Multi-dimensional regional economic index long and short term co-evolution mechanism analysis method, device and medium

Abstract

The invention relates to the field of data analysis, and discloses a multi-dimensional regional economic indicator long and short term co-evolution mechanism analysis method and device, and a medium, and the method comprises the steps: obtaining the index data of a research region; screening the index data of the research region by adopting an optimal parameter geographic detector (OPGD) to obtain a multi-dimensional regional economic index; inputting the multi-dimensional economic indexes into a space-time geographically weighted regression model GTWR to obtain a space-time evolution relationship of the multi-dimensional regional economic indexes; performing causal analysis on the multi-dimensional economic indexes by using a directed acyclic graph (DAG) and a constraint-based PC algorithm to obtain a causal path of the multi-dimensional regional economic indexes; and in combination with the spatio-temporal evolution relationship and the causal path of the multi-dimensional regional economic indexes, displaying and expressing a long-short term co-evolution mechanism of the multi-dimensional regional economic indexes by using an event ontology model and a affair map, and analyzing a complex driving mechanism. The method clearly reveals a multi-dimensional regional economic index co-evolution mechanism under different time and space.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data analysis, and in particular to a multi-dimensional regional economic indicator long-term and short-term collaborative evolution mechanism analysis method, device and medium. BACKGROUND

[0002] Regional economic development is one of the core fields of modern economic research. With the continuous advancement of globalization and regional integration, the diversity and complexity of regional economic activities are increasing. Traditional regional economic statistical indicators, such as GDP, employment rate and industrial structure, can provide certain quantitative information on economic activities, but their limitations in revealing the long-term evolution relationship and short-term fluctuations of regional economy have gradually emerged. In particular, in the current era of big data, how to introduce non-statistical economic indicators, build a multi-dimensional economic indicator system, and effectively quantify and analyze the collaborative evolution mechanism between these indicators has become an important challenge in measuring the dynamic development of regional economy.

[0003] Existing research mainly focuses on using vector autoregressive (VAR) models and Granger causality tests to explore the time lag effects between economic indicators (Shojaie & Fox, 2022), or through geographic detectors and geographic weighted regression (GWR) models to capture the spatial heterogeneity of economic indicators (Ding Yue et al., 2014). However, these studies often fail to fully combine the time and space dimensions, making it difficult to fully reveal the spatio-temporal dynamic evolution rules of various indicators in regional economy. The spatio-temporal weighted regression (GTWR) model is an extension of the GWR model (Fotheringham et al., 2015), which considers the time dimension of variables and can capture the spatio-temporal evolution relationship between multiple variables. However, although spatio-temporal analysis can reveal trends and changes in data, it still faces a key problem: how to accurately capture the causal relationship between multiple variables. The GTWR model can quantify the spatio-temporal evolution relationship between variables, but its structural analysis is insufficient and cannot effectively distinguish the causal paths between different variables. SUMMARY

[0004] The purpose of the present application is to provide a multi-dimensional regional economic indicator long-term and short-term collaborative evolution mechanism analysis method, device and medium, which solves the technical problem that existing analysis models lack structural analysis of economic indicators and cannot effectively distinguish the causal paths between different variables.

[0005] Specifically, the present application provides a multi-dimensional regional economic indicator long-term and short-term collaborative evolution mechanism analysis method, device and medium, the method comprising the following steps: S1, obtaining regional indicator data; S2, filtering the regional indicator data using the optimal parameter geographic detector (OPGD) to obtain multi-dimensional regional economic indicators; S3, input the multi-dimensional economic indicators into the spatio-temporal geographical weighted regression model GTWR to obtain the spatio-temporal evolution relationship of the multi-dimensional regional economic indicators; S4, utilize the directed acyclic graph DAG and the PC algorithm based on constraints to perform causal analysis on the multi-dimensional economic indicators to obtain the causal path of the multi-dimensional regional economic indicators; S5, in combination with the spatio-temporal evolution relationship and the causal path of the multi-dimensional regional economic indicators, utilize the event ontology model and the event-logic graph to display and express the long-term and short-term collaborative evolution mechanism of the multi-dimensional regional economic indicators, and analyze the complex driving mechanism.

