space matching on disaster-prone environmental data to obtain raster data of disaster-prone environmental data;

[0036] The raster data of the disaster-prone body data and the raster data of the disaster-prone environment data are integrated to obtain the characteristic quantity data of the disaster-prone body and the disaster-prone environment.

[0037] In a third aspect, the present invention provides a system for estimating losses caused by a snowstorm disaster, comprising:

[0038] The acquisition module is used to obtain the basic parameter data of blizzard disaster losses in the target area. The basic parameter data include blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster data, disaster-bearing body data and disaster-prone environment data;

[0039] A processing module, used for determining characteristic quantity data of a target area based on basic parameter data;

[0040] The estimation module is used to input the characteristic quantity data of the target area into the blizzard disaster loss estimation model to obtain the blizzard disaster loss level of the target area, wherein the blizzard disaster loss estimation model is trained using the above-mentioned blizzard disaster loss estimation model training method.

[0041] In a fourth aspect, the present invention further provides an electronic device, including a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor can execute the machine executable instructions to implement the above-mentioned snowstorm disaster loss estimation method.

[0042] In a fifth aspect, the present invention further provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to execute the above-mentioned blizzard disaster loss estimation method.

[0043] The present invention provides a blizzard disaster loss prediction model training method, prediction method and system, which obtains basic parameter data of blizzard disaster losses in a target area, the basic parameter data including blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster situation data, disaster-bearing body data and disaster-pregnant environment data; based on the basic parameter data, determines the characteristic quantity data of the target area; inputs the characteristic quantity data of the target area into the blizzard disaster loss prediction model to obtain the blizzard disaster loss level of the target area, thereby accurately predicting future blizzard disaster situations and achieving universality.

[0044] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 A schematic diagram of a process for estimating a snowstorm disaster loss provided by an embodiment of the present invention is shown;

[0047] Figure 2 A schematic diagram of a process for training a snowstorm disaster loss prediction model provided by an embodiment of the present invention is shown;

[0048] Figure 3 A schematic diagram of the structure of a snowstorm disaster loss prediction system provided by an embodiment of the present invention is shown;

[0049] Figure 4 A schematic structural diagram of an electronic device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0050] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0051] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0052] Existing blizzard disaster loss evaluation methods tend to adopt linear or semi-linear assumptions to simplify complex disaster dynamic systems and their response characteristics. Under extreme disaster conditions, especially when large-scale and high-intensity blizzard disasters occur, this model is difficult to capture the drastic changes in disaster losses, which often leads to a large gap between the evaluation results and the actual situation, and the determination of indicator weights depends largely on the subjective selection of indicator assignment methods. This subjectivity may introduce bias and affect the accuracy of the final evaluation results. Based on this, the present invention provides a blizzard disaster loss prediction model training method, prediction method and system, by obtaining basic parameter data of blizzard disaster losses in a target area, the basic parameter data includes blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster data, disaster-bearing body data and disaster-pregnant environment data; based on the basic parameter data, determining the characteristic quantity data of the target area; inputting the characteristic quantity data of the target area into the blizzard disaster loss model to obtain the blizzard disaster loss level of the target area, so that the future blizzard disaster situation can be accurately predicted, and universality is achieved.

[0053] Figure 1 A schematic diagram of a process for estimating a snowstorm disaster loss provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method comprises the following steps:

[0054] Step S110, obtaining basic parameter data of blizzard disaster losses in the target area, the basic parameter data including blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster data, disaster-bearing body data and disaster-prone environment data.

[0055] Among them, the blizzard disaster event data includes the original information such as the site name, longitude, latitude, blizzard disaster event start date, blizzard disaster event end date and blizzard disaster event intensity index that records each blizzard disaster event;

[0056] The meteorological station observation data includes, but is not limited to, national stations, regional stations, and automatic stations, and the observation data include daily snowfall, cumulative snowfall, average snow depth, maximum snow depth, daily average wind speed, daily maximum wind speed, visibility, etc.;

[0057] The land surface reanalysis data are obtained by analyzing the daily snowfall, cumulative snowfall, average snow depth, maximum snow depth, daily average wind speed, daily maximum wind speed, visibility and other data;

[0058] Snowstorm disaster data includes direct economic losses and administrative districts.

