CN-121997256-A - Risk prediction method and related equipment applied to typhoon and storm weather

Abstract

The embodiment of the application belongs to the technical field of artificial intelligence, and relates to a risk prediction method and related equipment applied to typhoon and storm weather; the method comprises the steps of carrying out sliding time window cutting processing on typhoon paths and rainfall intensities of dynamic meteorological data to obtain time sequence feature data, carrying out dangerous situation prediction processing on three-dimensional risk indexes according to an integrated learning part to obtain a basic prediction result, carrying out time sequence correction processing on the time sequence feature data according to a cyclic neural network part to obtain time sequence correction data, and carrying out weighting fusion processing on the basic prediction result according to the time sequence correction data to obtain a target prediction result. The application can be used for carrying out related risk prediction processing in a financial science and technology business system, and can reflect the risk condition of the target grid more comprehensively and accurately.

Inventors

• LUO LEILEI

Assignees

• 中国平安财产保险股份有限公司

Dates

Publication Date

20260508

Application Date

20260112

Claims (10)

1. 1. A risk prediction method applied to typhoon storm weather, comprising the following steps: collecting dynamic meteorological data, current risk data and current environment data of a target grid site in real time; Calculating the product of dangerous case density and dangerous case rate according to the current dangerous case data and the current environment data to obtain a three-dimensional risk index; performing sliding time window cutting processing on typhoon paths and rainfall intensities of the dynamic meteorological data to obtain time sequence characteristic data; Invoking a trained hybrid prediction model, wherein the hybrid prediction model comprises an integrated learning part and a cyclic neural network part; Carrying out dangerous case prediction processing on the three-dimensional risk index according to the integrated learning part to obtain a basic prediction result; performing time sequence correction processing on the time sequence characteristic data according to the cyclic neural network part to obtain time sequence correction data; and carrying out weighted fusion processing on the basic prediction result according to the time sequence correction data to obtain a target prediction result.
2. 2. The method for predicting risk for typhoon and heavy rain weather according to claim 1, further comprising the steps of, before the step of invoking the trained hybrid prediction model: Acquiring historical meteorological data, historical risk-giving data and historical environment data corresponding to the target grid site; Preprocessing the historical meteorological data, the historical risk data and the historical environment data to obtain preprocessed historical meteorological data, preprocessed historical risk data and preprocessed historical environment data; constructing a history static feature according to the preprocessing history risk data and the preprocessing history environment data; constructing historical dynamic characteristics according to the historical meteorological data; constructing an initial hybrid prediction model, wherein the initial hybrid prediction model comprises an initial integrated learning part and an initial cyclic neural network part; Performing first model training processing on the initial integrated learning part according to the historical static characteristics; Performing second model training processing on the initial cyclic neural network part according to the historical dynamic characteristics; And after the first model training process and the second model training process are completed, obtaining the trained hybrid prediction model.
3. 3. The method for predicting risk of typhoon and stormy weather according to claim 2, wherein the step of preprocessing the historical meteorological data, the historical risk data and the historical environmental data to obtain preprocessed historical meteorological data, preprocessed historical risk data and preprocessed historical environmental data specifically comprises the following steps: Time alignment processing the historical meteorological data, the historical risk data and the historical environment data, and/or Performing missing value processing on the historical meteorological data, the historical risk data and the historical environment data, and/or And carrying out outlier correction processing on the historical meteorological data, the historical risk data and the historical environment data to obtain the preprocessing historical meteorological data, the preprocessing historical risk data and the preprocessing historical environment data.
4. 4. The risk prediction method for typhoon and heavy rain weather according to claim 1, wherein the step of performing time sequence correction processing on the time sequence characteristic data according to the cyclic neural network part to obtain time sequence correction data specifically comprises the following steps: constructing a spatial weight according to the time sequence characteristic data and a Gaussian kernel function, wherein the spatial weight is expressed as: Wherein, the Representing the time from the target grid location (x, y) to the typhoon center Is used for the distance of euclidean distance, A typhoon-influencing radius parameter is represented; Constructing a time weight according to the time sequence characteristic data and the risk time sequence attenuation function, wherein the real-time weight is expressed as follows: Wherein, the The login time of typhoons is indicated, Representing a sensitive time window prior to login, Representing a decay time constant; And carrying out space-time weight fusion processing on the space weight and the time weight to obtain the time sequence correction data.
5. 5. The method for predicting risk of typhoon and heavy rain weather according to claim 4, further comprising the steps of, after the step of constructing spatial weights from the time series characteristic data and gaussian kernel function: acquiring target terrain data corresponding to the target grid location according to the current environment data; and carrying out space weight optimization processing on the space weight according to the digital elevation model and the target terrain data.
6. 6. The method for predicting risk of typhoon and heavy rain weather as recited in claim 4, wherein at said time weighting Wherein the decay time constant is set when the typhoon moving speed is greater than 20km/h 。
7. 7. The method for predicting risk of typhoon and heavy rain weather according to claim 4, wherein the step of performing space-time weight fusion processing on the spatial weight and the temporal weight to obtain the time sequence correction data specifically comprises the following steps: Constructing a three-dimensional tensor according to the space weight and the time weight, wherein the three-dimensional tensor is expressed as: normalizing the three-dimensional tensor to obtain the time sequence correction data, wherein the time sequence correction data is expressed as: 。
8. 8. A risk prediction device applied to typhoon storm weather, comprising: the real-time data acquisition module is used for acquiring dynamic meteorological data, current risk data and current environment data of the target grid site in real time; The three-dimensional risk index calculation module is used for calculating the product of the dangerous case density and the dangerous case rate according to the current dangerous case data and the current environment data to obtain a three-dimensional risk index; the time sequence characteristic acquisition module is used for carrying out sliding time window cutting processing on the typhoon path and rainfall intensity of the dynamic meteorological data to obtain time sequence characteristic data; the model calling module is used for calling a trained hybrid prediction model, wherein the hybrid prediction model comprises an integrated learning part and a circulating neural network part; The dangerous case prediction module is used for carrying out dangerous case prediction processing on the three-dimensional risk indexes according to the integrated learning part to obtain a basic prediction result; the time sequence correction module is used for performing time sequence correction processing on the time sequence characteristic data according to the cyclic neural network part to obtain time sequence correction data; And the weighted fusion module is used for carrying out weighted fusion processing on the basic prediction result according to the time sequence correction data to obtain a target prediction result.
9. 9. A computer device comprising a memory and a processor, wherein the memory has stored therein computer readable instructions which when executed by the processor implement the steps of the risk prediction method of any one of claims 1 to 7 for use in typhoon heavy rain weather.
10. 10. A computer readable storage medium having stored thereon computer readable instructions which when executed by a processor implement the steps of the risk prediction method of any one of claims 1 to 7 for use in typhoon storm weather.