[0006] A storage medium stores instructions and data for implementing a multi-dimensional regional economic indicator long-term and short-term collaborative evolution mechanism analysis method.

[0007] A multi-dimensional regional economic indicator long-term and short-term collaborative evolution mechanism analysis device comprises a processor and the storage medium; the processor loads and executes the instructions and data in the storage medium to implement a multi-dimensional regional economic indicator long-term and short-term collaborative evolution mechanism analysis method.

[0008] The present application provides the beneficial effects are: 1. In the OPGD-GTWR model, based on spatial heterogeneity, a plurality of economic indicators are screened, and the screened multi-dimensional economic indicators are used as input, which can effectively improve the fitting effect of the GTWR model, and capture the spatio-temporal heterogeneity between indicators, so that the spatio-temporal evolution relationship of the multi-dimensional economic indicators can be effectively analyzed.

[0009] 2. The DAG causal discovery algorithm is adopted, and the causal relationship between the multi-dimensional economic indicators is further clarified. In combination with the long-term and short-term spatio-temporal evolution quantitative relationship explored by the OPGD-GTWR and the clear structure of the DAG causal graph, the present method can accurately identify the multi-level and multi-channel action path between the multi-dimensional economic indicators, and reveal the action mechanism existing in the long-term and short-term collaborative evolution.

[0010] 3. The event ontology model and the spatio-temporal event-logic graph are utilized to explicitly express the spatio-temporal evolution relationship and the causal action path, and the collaborative evolution mechanism of the multi-dimensional regional economic indicators under different spatio-temporal conditions is clearly revealed. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 is a method flowchart of the present application; Figure 2 is a spatio-temporal evolution pattern of digital economy of a certain province; Figure 3 is a multi-dimensional economic indicator causal graph; Figure 4 is an event ontology model schematic diagram; Figure 5 is a spatio-temporal event-logic graph; Figure 6 is a schematic diagram of the coordinated evolution law of events at different times and spaces; Figure 7 is a schematic diagram of the working of the hardware device of the embodiment of the present application. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical scheme and advantages of the present application clearer, the embodiments of the present application will be further described below with reference to the drawings.

[0013] Before formally describing the present application, a general description of the scheme of the present application is given first for the convenience of understanding.

[0014] Please refer to Figure 1 , Figure 1 is a schematic diagram of the method flow of the present application; The multi-dimensional regional economic index long-term and short-term coordinated evolution mechanism analysis method provided by the present application comprises the following steps: S1, obtaining regional index data; It should be noted that the regional data in step S1 includes macro statistical indexes and non-uniform indexes in the statistical bureau and yearbook of each place, and the financial revenue is used as a proxy index of regional economic development.

[0015] As an embodiment, the present application obtains regional data, takes a province as an example, and verifies the effectiveness of the method. Due to the availability of data, the financial revenue of a province is selected as a proxy index of regional economic development in this study, which aims to explore the synergistic relationship between the index and other multi-dimensional economic indexes. The data used mainly include statistical indexes of a province, such as financial revenue, industrial structure (total value of the first, second and third industries, total import and export), financial institution deposit, total retail sales of social consumer goods, GDP growth rate, fixed asset investment growth rate, resident consumer price increase, and per capita disposable income, as well as non-statistical indexes, such as mass comment, market sentiment, population flow, digital economy, etc. Finally, the data set X is obtained: .

[0016] S2, screening the regional index data by using the optimal parameter geographic detector OPGD to obtain multi-dimensional regional economic indexes; It should be noted that in step S2, the financial revenue is used as the dependent variable, and the optimal parameter geographic detector OPGD is used to screen indexes with high spatial similarity with the financial revenue as the screened multi-dimensional regional economic indexes; the high spatial similarity specifically refers to the similarity exceeding a preset value.