[0059] The disaster-prone body data selects population density and gross national economic value as the basic disaster-prone body, and adds information such as the output value of the primary industry, the output value of the secondary industry, and the output value of the tertiary industry;

[0060] The disaster-prone environment data is obtained by analyzing factors such as topography, water system, vegetation and soil.

[0061] Step S120: Determine the characteristic data of the target area based on the basic parameter data.

[0062] In an optional embodiment, in step S120, determining the characteristic quantity data of the target area based on the basic parameter data includes:

[0063] Clean and grade the blizzard disaster event data.

[0064] Specifically, by checking the original information of each blizzard disaster event, such as the site name, longitude, latitude, blizzard disaster event start date, blizzard disaster event end date, and blizzard disaster event intensity index, the information that lacks any field information or has obvious outliers is removed to complete the cleaning of the blizzard disaster event data;

[0065] Statistics are collected for each blizzard disaster event, and the event intensity index is ranked from small to large. The 50%, 70%, and 90% blizzard disaster events are used as the three main reference points A, B, and C according to the ranking. Then, the blizzard disaster events are graded according to the reference points to form a blizzard disaster grade. The classification can be comprehensively considered in combination with the specific index value and the actual impact to ensure the rationality and practicality of the classification.

[0066] In this application, blizzard disaster events with an intensity index less than A are classified as a blue warning level; blizzard disaster events with an intensity index greater than A but less than B are classified as a yellow warning level; blizzard disaster events with an intensity index greater than B but less than C are classified as an orange warning level; blizzard disaster events with an intensity index greater than C are classified as a red warning level.

[0067] In an optional embodiment, in step S120, determining the characteristic quantity data of the target area based on the basic parameter data includes:

[0068] Perform time-space matching and data quality control processing on meteorological station observation data.

[0069] Specifically, the observation time of the meteorological station observation data is obtained, and the observation time is matched with the start date t1 of the blizzard disaster event and the end date t2 of the blizzard disaster event, and the meteorological station observation data between the start date t1 of the blizzard disaster event and the end date t2 of the blizzard disaster event are screened out to complete the spatiotemporal matching of the meteorological station observation data;

[0070] Then, quality control is carried out on the screened meteorological station observation data. The meteorological station observation data can be processed using inspection methods such as climatological limit value check, climate extreme value check, internal consistency check, spatial consistency check and temporal consistency check to determine the accuracy of the meteorological station observation data.

[0071] Among them, the climatological limit value check is to compare the observed data with the known climatological range and mark the abnormal data beyond these ranges; the climate extreme value check is to check whether the observed data exceeds the statistical extreme value range; the internal consistency check is to check whether the data at different time points of the same station are logical; the spatial consistency check is to check whether the data at different stations in the same area are reasonably consistent; the temporal consistency check is to check the internal consistency of the observed data at the same station at the same time point.

[0072] In an optional embodiment, in step S120, determining the characteristic quantity data of the target area based on the basic parameter data includes:

[0073] Perform spatiotemporal matching processing on land surface reanalysis data.

[0074] Specifically, the land surface reanalysis data time is matched with the start date t1 of the blizzard disaster event and the end date t2 of the blizzard disaster event, and the land surface analysis element data within the latitude and longitude range of the blizzard disaster event between the start date t1 of the blizzard disaster event and the end date t2 of the blizzard disaster event are screened out;

[0075] Since the land surface reanalysis data are grid point data, it is necessary to project the land surface reanalysis data to all grid point data within the administrative vector based on the county-level administrative vector data, and perform mean, maximum and minimum value statistics. Finally, the land surface analysis data are output according to the county-level units. That is, each land surface analysis data is projected to the corresponding county-level boundary, and the mean, maximum and minimum values ​​are counted by time period (such as day, week, month and year) and each county-level unit, so as to complete the spatiotemporal matching processing of the land surface reanalysis data.

