Modeling method for spatial lag relationship of geographic event

Abstract

The embodiment of the invention provides a modeling method for a spatial lag relationship of geographic events, which belongs to the technical field of data processing, and specifically comprises the following steps of: 1, modeling the geographic events of a target area into spatial point elements, and modeling geographic factors into spatial surface elements; 2, screening effective causal association edges among different geographic events in the spatial point elements based on a space-time rule; and step 3, modeling an adaptive Gaussian field according to the effective causal association edge and the spatial surface element, and outputting a space lag effect quantization result. According to the scheme of the invention, the precision and robustness of spatial causal inference are significantly improved through refined filtering and dynamic modeling.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of data processing, and in particular, to a method for modeling spatial lag relationships of geographical events. Background Art

[0002] Currently, in the prior art, the spatial lag effect is often quantified by the spatial distance between geographical events, but there are the following defects: Invalid distance interference: It fails to effectively distinguish the true causal association from the noise distance of random spatial proximity, resulting in a decrease in model accuracy; Unmodeled heterogeneity effects: The heterogeneity of the geographical environment and event types is not incorporated into the spatial lag function, affecting the accuracy of parameter estimation.

[0003] It can be seen that there is an urgent need for a method for modeling spatial lag relationships of geographical events that can improve the accuracy and robustness of causal inference. Summary of the Invention

[0004] In view of this, the embodiments of the present invention provide a method for modeling spatial lag relationships of geographical events, which at least partially solves the problem of poor accuracy and robustness of causal inference in the prior art.

[0005] The embodiments of the present invention provide a method for modeling spatial lag relationships of geographical events, including: Step 1, modeling the geographical events in the target area as spatial point features and modeling the geographical factors as spatial surface features; Step 2, screening the effective causal association edges between different geographical events in the spatial point features based on spatio-temporal rules; Step 3, modeling an adaptive Gaussian field according to the effective causal association edges and the spatial surface features, and outputting the quantization result of the spatial lag effect.

[0006] According to a specific implementation manner of the embodiments of the present invention, the geographical events include longitude, latitude, occurrence time, and event type, and the geographical factors include natural factors and human factors. The natural factors include terrain, climate, soil type, and vegetation distribution, and the human factors include social and economic activities and engineering construction.

[0007] According to a specific implementation manner of the embodiments of the present invention, the specific content of the step 2 includes: Step 2.1, connecting the pairwise geographical events in the spatial point features to form spatial connection edges; Step 2.2, setting the time sequence rule as: taking any geographical event as the central event, for the central event and its adjacent events , if the occurrence time of the adjacent event is earlier than that of the central event Time of occurrence ,Right now , then delete the central event Adjacent events Corresponding spatial connection edges ; Step 2.3, setting the geographic prior knowledge rule as follows: deleting the connection edges that do not conform to the geographic process logic according to the preset invalid causal type table. The invalid causal types include geological disaster events leading to rainfall events, surface deformation events leading to rainfall events, and earthquake events leading to rainfall events. Step 2.4: Set the spatial density distribution rule as follows: divide the distance from the central event to the affected event into equal-width annular intervals, calculate the normalized spatial density of each interval, and screen valid connecting edges based on the distance decay, long-range dependency preservation, and noise edge removal rules; In step 2.5, the spatial connection edges are screened according to the time sequence rules, geographical prior knowledge rules, and spatial density distribution rules to obtain effective causal association edges.

[0008] According to a specific implementation of an embodiment of the present invention, the expression of the normalized spatial density is: in, The ring interval divided Number of geographical events affected within For the annular area The area, is the width of the divided ring, that is, the equal division step length of the ring.

[0009] According to a specific implementation of the embodiment of the present invention, step 3 specifically includes: Step 3.1, calculate the distance of each effective causal relationship edge and the ring interval index ; Step 3.2, calculate the hysteresis intensity parameter according to the distance and the index of the ring interval to which it belongs; Step 3.3, optimize the standard deviation based on MLE , and combined the hysteresis intensity parameters and spatial surface elements to fit the Gaussian field function and model the adaptive Gaussian field; In step 3.4, the spatial hysteresis effect is quantified by outputting the adaptive Gaussian field.

[0010] According to a specific implementation of the embodiment of the present invention, the optimized standard deviation The steps include: Construct a likelihood function based on the filtered effective connection edge dataset; Optimized by gradient descent or expectation maximization algorithm to maximize the probability distribution of the observed data.

[0011] According to a specific implementation manner of an embodiment of the present invention, the expression of the lag intensity parameter is wherein, represents the number of divided annular intervals.