[0017] As an embodiment, the screening process of the present application is carried out by using OPGD, in particular, the optimal parameter geographic detector (OPGD) is an extension of the geographic detector, aiming to improve the accuracy of the model in analyzing spatial heterogeneity and the relationship between variables by optimizing spatial grouping and parameters. The core idea of the model is: if a certain independent variable has an important influence on the dependent variable, the spatial distribution of the independent variable and the dependent variable should have high similarity. In other words, if the significant difference of fiscal revenue is caused by a certain economic indicator, the spatial distribution of the economic indicator should have certain similarity with the spatial distribution of the fiscal revenue.

[0018] The output result of the OPGD is a value ranging from 0 to 1, q q The greater the value, the stronger the explanatory power of the independent variable on the dependent variable. By analyzing the change of the value, the influence and correlation of different factors on the dependent variable can be revealed. The calculation formula of the value is as follows: q q

[0019] Wherein, is the explanatory power of the variable X on the dependent variable Y (between 0 and 1); N is the total number of samples; X is the number of samples in the i-th category (region); Y is the number of categories (regions) in which the variable X is divided; is the variance of the i-th category (region); h is the total variance. L The present application realizes OPGD by using R language, analyzes the contribution and synergistic effect of statistical indicators and non-statistical indicators on regional development, selects the main factors affecting regional economic spatial heterogeneity from the collected indicators, and forms a multi-dimensional economic indicator system with fiscal revenue h , the formula is as follows: of

[0020] Wherein, is the q value of the variable X; is the selected threshold value. S3, input the multi-dimensional economic indicators into the spatio-temporal geographic weighted regression model GTWR to obtain the spatio-temporal evolution relationship of the multi-dimensional regional economic indicators;

[0021] It should be noted that the spatio-temporal geographic weighted regression model GTWR in step S3 is as follows:

[0022] S3, input the multi-dimensional economic indicators into the spatio-temporal geographic weighted regression model GTWR to obtain the spatio-temporal evolution relationship of the multi-dimensional regional economic indicators; It should be noted that the spatio-temporal geographic weighted regression model GTWR in step S3 is as follows: ​​​​​​​

[0023] wherein, is the dependent variable; is the spatial coordinate and time point of the observation point ; is the intercept term; is the sample number; m is the intercept term; is the intercept term; is the intercept term; is the intercept term; k is the intercept term; is the intercept term; is the intercept term; is the error term; represents a multi-dimensional regional economic indicator.

[0024] As an embodiment, the spatio-temporal geographically weighted regression model (GTWR) is used to capture the spatio-temporal heterogeneity of the multi-dimensional economic indicator system The screened multi-dimensional economic indicators can effectively enhance the effect of the spatio-temporal geographically weighted regression model (GTWR), and the GTWR can improve the capture ability of the spatio-temporal dynamic relationship by introducing the time stamp, so that the regression coefficient changes with time and space.

[0025] Specifically, the GTWR first performs weighted calculation on the spatial coordinates and time to obtain local regression coefficients, which can capture the differences of local regions and time periods. The GTWR is used to reveal the dynamic changes of the multi-dimensional economic indicators in the spatial and temporal dimensions. The equation of the OPGD-GTWR model is: (4) The present application realizes the OPGD-GTWR by using R language, tests two GTWR modeling methods of fixed Gaussian and adaptive double square kernel functions, finds that the latter is better, and therefore uses the adaptive double square kernel function as the distance attenuation function and adopts the golden bandwidth selection method in the GTWR estimation.

[0026] S4, using directed acyclic graph (DAG) and constraint-based PC algorithm to perform causal analysis on the multi-dimensional economic indicators, to obtain the causal path of the multi-dimensional regional economic indicators; It should be noted that step S4 is specifically as follows: S41, initializing a complete undirected graph; S42, using statistical test to judge the conditional independence between variables; according to the set significance level, if the variables X and Y are independent under the given set Z, the edge between X and Y is removed; wherein the variable X and the variable Y belong to the multi-dimensional regional economic indicators; S43, after completing the removal of all edges, using constraint condition and data driving to determine the direction of the edge; S44. Output graph structure. The final graph structure is a directed acyclic graph (DAG), which represents the causal relationship between variables.