[0076] In an optional embodiment, in step S120, determining the characteristic quantity data of the target area based on the basic parameter data includes:

[0077] Based on the blizzard disaster data, the administrative district corresponding to the blizzard disaster is obtained;

[0078] Based on the administrative districts, meteorological station observation data and blizzard disaster event data, the direct economic losses corresponding to the administrative districts were obtained.

[0079] Specifically, the blizzard disaster data is processed to obtain the name of the administrative region where the blizzard disaster is located and the direct economic losses caused by the blizzard;

[0080] Among them, the administrative district to which the blizzard disaster belongs can be obtained through the data observed by the meteorological observation station. Since the meteorological observation station data are mostly based on county-level units, that is, there is almost a meteorological observation station in a county-level administrative unit. Therefore, in order to match the meteorological observation station data and make all disaster data belong to the same level, it is necessary to clean the disaster data, extract data only for each county-level unit, delete provincial and municipal data, and complete the cleaning of the blizzard disaster data;

[0081] Then, by matching the meteorological station observation information with the disaster data, we can obtain the direct economic loss corresponding to the administrative area where the meteorological station is located. The meteorological station observation information includes the name of the station and the longitude and latitude of the station.

[0082] In an optional embodiment, in step S120, determining the characteristic quantity data of the target area based on the basic parameter data includes:

[0083] Performing time-space matching on the disaster-affected volume data to obtain the grid data of the disaster-affected volume data;

[0084] Perform time-space matching on disaster-prone environmental data to obtain raster data of disaster-prone environmental data;

[0085] The raster data of the disaster-prone body data and the raster data of the disaster-prone environment data are integrated to obtain the characteristic quantity data of the disaster-prone body and the disaster-prone environment.

[0086] Specifically, in this application, the disaster-bearing body selects population density and gross national economic value as the basic disaster-bearing body, and adds information such as the output value of the primary industry, the output value of the secondary industry and the output value of the tertiary industry. Among them, the indicators such as population density and gross national economic value are obtained through the statistical data of the county-level administrative units, which can quantitatively reflect the economic development and social structure characteristics of the county-level area. The output value of the primary industry, the output value of the secondary industry and the output value of the tertiary industry understand the impact of different industries on the change of disaster conditions. By statistically analyzing the industrial data at the county-level administrative units, the characteristics of the regional economic structure can be more comprehensively grasped.

[0087] The disaster-prone environment can select element data such as terrain, water system, vegetation and soil. These element data play a fundamental role in understanding the occurrence and development of disasters. Among them, the element data include elevation data, terrain undulation, river network density, vegetation coverage, and clay content of 30cm soil. These data are important factors affecting the disaster-prone environment;

[0088] Since both the disaster-prone body data and the disaster-prone environment data are grid point data, it is necessary to count the disaster-prone body data and the disaster-prone environment data at the county-level administrative unit. That is, based on the county-level administrative vector data, all the grid point data falling within the administrative vector are counted for mean, maximum and minimum values, and output according to the county-level units, thereby completing the fusion processing of the disaster-prone body data and the disaster-prone environment data.

[0089] Furthermore, the raster data of disaster-prone body data and disaster-prone environment data are spatially overlaid and analyzed with the county-level administrative division data; and based on the county-level administrative division data, all grid point data falling within each county-level unit are statistically analyzed to obtain statistical values ​​such as the mean, maximum value, and minimum value; finally, the statistical results within the county are summarized to form analyzable data tables or maps.

[0090] Step S130: input the characteristic quantity data of the target area into the blizzard disaster loss estimation model to obtain the blizzard disaster loss level of the target area, wherein the blizzard disaster loss estimation model is trained using the blizzard disaster loss estimation model training method.

