Method for identifying influence of mining intensity on mining industry urban social-ecological system

Abstract

The invention relates to the technical field of remote sensing and geographic information, and discloses a method for identifying the influence of mining intensity on a mining industry urban society-ecosystem, which comprises the following steps: extracting year-by-year mining industry activity land vector data of a mining industry centralized development area, and establishing a space density field model; noctilucent remote sensing data and GDP statistical data are fused, and a social economic development index is constructed; integrating four key ecological factors of greenness, humidity, dryness and heat, and constructing comprehensive indexes for evaluating the regional ecological environment quality; constructing a mining city society-ecosystem analysis framework to systematically quantify an interaction relationship between social economy and ecological environment components; and analyzing the influence of the mining strength index on the interaction relationship by adopting a panel quantile regression model. According to the method, quantitative evaluation of the social-ecological system interaction relationship is realized, and the threshold feature of the influence of the mining intensity index on the system is revealed.

Description

Technical Field

[0001] This invention relates to the field of remote sensing and geographic information technology, specifically representing a method for identifying the impact of mining intensity on the socio-ecological system of mining cities. Background Technology

[0002] Mining cities, as typical socio-ecological systems, have long faced a prominent contradiction between resource development, socio-economic development, and ecological environmental protection in their sustainable development. Existing research on such systems has significant shortcomings: on the one hand, there is a general lack of systematic assessments of long-term socio-ecological system changes, making it difficult to fully reveal their dynamic evolution patterns; on the other hand, most studies neglect the dynamic processes of interaction between the social system and the ecosystem under the influence of mining activities, leading to significant limitations in understanding the mechanisms of system evolution. Although scholars have conducted considerable research on karst regions and mangrove ecosystems, the application of socio-ecological system research in the mining field remains relatively lacking. Existing studies mainly focus on static analyses such as assessment techniques and current characteristics, failing to deeply analyze the dynamic coupling relationships between system elements.

[0003] At the methodological level, existing research largely relies on qualitative indicators such as vulnerability and resilience, or traditional linear models. These methods are insufficient for effectively quantifying the complexity, nonlinearity, and multilateral characteristics of socio-ecological systems. Due to the inherent threshold characteristics and dynamic feedback mechanisms of systems, traditional assessment methods have inherent limitations in capturing the dynamic processes of system interactions and identifying key thresholds, particularly when dealing with long-term series and multi-source heterogeneous data. This methodological deficiency directly restricts the scientific formulation of sustainable development strategies for mining cities and makes it difficult to provide accurate decision support for the coordinated advancement of regional ecological protection and socio-economic development. Summary of the Invention

[0004] Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a method for identifying the impact of mining intensity on the socio-ecological system of mining cities. It aims to solve three key technical problems in current research on the socio-ecological system of mining cities: first, the lack of a systematic assessment method for long-term changes in the socio-ecological system; second, the neglect of the dynamic processes of interaction between the social system and the ecosystem under the influence of mining; and third, the difficulty of effectively quantifying the nonlinear characteristics and threshold effects of the system using traditional linear models. By integrating multi-source remote sensing data, socio-economic statistics, and machine learning methods, this invention achieves a comprehensive understanding of the socio-ecological system interaction. The quantitative assessment reveals the mining intensity index. The threshold characteristics of the impact on the system provide a scientific basis and technical support for mining cities to formulate differentiated sustainable development strategies and promote the coordinated development of human-mining social systems and ecosystems.

[0005] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a method for identifying the impact of mining intensity on the socio-ecological system of mining cities, comprising the following steps: Step 1: Based on Landsat series satellite remote sensing images, a combination of computer-automated identification and expert-guided manual visual interpretation method is used to extract annual mining activity land use vector data of concentrated mining development areas and establish a spatial density field model. The mining activity land use vector data includes mining sites, solid waste, transfer sites, tailings ponds, and ecological restoration areas. Step 2: Integrate nighttime light remote sensing data with GDP statistics to construct a socio-economic development index. ; Step 3: Integrate four key ecological factors—greenness, humidity, aridity, and temperature—to construct a comprehensive index for assessing the quality of the regional ecological environment. ; Step 4: Construct a socio-ecological system analysis framework for mining cities to systematically quantify the interactions between socio-economic and ecological environmental components. ; Step 5: Analyze the mining intensity index using a panel quantile regression model. Interaction relationship The impact.