[0027] As one example, a directed acyclic graph (DAG) is a graphical representation used to display the flow of causal relationships between variables considered relevant in a theory or research. In a DAG, the causal flow does not contain directed cycles, meaning that starting from a node (or vertex), a directed path cannot return to the same node. Nodes in a DAG represent collected variables, while the edges connecting the nodes (often called direct edges or arrows) are generated by estimating the conditional statistical dependencies or independence between the variables.

[0028] Assume that there is a causal relationship between economic variables X, Y, and Z.

[0029] The first type of causal relationship is causal bifurcation: for example, the digital economy leads to changes in imports and exports and the tertiary industry, represented as: imports and exports ← digital economy → tertiary industry. In this case, the digital economy as a common cause implies an unconditional correlation between imports and exports and the tertiary industry. However, after controlling for the digital economy, the conditional correlation between imports and exports and the tertiary industry disappears, indicating that the digital economy as a cause masks the correlation of the common effects on imports and exports and the tertiary industry.

[0030] The second type of causal relationship is the collision relationship: Suppose that Y and Z together cause X, which can be expressed as: Y→X←Z.

[0031] In one specific embodiment of this application, the clash relationship, such as the secondary and tertiary industries jointly causing changes in social consumption, is represented as: secondary industry → social consumption ← tertiary industry. This indicates that there is no unconditional correlation between the secondary and tertiary industries, but after controlling for social consumption, a conditional correlation exists between the secondary and tertiary industries, demonstrating that the joint effect does not mask the correlation between their causes. A Directed Acyclic Graph (DAG) can be used to represent conditional independence, thus representing the flow of causal relationships. Its recursive multiplication and integration solution formula is as follows: (5) in It is the joint probability distribution of all variables; It is the first One random variable; It is a variable The parent node, that is, the node that directly affects Nodes; It is known that its parent node is In the case of variables conditional probability The application adopts Python to realize a constraint-based PC algorithm, infers the causal relationship between variables through a DAG graph structure, so that the mutual influence of each economic indicator can be clearly distinguished. The specific steps are as follows: ①Initialization of a complete undirected graph, starting from a completely connected graph, each node representing a variable; ②Conditional independence test: use statistical tests (such as chi-square test, Fisher's Z test, etc.) to judge the conditional independence between variables. According to the set significance level, if variables X and Y are independent given the set Z, remove the edge between X and Y; ③Edge direction determination: after completing the removal of all edges, use constraint conditions and data driving to determine the direction of the edges; ④Output graph structure: the final graph structure is a directed acyclic graph (DAG), representing the causal relationship between variables.

[0032] S5, combined with the spatio-temporal evolution relationship and causal path of multi-dimensional regional economic indicators, the long-term and short-term collaborative evolution mechanism of multi-dimensional regional economic indicators is displayed and expressed by using the event ontology model and the event logic graph, and the complex driving mechanism is analyzed.

[0033] It should be noted that step S5 is as follows: S51, based on steps S3 and S4, identify the relevant event elements in the regional economic collaborative evolution process; S52, after extracting the preliminary structured events, construct an event ontology model; S53, use the event ontology model to map the extracted structured events to the event logic graph, and establish the spatio-temporal event logic graph between events in combination with the spatio-temporal evolution relationship of S3 and the causal path of S4.