[0091] Figure 2 A flowchart of a snowstorm disaster loss prediction model training method provided by an embodiment of the present invention is shown in FIG. Figure 2 As shown, the method includes: constructing a sample set and a model for predicting the loss of a snowstorm disaster;

[0092] Among them, the sample set for building the snowstorm disaster loss estimation model includes:

[0093] A sample set is constructed based on historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, historical blizzard disaster data, disaster-bearing bodies and disaster-pregnant environment data;

[0094] The sample set is processed by eliminating voids and dimensionless processing to obtain the snowstorm disaster and loss sample set;

[0095] The construction of a snowstorm disaster loss estimation model includes:

[0096] A neural network model is created with a snowstorm disaster and loss sample set as input and snowstorm disaster loss level as output, and the parameter values ​​of the optimizer type, loss function, evaluation index, number of iterations, activation function, and batch number of the neural network model are set.

[0097] Furthermore, a sample set of the snowstorm disaster loss estimation model is constructed, including:

[0098] First, historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, historical blizzard disaster data, disaster-bearing bodies and disaster-pregnant environmental data are matched based on county-level units to form a sample set;

[0099] Then, the sample set is processed to eliminate empty data, that is, if a certain data is empty or missing, the sample is deleted to avoid the data used for training containing missing data;

[0100] Finally, the sample set after the void elimination process is normalized to the maximum and minimum values. All feature quantity data are normalized to the maximum and minimum values, converted into dimensionless values, and the dimension differences of each feature quantity data are eliminated to form a snowstorm disaster and loss sample set, where the feature quantity data includes historical snowstorm disaster event data, meteorological station observation data, land surface reanalysis data, disaster-bearing body and disaster-pregnant environment data;

[0101] Suppose the original value of a feature is x, and the minimum value in the data is x min , the maximum value is x max , then the normalized calculation formula is:

[0102]

[0103] Furthermore, the neural network model selected in this application is BP neural network (Back Propagation Neural Network), which is a multi-layer feedforward neural network trained according to error back propagation. BP neural network consists of an input layer, a hidden layer and an output layer, wherein the hidden layer can be single-layer or multi-layer.

[0104] Since the problem involved in the present invention is a regression model, the following parameters need to be set to construct the neural network model in this application:

[0105] Set optimizer: Adam;

[0106] Set the loss function: mean_squared_error;

[0107] Set the evaluation metric: accuracy;

[0108] Set the number of iterations: 200;

[0109] Set activation function: ReLU

[0110] Set the batch size, which is the amount of data input to the model each time it is trained: 64.

[0111] In an optional embodiment, the weights and thresholds of the neural network model are optimized by the following optimization method:

[0112] Set parameters, including population size, search range, search dimension, scaling factor range, maximum number of iterations, and crossover probability;

[0113] Initialize through mutation selection operation based on initialization parameters, calculate the objective function value of the individual, and select the best individual;

[0114] Calculate the distance between other individuals and the optimal individual and update the position;

[0115] Update the parameters, initialize the parameters, cross-check the individual positions, select new individuals and calculate the objective function value;

[0116] Update the position of the optimal individual until the maximum number of iterations is reached and output the optimal solution.

[0117] Furthermore, the differential evolution grey wolf algorithm is used in this application to optimize the neural network model. The specific process is as follows:

[0118] a. Set parameters: Set the population size, search range, search dimension, scaling factor range, maximum number of iterations, crossover probability and other parameters of the differential evolution algorithm;

[0119] b. Initialization: Initialize the gray wolf population through mutation selection operation based on the initialization parameters, calculate the objective function value of the gray wolf individual, and select the best individual;

[0120] c. Mutation selection: calculate the distance between other gray wolves and the optimal individual and update the position;

[0121] d. Update and crossover: Update the parameters and initialize the parameters, perform crossover operations on individual positions, select new individuals and calculate the objective function value;

[0122] e. Optimization update: Update the position of the optimal individual and determine whether the maximum number of iterations has been reached. If so, end the algorithm and output the optimal solution. Otherwise, return to step b to continue execution.

[0123] Based on the differential evolution grey wolf algorithm, the optimal initial weights and thresholds of the neural network model can be determined.

[0124] In an optional embodiment, after optimizing the neural network model, the method further includes:

[0125] Construct training set data based on historical blizzard disaster data;

[0126] Based on the training set data, iterative training is performed until the number of iterations is reached. The iterative training is to input the training set data into the neural network model and optimize the neural network model through optimization methods.