[0006] Preferably, establishing the spatial density field model in step 1 includes: S1.1. Convert the vector data of mining activity land use into a set of centroid points representing a regular distribution of 100-meter grid. S1.2 Constructing a spatialized model based on kernel density estimation.

[0007] Preferably, step S1.2 is specifically implemented as follows: Using the centroid points of the S1.1 grid as a set, each centroid point serves as the representative density calculation core. A search radius parameter of 2000 meters is set. Based on the quadratic kernel function, the cumulative density contribution value received by each output pixel from all mining points within the search range is calculated. Continuous mining density grid surface density field raster data is generated through a spatial weighted accumulation algorithm. The mathematical expression is as follows: In the formula, Representative location point Density field raster data at that location, This represents the number of all mining points within the search radius. Represents the kernel function. Representative location point To the The Euclidean distance between mining sites This represents the search radius.

[0008] Preferably, in step S1.2, the municipal-level administrative unit represents the statistical unit. First, the average mining density within each unit is calculated. Then, a statistical relationship model between raw coal production and the average mining density is constructed using a multinomial regression method. The mathematical expression of this model is as follows: In the formula, Represents the city-level raw coal production statistics. This represents the average mining density of the corresponding unit. , , These represent the regression coefficients obtained by fitting using the least squares method; The average mining density of a unit can be obtained by overlaying density field raster data with the boundary of a municipal administrative unit, extracting the density values ​​of all pixels in each unit, and calculating their arithmetic mean. Based on the established regression relationship, the density field raster data is recalibrated to represent the standardized mining intensity index.

[0009] Preferably, in step S1.2, spatial correction is performed using a total quantity control method: First, calculate the mining intensity index within each municipal unit. The spatial integral of the initial value is then compared with the statistical raw coal production of the corresponding unit to calculate the relative error; a correction coefficient is then introduced. =Statistical output / spatial integral value, adjust the value of each pixel in the unit proportionally, repeat this process until the relative error of all city-level units is less than 5%.

[0010] Preferably, in step 2, a socio-economic development index is constructed. The specific method is as follows: By integrating two types of nighttime light remote sensing data, DMSP-OLS and NPP-VIIRS, improved nighttime light intensity time-series data were generated after correction and fusion. Subsequently, this improved nighttime light intensity time-series data was combined with GDP statistics for various regions to construct a socio-economic development index as follows: Computational model: In the formula, This represents the total nighttime light intensity of each region. The normalized equation represents the period of study.

[0011] Preferably, in step 3, a comprehensive index for assessing the quality of the regional ecological environment is constructed. The specific method is as follows: Soil-based vegetation index adjustment The core characteristic variable representing the greenness factor is expressed mathematically as follows: In the formula, and These represent the reflectance of the red and near-infrared bands of the remotely sensed image, respectively. Represents soil regulating factors; In the factor integration stage, principal component analysis was used to reduce the dimensionality of the four key factors—greenness, humidity, dryness, and heat—after standardization. The first principal component, which carries the most information, was then selected as the representative comprehensive indicator for the final analysis. .

[0012] Preferably, in step 4, the interaction between socio-economic and ecological environment components is quantified. The specific method is as follows: Interaction relationship The mathematical expression is: In the formula, Represents the coefficients of the explanatory variables, representing right The intensity of the influence; conversely, Then it means right The intensity of the impact.

[0013] Preferably, in step 4, a random forest model is used to capture the complex nonlinear relationships between the components, and the mathematical expression is: In the formula, Input sample The final predicted value, i.e. the predicted socioeconomic development index. Comprehensive indicators for prediction , It is the first Tree samples The predicted value, Represents the total number of trees in the forest; I In the formula, I It is the first The importance score of each feature It is the first Error of trees on samples outside the bag, It is the first After the value of the first feature is randomly shuffled, the second feature... Error of each tree on the same sample.

[0014] Preferably, in step 5, the panel quantile regression model expression is as follows: In the formula, The conditional quantile function representing the intensity of socio-ecological system interactions. Represents the social-ecosystem interaction, Represents a constant. Represents the explanatory variable matrix. Represents the research period, Represents the sample size of the study. represent Influence coefficient under quantile Represents the defined quantile; In the formula, Represents the influence coefficient. Represents the number of quantiles in the array. The quantile number of the quantiles Group, Represents the quantile loss function. Representing the The weighting coefficients of quantiles, Representing the The influence coefficient of quantiles.