[0034] It should be noted that the event ontology is a kind of knowledge representation model with events as the core, which can clearly describe events, event elements and various relationships between events, and is the logical architecture of the event logic graph. The event logic graph is a logic knowledge base with dynamic events or events as the core, which describes the evolution rules and patterns between events, and helps to master the collaborative evolution rules of regional economic development. Combined with the spatio-temporal evolution relationship of S3 and the causal path of S4, the long-term and short-term collaborative evolution mechanism of multi-dimensional regional economic indicators is displayed and expressed by using the event ontology model and the event logic graph, the complex action mechanism between multi-dimensional economic indicators is analyzed in depth, the change mode of different action mechanisms in different space-time is judged, and the collaborative evolution characteristics in different stages and regions are revealed. The specific steps are as follows: ①Event element identification First, based on steps S3 and S4, identify the relevant event elements in the regional economic collaborative evolution process.

[0035] The event elements usually include: event concepts, such as economic activity events, social change events, policy intervention events, and external shock events; and key words, such as "growth", "decline", and "transition"; event subjects in event attributes, such as regional cities and governments at all levels; geographical locations, such as Fuzhou and Putian regions; and time intervals, such as the first quarter of 2018 to the fourth quarter of 2018; semantic relationships in event relationships, such as causal relationships, promotion relationships, and inhibition relationships; time sequence relationships, such as chronological relationships and time sequence relationships; and attribute relationships, such as geographical locations and time intervals.

[0036] ②Event relationship extraction and ontology construction After extracting the preliminary structured events, an event ontology model is constructed. The event ontology is a knowledge representation model for defining "regional economic events", and its core includes event type classification, event element attribute extraction, and event relationship definition.

[0037] ③Reasoning graph modeling Using the event ontology model, the extracted structured events are mapped into the reasoning graph, and the spatio-temporal reasoning graph between events is established by combining the spatio-temporal evolution relationship of S3 and the causal action path of S4.

[0038] To verify the effect of the present application, please refer to Figure 2 , Figure 2 which clearly shows the spatio-temporal evolution pattern of the digital economy of a certain province, and reveals how the digital economy indicators affect the regional economic development through spatio-temporal evolution.

[0039] The causal graph further verifies the interaction mechanism between the indicators, as shown in Figure 2 .

[0040] As a characteristic economy of a certain province, the digital economy plays a crucial role in optimizing the tertiary industry structure. Through multiple causal paths, the digital economy not only directly affects the fiscal revenue, but also indirectly affects the fiscal revenue through industrial structure optimization. The causal relationship shows that the causal relationship between multi-dimensional economic indicators is not single, but presents a multi-dimensional and multi-level interactive effect. Each indicator not only directly affects the fiscal revenue, but also promotes the optimization of the economic structure and the growth of the fiscal revenue through mutual interaction and cross-domain influence mechanism.

[0041] Combining the spatio-temporal evolution relationship of S3 and the causal action path of S4, the event ontology model of the multi-dimensional regional economic indicator cooperative evolution of a certain province is constructed as shown in Figure 4 .

[0042] Based on the event ontology model, the extracted structured events are mapped into the event graph, and the spatio-temporal event graph is established by combining the spatio-temporal evolution relationship of S3 and the causal action path of S4. The space layer and the time layer represent the attributes of the events in the ontology model and their attribute relationships, and the event layer models the specific event concepts and relationships. The event graph is as follows Figure 5 .

[0043] From the event graph, the following mechanisms exist in the coordinated evolution mechanism of the multi-dimensional economic indicators of a certain province: ① The structural linkage characteristics of the "industry-consumption-finance" path The first industry, the second industry, the third industry, population flow, total import and export, and social consumption constitute the basic path system of the financial revenue of a certain province. In regional economic fluctuations, reasonable population flow, balanced development of industries between regions, coordination, and healthy development are the "front-end variables" of financial pressure, and are also the "bottom driving force" of consumption and trade activity. The influence of the front-end variables on financial revenue presents a synchronous evolution of the coordinated structure.