[0127] In an optional embodiment, in step S130, after training the neural network model, the following further includes:

[0128] Construct test data sets, including historical snowstorm disaster event data, meteorological station observation data, land surface reanalysis data, disaster-bearing body and disaster-pregnant environment data;

[0129] Input the test set data into the blizzard disaster loss model to obtain the predicted data;

[0130] Compare the predicted data with the real data to get the accuracy of the neural network model;

[0131] Based on the accuracy of the neural network model, it is determined whether the quality of the neural network model reaches the preset target.

[0132] Specifically, Figure 2 As shown, the optimization, training and verification of the neural network model can be carried out through the following steps:

[0133] 1. Obtain historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, historical blizzard disaster data, disaster-bearing body and disaster-pregnant environment data, and pre-process each data. The pre-processing method can be based on the above method;

[0134] 2. Divide the sample set constructed by historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, historical blizzard disaster data, disaster-bearing body and disaster-pregnant environment data into a model training set and a model test set, and distinguish between label data and feature data, that is, assign labels to the sample set and the training set, assign the training set to the Y label, and assign the test set to the X label; in this application, the ratio of dividing the training set and the test set is 8:2, that is, if there are 10 sample sets, 8 of them are used as training sets, and 2 sample sets are used as test sets;

[0135] 3. Optimize the parameters of the BP neural network model through the differential evolution gray wolf algorithm, so as to determine the optimal initial weights and thresholds of the neural network model and assign them;

[0136] 4. Train the data in the training set y And the data in the test set Train x Input into the optimized neural network model for iterative training until all the training set data and test set data are iterated;

[0137] 5. Test the data in the test set x Input into the blizzard disaster loss model to obtain the predicted value Test Y′ ;

[0138] 6. Test the predicted value Y′ Test with real values Y By comparison, the snowstorm disaster loss model can be evaluated;

[0139] Furthermore, according to the characteristics of the neural network model, this application uses mean absolute error and root mean square error to evaluate the accuracy of the model, where the mean absolute error is used to measure the average absolute difference between the predicted value and the true value, indicating the average degree of deviation between the predicted value and the true value, and its calculation formula is as follows:

[0140]

[0141] Where N is the total number of samples, y i is the predicted value, x i is the real value;

[0142] The root mean square error is the square root of the ratio of the square of the deviation between the predicted value and the actual value to the number of observations. The calculation formula is as follows:

[0143]

[0144] 7. The quality of the model is judged according to the evaluation indicators NAE and RMSE of the neural network model. The closer NAE and RMSE are to 0, the higher the accuracy of the neural network model. Otherwise, the differential evolution gray wolf algorithm is used to further tune the neural network model and retrain it to continuously improve the model training effect.

[0145] It should be noted that when training the neural network model, the data of the training set Y is Train y , the test set X is Train x ; When verifying the neural network model, the test set X is Test x , that is, input the historical blizzard disaster data into the blizzard disaster loss model, obtain the predicted value of the blizzard disaster level corresponding to the historical blizzard disaster data, compare the predicted value of the blizzard disaster level with the actual value, so as to verify the accuracy of the blizzard disaster loss model.

[0146] Furthermore, the neural network model was optimized through the differential evolution grey wolf algorithm, and trained and verified through historical data, thereby ensuring that the neural network model can effectively learn from historical data and accurately predict future blizzard disasters, thereby improving the prediction accuracy and versatility of the model.

[0147] In this application, a blizzard disaster loss estimation model based on an improved BP neural network is constructed, which comprehensively considers the data-driven danger, exposure of the disaster-bearing body, vulnerability of the disaster-bearing body, and disaster-prone environment. The advantage of using the neural network method is that it can effectively extract complex characteristic parameters and information from massive data, avoiding unnecessary physical simulation analysis that wastes computing time; the advantage of using data-driven is that it can realize real-time update of dynamic loss estimates during the evolution of the disaster, providing efficient and accurate loss assessment technology for disaster prevention and mitigation of blizzard disasters.