[0012] According to a specific implementation manner of an embodiment of the present invention, the expression of the adaptive Gaussian field is wherein, represents the spatial surface feature in which the distance is spatially included.

[0013] The modeling scheme for the spatial lag relationship of geographical events in the embodiments of the present invention includes: Step 1, modeling the geographical events in the target area as spatial point features and modeling the geographical factors as spatial surface features; Step 2, screening the effective causal association edges between different geographical events in the spatial point features based on spatio-temporal rules; Step 3, modeling an adaptive Gaussian field according to the effective causal association edges and the spatial surface features, and outputting the quantization result of the spatial lag effect.

[0014] The beneficial effects of the embodiments of the present invention are as follows: Through the solution of the present invention, effective spatial connection edges are screened through a three-level filtering mechanism: 1) The time sequence rule deletes the invalid edges where the result is earlier than the cause; 2) The geographical prior rule eliminates the edges that violate geographical logic such as rainfall caused by geological disasters; 3) The spatial density distribution rule combines the normalized density calculation of the annular interval to retain the edges that conform to the distance decay and long-range dependence characteristics and eliminate the noise. Subsequently, an adaptive Gaussian field model is constructed to quantify the spatial lag effect, the intensity parameter of which is dynamically determined by the mean density of the annular interval, and the standard deviation is optimized by maximum likelihood estimation to characterize the influence range. The present invention significantly improves the accuracy of spatial causal inference through refined filtering and dynamic modeling. Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a schematic flowchart of a method for modeling the spatial lag relationship of geographical events provided by an embodiment of the present invention; Figure 2A schematic diagram of a spatial connection edge filtering process provided by an embodiment of the present invention, where (a) represents the time sequence rule, (b) represents the geographical prior knowledge rule, and (c) represents the spatial density distribution rule; Figure 3 A schematic diagram of parameter fitting of an adaptive Gaussian field model provided by an embodiment of the present invention; Figure 4 A schematic diagram of the spatial lag effect of heavy rain events on landslide events provided by an embodiment of the present invention. Detailed implementation manners

[0017] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] The following uses specific specific examples to illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0019] It should be noted that the following describes various aspects of the embodiments within the scope of the appended claims. It should be obvious that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is illustrative only. Based on the present invention, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. In addition, this device and / or this method can be implemented using other structures and / or functions in addition to one or more of the aspects described herein.

[0020] It also should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. The diagrams only show the components related to the present invention, rather than being drawn according to the number, shape and size of the components in actual implementation. The type, quantity and proportion of each component in its actual implementation can be an arbitrary change, and the component layout type may also be more complex.

[0021] In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0022] The disadvantages of the prior art include: (1) Experimental dependence limitation: The Potential Outcomes Framework needs to estimate the causal effect through intervention experiments, but geographical events (such as rainfall, earthquakes) are not intervenable, making it difficult to apply this model in practice; (2) Insufficient time series data: The Granger Causality model depends on continuous time series data, while geological disaster data (such as landslides, debris flows) usually only provides the spatial distribution and sparse timestamps of events, unable to meet the requirements of time series analysis; (3) Ignoring spatial lag and geographical environment: Although the Structural Causal Models can express causal relationships through graph models, they do not quantify the spatial lag effect (such as the propagation distance of seismic waves in vulnerable rock formations) and geographical environment heterogeneity (such as the influence of steep slopes and flatlands on the rainfall lag distance), resulting in biases in the construction of causal networks.

[0023] An embodiment of the present invention provides a method for modeling the spatial lag relationship of geographical events, which can be applied to the geographical time causal inference process in scenarios such as geological disaster early warning and environmental monitoring.

[0024] See Figure 1 , which is a schematic flowchart of a method for modeling the spatial lag relationship of geographical events provided by an embodiment of the present invention. As Figure 1 shown, the method mainly includes the following steps: Step 1, model the geographical events in the target area as spatial point features and the geographical factors as spatial surface features; In specific implementation, model the geographical entities. Geographical events are specific and observable phenomena in geographical processes, with a certain spatial and temporal range and influence, and are the basic units for describing and modeling geographical processes. Therefore, model the geographical events as spatial point features. The geographical environment refers to the spatial background that affects the occurrence of geographical events, including natural factors such as terrain, climate, soil type, vegetation distribution, etc. and human factors such as social and economic activities, engineering construction, etc. The geographical environment not only limits the occurrence range of geographical events in space but also affects the formation of spatial lag relationships through the interaction process with geographical events. Therefore, model it as a spatial surface feature.