[0044] ② The coupled evolution mechanism of the "digital economy-service industry-finance" path As the development focus of the characteristic economy of a certain province, digital economy exhibits a "multi-channel, multi-level" influence path, which not only directly affects financial revenue (GPBR), but also builds a complex action mechanism through the "third industry" and "import and export" channels. This influence structure shows that digital economy has played a substantial role in promoting the transformation of the service industry and the digitization of trade. The differences in the digital economy empowerment capacity of different cities in a certain province determine the spatio-temporal heterogeneity of the coordinated evolution of the "digital-service-finance" path. Therefore, the spillover effect of digital economy needs to gradually spread from the core coastal cities to the non-core mountainous cities. Overall, digital economy gradually transforms into a new dynamic force for financial revenue, and is also the core driving force for promoting the reconstruction of regional factors and the optimization of financial structure.

[0045] ③ "Intermediate variables" and nonlinear feedback mechanism Social consumption and import and export in the causal mechanism not only serve as important factors for financial revenue, but also as intermediate nodes for multi-variable action, showing a chain-like conduction feature. This structure makes the influence of multi-dimensional economic indicators on GPBR mostly not one-time, but through the conduction and feedback mechanism of the action chain. Under this mechanism, the coordination between variables is not linear superposition, but nonlinear transition effect under specific policies and event backgrounds, which is characterized by path dependence, marginal enhancement or weakening, time lag, etc. The influence of financial deposits on financial revenue also needs to be understood in this context, which may form implicit support through financing channels or policy investment.

[0046] Figure 6The cooperative evolution rule of events in different time and space is shown, under the impact of external sudden events, the lack of population mobility and the weakening of the support of most urban industries to import and export and social consumption indirectly inhibits the growth of fiscal revenue, which changes from sustained growth to low-speed growth. This change reflects that when the entity industry suffers from capacity contraction and external demand uncertainty, not only the regional economic development of the entity industry itself is weakened, but also the downstream consumption and trade decline, forming a negative linkage chain, and the heterogeneity of different action paths between regional economic development events in different time and space.

[0047] In summary, the long-term cooperative evolution mechanism of a certain province specifically embodies a multi-channel, multi-level variable cooperative evolution pattern. There are multiple cross paths and feedback mechanisms between industrial activities, consumption behavior and digital transformation, and the position and function of each core variable in the action path also dynamically adjust with the economic stage and external shocks. Therefore, the key to achieving sustainable growth of fiscal revenue lies in opening up the cooperative channels between the two main paths of "industry-consumption-fiscal revenue" and "digital-service-fiscal revenue", and optimizing the structural efficiency and reaction sensitivity of the intermediate variables, improving the structural flexibility and cooperative ability of the whole system.

[0048] See Figure 7 , Figure 7 is a hardware device working schematic diagram of the embodiment of the present application, and the hardware device specifically comprises: a multi-dimensional regional economic index long-term and short-term cooperative evolution mechanism analysis device 401, a processor 402 and a storage medium 403.

[0049] The multi-dimensional regional economic index long-term and short-term cooperative evolution mechanism analysis device 401: the multi-dimensional regional economic index long-term and short-term cooperative evolution mechanism analysis device 401 realizes the multi-dimensional regional economic index long-term and short-term cooperative evolution mechanism analysis method.

[0050] The processor 402: the processor 402 loads and executes the instructions and data in the storage medium 403 for realizing the multi-dimensional regional economic index long-term and short-term cooperative evolution mechanism analysis method.

[0051] The storage medium 403: the storage medium 403 stores instructions and data; the storage medium 403 is used for realizing the multi-dimensional regional economic index long-term and short-term cooperative evolution mechanism analysis method.

[0052] The beneficial effects of the present application are: 1. In the OPGD-GTWR model, multiple economic indicators are screened based on spatial heterogeneity, and the screened multi-dimensional economic indicators are used as input, which can effectively improve the fitting effect of the GTWR model, and capture the spatio-temporal heterogeneity between indicators, so that the spatio-temporal evolution relationship of the multi-dimensional economic indicators can be effectively analyzed.