[0148] Figure 3 A schematic diagram of a snowstorm disaster loss prediction system provided by an embodiment of the present invention is shown in FIG. Figure 3 As shown, the system includes:

[0149] The acquisition module 210 is used to acquire basic parameter data of blizzard disaster losses in the target area, wherein the basic parameter data includes blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster data, disaster-bearing body data and disaster-pregnant environment data;

[0150] A processing module 220 is used to determine characteristic quantity data of a target area based on basic parameter data;

[0151] The estimation module 230 is used to input the characteristic quantity data of the target area into the blizzard disaster loss estimation model to obtain the blizzard disaster loss level of the target area, wherein the blizzard disaster loss estimation model is trained using the blizzard disaster loss estimation model training method.

[0152] In an optional embodiment, the processing module 220 is further configured to:

[0153] Based on the blizzard disaster data, the administrative district corresponding to the blizzard disaster is obtained;

[0154] Based on the administrative districts, meteorological station observation data and blizzard disaster event data, the direct economic losses corresponding to the administrative districts were obtained.

[0155] In an optional embodiment, the processing module 220 is further configured to:

[0156] Performing time-space matching on the disaster-affected volume data to obtain the grid data of the disaster-affected volume data;

[0157] Perform time-space matching on disaster-prone environmental data to obtain raster data of disaster-prone environmental data;

[0158] The raster data of the disaster-prone body data and the raster data of the disaster-prone environment data are integrated to obtain the characteristic quantity data of the disaster-prone body and the disaster-prone environment.

[0159] The implementation principle and technical effects of the system provided in the embodiment of the present application are the same as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the system embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0160] like Figure 4 As shown, an electronic device 600 provided in an embodiment of the present application includes: a processor 601, a memory 602 and a bus, the memory 602 stores machine-readable instructions executable by the processor 601, when the electronic device is running, the processor 601 and the memory 602 communicate through the bus, and the processor 601 executes the machine-readable instructions to perform the steps of the above-mentioned blizzard disaster loss estimation method.

[0161] Specifically, the above-mentioned memory 602 and processor 601 can be general-purpose memories and processors, which are not specifically limited here. When the processor 601 runs the computer program stored in the memory 602, the above-mentioned blizzard disaster loss estimation method can be executed.

[0162] The processor 601 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 601. The above processor 601 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module may be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 602, and the processor 601 reads the information in the memory 602 and completes the steps of the above method in combination with its hardware.

[0163] Corresponding to the above-mentioned blizzard disaster loss estimation method, an embodiment of the present application also provides a computer-readable storage medium, which stores machine-executable instructions. When the computer-executable instructions are called and executed by the processor, the computer-executable instructions prompt the processor to execute the steps of the above-mentioned blizzard disaster loss estimation method.

[0164] The blizzard disaster loss estimation system based on neural network provided in the embodiment of the present application can be specific hardware on the device or software or firmware installed on the device. The implementation principle and technical effects of the device provided in the embodiment of the present application are the same as those of the aforementioned method embodiment. For the sake of brief description, the parts not mentioned in the device embodiment can refer to the corresponding contents in the aforementioned method embodiment. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can all refer to the corresponding processes in the aforementioned method embodiment, and will not be repeated here.

[0165] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0166] For another example, the flowchart and block diagram in the accompanying drawings show the possible architecture, functions and operations of the device, method and computer program product according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and a part of the module, program segment or code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0167] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0169] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can essentially or in other words, the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions for an electronic device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0170] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0171] Finally, it should be noted that the above embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the aforementioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the aforementioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A snowstorm disaster loss estimation model training method, characterized in that: include: Construct sample sets and snowstorm disaster loss estimation models; The construction of the sample set includes: A sample set is constructed based on historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, historical blizzard disaster data, disaster-bearing bodies and disaster-pregnant environment data; Performing void elimination and dimensionless processing on the sample set to obtain a snowstorm disaster and loss sample set; The construction of a snowstorm disaster loss estimation model includes: A neural network model is formed with the snowstorm disaster and loss sample set as input and the snowstorm disaster loss level as output, and parameter values ​​of the optimizer type, loss function, evaluation index, number of iterations, activation function and batch number of the neural network model are set.