Description

Risk prediction method and related equipment applied to typhoon and storm weather Technical Field The application relates to the technical field of artificial intelligence, in particular to a risk prediction method and related equipment applied to typhoon and storm weather. Background When dealing with various natural disasters and sudden dangerous cases, accurate risk prediction is important to taking precautionary measures in advance and reducing loss. The traditional risk prediction method only considers single type of data, such as prediction is performed only according to meteorological data or historical risk data, and relevance and complementarity between different data are ignored. For example, predicting the possible impact of typhoons based on weather data alone may not fully take into account the amplifying or weakening effects of local environmental factors and historical risk situations, but based on historical risk data alone, it may be difficult to capture the dynamic impact of real-time weather changes on risk. Therefore, the traditional risk prediction method has the problem of low prediction accuracy. Disclosure of Invention The embodiment of the application aims to provide a risk prediction method and related equipment applied to typhoon and storm weather, so as to solve the problem of low prediction accuracy of the traditional risk prediction method. In order to solve the technical problems, the embodiment of the application provides a risk prediction method applied to typhoon and storm weather, which adopts the following technical scheme: collecting dynamic meteorological data, current risk data and current environment data of a target grid site in real time; Calculating the product of dangerous case density and dangerous case rate according to the current dangerous case data and the current environment data to obtain a three-dimensional risk index; performing sliding time window cutting processing on typhoon paths and rainfall intensities of the dynamic meteorological data to obtain time sequence characteristic data; Invoking a trained hybrid prediction model, wherein the hybrid prediction model comprises an integrated learning part and a cyclic neural network part; Carrying out dangerous case prediction processing on the three-dimensional risk index according to the integrated learning part to obtain a basic prediction result; performing time sequence correction processing on the time sequence characteristic data according to the cyclic neural network part to obtain time sequence correction data; and carrying out weighted fusion processing on the basic prediction result according to the time sequence correction data to obtain a target prediction result. In order to solve the technical problems, the embodiment of the application also provides a risk prediction device applied to typhoon storm weather, which adopts the following technical scheme: the real-time data acquisition module is used for acquiring dynamic meteorological data, current risk data and current environment data of the target grid site in real time; The three-dimensional risk index calculation module is used for calculating the product of the dangerous case density and the dangerous case rate according to the current dangerous case data and the current environment data to obtain a three-dimensional risk index; the time sequence characteristic acquisition module is used for carrying out sliding time window cutting processing on the typhoon path and rainfall intensity of the dynamic meteorological data to obtain time sequence characteristic data; the model calling module is used for calling a trained hybrid prediction model, wherein the hybrid prediction model comprises an integrated learning part and a circulating neural network part; The dangerous case prediction module is used for carrying out dangerous case prediction processing on the three-dimensional risk indexes according to the integrated learning part to obtain a basic prediction result; the time sequence correction module is used for performing time sequence correction processing on the time sequence characteristic data according to the cyclic neural network part to obtain time sequence correction data; And the weighted fusion module is used for carrying out weighted fusion processing on the basic prediction result according to the time sequence correction data to obtain a target prediction result. In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical schemes: Comprising a memory having stored therein computer readable instructions which when executed by a processor implement the steps of the risk prediction method as described above for use in typhoon storm weather. In order to solve the above technical problems, an embodiment of the present application further provides a computer readable storage medium, which adopts the following technical schemes: th