[0053] 2. Adopting DAG causal discovery algorithm, the causal relationship between multi-dimensional economic indicators is further clarified. Combined with the quantitative relationship of long and short-term spatio-temporal evolution explored by OPGD-GTWR and the clear structure of DAG causal diagram, this method can accurately identify the multi-level and multi-channel action path between multi-dimensional economic indicators, and reveal the action mechanism existing in long and short-term collaborative evolution.

[0054] 3. The event ontology model and spatio-temporal matter map are used to explicitly express the spatio-temporal evolution relationship and causal action path, and the collaborative evolution mechanism of multi-dimensional regional economic indicators under different space-time is clearly revealed.

[0055] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for analyzing the long-term and short-term synergistic evolution mechanism of multidimensional regional economic indicators, characterized by: Includes the following steps: S1. Obtain indicator data for the study area; S2. The OPGD (Optimal Parameter Geographic Detector) was used to filter the indicator data of the study area to obtain multidimensional regional economic indicators. S3. Input the multidimensional economic indicators into the spatiotemporal geographic weighted regression model GTWR to obtain the spatiotemporal evolution relationship of the multidimensional regional economic indicators. S4. Using a directed acyclic graph (DAG) and a constraint-based PC algorithm, causal analysis is performed on multidimensional economic indicators to obtain the causal paths of multidimensional regional economic indicators. S5. Combining the spatiotemporal evolution relationship and causal path of multidimensional regional economic indicators, the event ontology model and event graph are used to express the long-term and short-term synergistic evolution mechanism of multidimensional regional economic indicators, and the complex driving mechanism is analyzed.

2. The method for analyzing the long- and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in claim 1, characterized in that: The data for the study area in step S1 includes macroeconomic statistical indicators and non-statistical indicators from local statistical bureaus and yearbooks, and uses fiscal revenue as a proxy indicator for regional economic development.

3. The method for analyzing the long- and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in claim 2, characterized in that: In step S2, fiscal revenue is used as the dependent variable, and the optimal parameter geographic detector OPGD is used to screen indicators that have a high spatial similarity with fiscal revenue as the screened multidimensional regional economic indicators; high spatial similarity specifically refers to similarity exceeding a preset value.

4. The method for analyzing the long- and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in claim 3, characterized in that: The specific details of the spatiotemporal geographic weighted regression model GTWR in step S3 are as follows: in, It is the dependent variable; For observation point Spatial coordinates and time points; For the intercept term; m The number of samples; for The Middle The first observation point k One independent variable; Independent variable The spatiotemporal weighted regression coefficients; This is the error term; This represents a multidimensional regional economic indicator.

5. The method for analyzing the long-term and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in claim 1, characterized in that: Step S4 is as follows: S41. Initialize a completely undirected graph; S42. Use statistical tests to determine the conditional independence between variables; based on the set significance level, if variables X and Y are independent under the given set Z, then remove the edge between X and Y; where variables X and Y belong to multidimensional regional economic indicators. S43. After removing all edges, use constraints and data-driven methods to determine the direction of the edges. S44. Output graph structure. The final graph structure is a directed acyclic graph (DAG), which represents the causal relationship between variables.

6. The method for analyzing the long-term and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in claim 1, characterized in that: Step S5 is as follows: S51. Based on steps S3 and S4, identify relevant event elements in the process of regional economic synergistic evolution; S52. After extracting the initial structured events, construct the event ontology model; S53. Using the event ontology model, the extracted structured events are mapped onto the event graph. The spatiotemporal event graph between events is established by combining the spatiotemporal evolution relationship in S3 and the causal action path in S4.

7. A storage medium, characterized in that: The storage medium stores instructions and data to implement the method for analyzing the long-term and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in any one of claims 1 to 6.

8. A device for analyzing the long-term and short-term synergistic evolution mechanism of multidimensional regional economic indicators, characterized in that: include: A processor and a storage medium; the processor loads and executes instructions and data in the storage medium to implement the method for analyzing the long- and short-term synergistic evolution mechanism of multidimensional regional economic indicators as described in any one of claims 1 to 6.

Images