2. The snowstorm disaster loss estimation model training method according to claim 1 is characterized in that: After the snowstorm disaster loss estimation model is constructed, the method further includes: optimizing the weights and thresholds of the snowstorm disaster loss estimation model by the following optimization method: Setting parameters, including population size, search range, search dimension, scaling factor range, maximum number of iterations, and crossover probability; Initialize through mutation selection operation based on initialization parameters, calculate the objective function value of the individual, and select the best individual; Calculate the distance between other individuals and the optimal individual and update the position; Update the parameters, initialize the parameters, cross-check the individual positions, select new individuals and calculate the objective function value; Update the position of the optimal individual until the maximum number of iterations is reached and output the optimal solution.

3. The snowstorm disaster loss estimation model training method according to claim 2 is characterized in that: After the weights and thresholds of the snowstorm disaster loss estimation model are optimized, the following steps are also included: Construct training set data based on historical blizzard disaster data; Based on the training set data, iterative training is performed until the number of iterations is reached. The iterative training is to input the training set data into the neural network model and optimize the neural network model through the optimization method.

4. The snowstorm disaster loss prediction model training method according to claim 3 is characterized in that: The optimization of the weights and thresholds of the snowstorm disaster loss estimation model also includes: Constructing a test set of data, the test set of data including: historical blizzard disaster event data, meteorological station observation data, land surface reanalysis data, disaster-bearing body and disaster-pregnant environment data; Inputting the test set data into the blizzard disaster loss model to obtain predicted data; Comparing the predicted data with the real data to obtain the accuracy of the neural network model; Based on the accuracy of the neural network model, it is determined that the quality of the neural network model reaches a preset target.

5. A method for estimating losses caused by a snowstorm disaster, characterized in that: include: Obtaining basic parameter data of blizzard disaster losses in the target area, the basic parameter data including blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster situation data, disaster-bearing body data and disaster-prone environment data; Determining characteristic quantity data of the target area based on the basic parameter data; The characteristic quantity data of the target area is input into a blizzard disaster loss estimation model to obtain the blizzard disaster loss level of the target area, wherein the blizzard disaster loss estimation model is trained using the blizzard disaster loss estimation model training method described in any one of claims 1-4.

6. The method for estimating snowstorm disaster losses according to claim 5, characterized in that: Determining the characteristic quantity data of the target area based on the basic parameter data also includes: Based on the blizzard disaster data, obtaining the administrative district corresponding to the blizzard disaster; Based on the administrative district, the meteorological station observation data and the blizzard disaster event data, the direct economic loss corresponding to the administrative district is obtained.

7. The method for estimating snowstorm disaster losses according to claim 5, characterized in that: Determining the characteristic quantity data of the target area based on the basic parameter data also includes: Performing time-space matching on the disaster-affected body data to obtain raster data of the disaster-affected body data; Performing time-space matching on the disaster-prone environmental data to obtain grid data of the disaster-prone environmental data; The grid data of the disaster-prone body data and the grid data of the disaster-prone environment data are integrated to obtain characteristic quantity data of the disaster-prone body and the disaster-prone environment.

8. A snowstorm disaster loss prediction system, characterized in that: include: An acquisition module is used to acquire basic parameter data of blizzard disaster losses in the target area, wherein the basic parameter data includes blizzard disaster event data, meteorological station observation data, land surface reanalysis data, blizzard disaster data, disaster-bearing body data and disaster-pregnant environment data; A processing module, used for determining characteristic quantity data of the target area based on the basic parameter data; An estimation module is used to input the characteristic quantity data of the target area into a blizzard disaster loss estimation model to obtain the blizzard disaster loss level of the target area, wherein the blizzard disaster loss estimation model is trained using the blizzard disaster loss estimation model training method described in any one of claims 1-4.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor can execute the machine executable instructions to implement the method according to any one of claims 5 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions cause the processor to execute the method described in any one of claims 5 to 7.

Images