[0025] Step 2, screen the effective causal association edges between different geographical events in the spatial point features based on spatio-temporal rules; In specific implementation, as Figure 2As shown, spatial connection edge filtering is to screen effective causal association edges between geographical events based on spatio-temporal rules. The specific steps mainly include: Time sequence rule: For the central event and its adjacent events , if the occurrence time is earlier, that is, the result is earlier than the cause, then delete the corresponding spatial connection edge ; Geographical prior knowledge rule: Delete the connection edges that do not conform to the geographical process logic according to the preset invalid causal type table. The invalid causal types include geological disaster events causing rainfall events, surface deformation events causing rainfall events, and earthquake events causing rainfall events. For example, in the geographical prior knowledge rule, the invalid causal types include: connection edges with landslides, debris flows, collapses, ground fissures, land subsidence, or unstable slope events as cause events and rainfall events as result events; connection edges with surface deformation events as cause events and rainfall events as result events; connection edges with earthquake events as cause events and rainfall events as result events. The rule explanation for deleting invalid causal connection edges based on geographical prior knowledge is shown in Table 1. Taking geological disaster prior knowledge as an example, delete the invalid causal connection edges according to the listed rules: Table 1

[0026] Spatial density distribution rule: Divide the distance from the central event to the affected event into equally wide annular intervals The equally wide annular interval has a fixed division step of , and the number of intervals is dynamically adjusted according to the maximum observation distance.

[0027] Calculate the normalized spatial density of each interval: Where: And screen effective spatial lag edges based on distance decay, long-range dependence retention, and noise edge elimination rules.

[0028] Specifically, the rule explanation of the spatial density distribution rule for screening effective connection edges is shown in Table 2. The rules for screening effective connection edges include: Distance decay and proximity: If the normalized density of the adjacent annular interval (with a smaller value) is significantly higher than that of the far neighbor interval, then retain the corresponding connection edge; Long-range dependence retention: If the normalized density of the long-distance interval ( with a larger value) Exceed the preset threshold , then retain the corresponding connection edge; Noise edge elimination: If the density of the near-neighbor interval approaches zero while the density of the far-neighbor interval is significantly non-zero, then delete the corresponding connection edge.

[0029] Table 2

[0030] Step 3: Model an adaptive Gaussian field based on the effective causal association edges and spatial surface elements, and output the quantification result of the spatial lag effect.

[0031] In specific implementation, an adaptive Gaussian function is used to quantify the spatial lag effect. According to the first law of geography, the closer the distance between two geographical events, the greater the influence on each other. Therefore, the spatial lag effect usually decays with the increase of distance. To capture this distance decay effect, an adaptive Gaussian random field is constructed based on the spatial distance between geographical events. Specific event lag distance and intensity: The spatial lag distance and the resulting influence intensity are not uniform among all geographical events. Different types of events will have different spatial influence ranges and have different degrees of influence on their surrounding environments. The spatial lag effect is also affected by the geographical environment where the event occurs. When the spatial distance between two events is the same, different lag intensities will be generated according to the characteristics of the surrounding environment. An adaptive Gaussian field is constructed for different geographical environments. Its expression is: Where in the formula is the intensity of the spatial lag effect, reflecting the source event 's influence degree on the surrounding events. Each distance corresponding spatial surface element has spatial heterogeneity differences. The intensity of the spatial lag effect is calculated from the spatial density distribution of the event connection points. Specifically, the intensity of the spatial lag effect can be estimated by the following formula: Such as Figure 3 shown, the fitting process of the standard deviation includes: Construct a likelihood function based on the filtered effective connection edge dataset; Optimize through the gradient descent or expectation maximization algorithm to maximize the probability distribution of the observed data.

[0032] Based on the constructed Gaussian field, an adaptive Gaussian field model is built according to the filtered connected edge data, and the intensity of the spatial lag effect and the influence range parameters are output. A spatial causal network diagram and a heat map of the lag effect distribution are generated.

[0033] As Figure 4 shown, the process of modeling the spatial lag effect of heavy rain events on landslide events includes: Model heavy rain events and landslide events as spatial points; Based on the rules of distance decay, long-range dependence retention, and noise edge elimination, screen the effective spatial lag edges of heavy rain events on landslide events; According to the different geographical environments where the effective spatial lag edges are located, construct an adaptive Gaussian field AGF, and use a circular heat map to express the spatial lag effect generated by heavy rain events on landslide events according to the output intensity of the spatial lag effect and the influence range parameters.

[0034] The method for modeling the spatial lag relationship of geographical events provided in this embodiment screens effective spatial connection edges through a three-level filtering mechanism: 1) The time sequence rule deletes invalid edges where the result is earlier than the cause; 2) The geographical prior rule eliminates edges that violate geographical logic such as rainfall caused by geological disasters; 3) The spatial density distribution rule combines the calculation of the normalized density in the circular interval, retains edges that conform to the characteristics of distance decay and long-range dependence, and eliminates noise. Subsequently, an adaptive Gaussian field model is constructed to quantify the spatial lag effect. Its intensity parameter is dynamically determined by the mean density of the circular interval, and the standard deviation is optimized by maximum likelihood estimation to represent the influence range. The present invention significantly improves the accuracy of spatial causal inference through refined filtering and dynamic modeling.

[0035] The method of the present invention will be further described below through a specific embodiment. The method mainly includes the following steps: Step 1: Model geographical entities. Model geographical events as spatial point features, including four pieces of information: longitude, latitude, occurrence time, and event type. Model them as spatial surface features, including spatial location and attribute information.

[0036] Step 2: Spatial connection edge filtering Input the geographical event data set, including event type, location, and time; Delete according to the time sequence rule the edges; Delete invalid causal edges according to the geographical prior rule in Table 1; Divide into circular regions, calculate the density of each interval , and apply the rules in Table 2 to screen effective edges.

[0037] Step 3: AGF model construction Calculate the distance for the retained edges and the affiliated circular interval index ; According to the formula calculate the lag strength parameter; Based on MLE optimization , fit the Gaussian field function; Output the quantization result of the spatial lag effect.

[0038] It should be understood that each part of the present invention can be implemented by hardware, software, firmware or a combination thereof.

[0039] As mentioned above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for modeling the spatial lag relationship of geographical events, characterized in that Including: Step 1: Model the geographical events in the target area as spatial point features and the geographical factors as spatial surface features; Step 2: Screen the effective causal association edges between different geographical events in the spatial point features based on spatio-temporal rules; Step 3: Model an adaptive Gaussian field according to the effective causal association edges and the spatial surface features, and output the quantification result of the spatial lag effect.

2. The method according to claim 1, characterized in that, The geographical events include longitude, latitude, occurrence time and event type. The geographical factors include natural factors and human factors. The natural factors include terrain, climate, soil type and vegetation distribution. The human factors include socio-economic activities and engineering construction.

3. The method according to claim 2, wherein The specific content of Step 2 includes: Step 2.1: Connect pairwise geographical events in the spatial point features to form spatial connection edges; Step 2.2, set the time sequence rule as: take any geographical event as the central event, for the central event Adjacent events , if the adjacent event Time of occurrence Central events that precede Time of occurrence ,Right now , then delete the central event Adjacent events Corresponding spatial connection edges ; Step 2.3: Set the geographical prior knowledge rule as: Delete the connection edges that do not conform to the geographical process logic according to the preset invalid causal type table. The invalid causal types include geological disaster events causing rainfall events, surface deformation events causing rainfall events, and earthquake events causing rainfall events; Step 2.4: Set the spatial density distribution rule as: Divide the distance from the central event to the affected event into equal-width circular intervals, calculate the normalized spatial density of each interval accordingly, and screen the effective connection edges based on the distance decay, long-range dependence retention and noise edge elimination rules; Step 2.5: Screen the spatial connection edges according to the time sequence rule, the geographical prior knowledge rule and the spatial density distribution rule to obtain the effective causal association edges.

4. The method according to claim 3, characterized in that, The expression of the normalized spatial density is Among them, is the number of affected geographical events within the divided circular interval , is the area of the circular region , is the divided circular width, that is, the equal division step length of the circular ring.

5. The method according to claim 4, wherein The specific content of Step 3 includes: Step 3.1, calculate the distance of each valid causal association edge and the index of the corresponding circular interval ; Step 3.2: Calculate the lag intensity parameter according to the distance and the circular interval index to which it belongs; Step 3.3, optimize the standard deviation based on MLE , and combine the lag intensity parameter and the spatial surface element to fit the Gaussian field function to model the adaptive Gaussian field; Step 3.4: Output the quantification result of the spatial lag effect through the adaptive Gaussian field.

6. The method according to claim 5, characterized in that, The optimized standard deviation The steps include: Construct a likelihood function based on the filtered effective connection edge dataset; Optimized by gradient descent or expectation maximization algorithm to maximize the probability distribution of the observed data.

7. The method according to claim 6, wherein The expression of the lag intensity parameter is Among them, represents the number of divided circular intervals.

8. The method according to claim 7, characterized in that, The expression of the adaptive Gaussian field is Among them, represents the distance spatial surface elements that are spatially contained.

Images