[0015] Compared with existing technologies, this invention provides a method for identifying the socio-ecological impacts of mining intensity on mining cities, which has the following beneficial effects: 1. This invention achieves long-term dynamic assessment of the socio-ecological system of mining cities through multi-source data fusion and machine learning models, overcoming the shortcomings of existing methods in long-term time-series analysis of the socio-ecological system of mining cities.

[0016] 2. This invention effectively quantifies the nonlinear interactions between various components of the mining area's socio-ecological system by employing the random forest algorithm.

[0017] 3. This invention clarifies the weakening effect and threshold characteristics of MII on SEII by using a panel quantile regression model, providing a scientific basis for mining cities to formulate differentiated sustainable development strategies. Attached Figure Description

[0018] Figure 1 This represents the technical flowchart of the present invention; Figure 2 This represents the socio-ecological system evolution analysis framework for mining cities in this invention; Figure 3 This diagram illustrates the steps of the method of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figures 1-3 A method for identifying the impact of mining intensity on the socio-ecological system of mining cities is proposed. This method quantifies the interaction between mining activities, socio-economic development, and the ecological environment by constructing a mining intensity index, a socio-economic development index, and a comprehensive indicator. Machine learning algorithms are then used to analyze the dynamic evolution and threshold effects of these interactions. The method includes the following steps: Step 1: Based on Landsat series satellite remote sensing imagery, and employing a method combining computer-automated identification and expert-guided visual interpretation, extract annual mining activity land use vector data for concentrated mining development areas. Establish a spatial density field model to quantitatively characterize the spatial concentration of mining activities. The mining activity land use vector data includes mining sites, solid waste, transfer sites, tailings ponds, and ecological restoration areas. The construction of the spatial density field model comprises the following two core steps: S1.1. Convert the vector data of mining activity land use into a set of 100-meter grid centroids representing a regular distribution. This step transforms the areal mining activity area into a discrete set of points, where each point represents a local mining activity unit, laying the foundation for subsequent density calculations. This includes: S1.2 Constructing a spatialized model based on kernel density estimation; Using the centroids of the S1.1 grid as a central point, each centroid serves as the core for density calculation. A search radius of 2000 meters is set. Based on a quadratic kernel function, the cumulative density contribution value of each output pixel within the search range from all mining points is calculated. The closer the centroids are, the greater their contribution value. Areas with higher density values ​​represent denser mining activities per unit area, indicating a higher degree of spatial concentration in mining. A spatially weighted cumulative algorithm is used to generate continuous mining density raster surface density field raster data. This process is fully spatialized using the "Kernel Density Analysis" tool on the ArcGIS platform. The mathematical expression is as follows: In the formula, Representative location point The density field raster data at that location is the value of each cell in the final output density field raster, which intuitively represents the degree of concentration of mining activities per unit area at that location; This represents the number of all mining points within the search radius. The kernel function describes how the influence of each mining point on its surrounding space decays with distance. Representative location point To the The Euclidean distance between mining sites Represents the search radius; To establish a quantitative relationship between mining density and actual output, municipal administrative units are used as statistical units. First, the average mining density within each unit is calculated. Then, a multinomial regression method is used to construct a statistical relationship model between raw coal output and the average mining density. The mathematical expression is as follows: In the formula, Represents the city-level raw coal production statistics. This represents the average mining density of the corresponding unit. , , These represent the regression coefficients obtained by fitting using the least squares method; The average mining density of a unit can be obtained by overlaying density field raster data with the boundary of a municipal administrative unit, extracting the density values ​​of all pixels in each unit, and calculating their arithmetic mean. The goodness of fit of the regression model is evaluated by statistical indicators such as the coefficient of determination and significance test. Based on the established regression relationship, the density field grid data is recalibrated to represent the standardized mining intensity index. To ensure that the index possesses both spatial heterogeneity and statistical consistency, a total control method is used for spatial correction: First, calculate the mining intensity index within each municipal unit. The spatial integral of the initial value is then compared with the statistical raw coal production of the corresponding unit to calculate the relative error; a correction coefficient is then introduced. =Statistical output / Spatial integral value, and adjust the value of each pixel in the unit proportionally. Repeat this process until the relative error of all city-level units is less than 5%. This correction mechanism not only preserves the spatial distribution differences revealed by the original density field, but also ensures that the aggregated data of administrative units at all levels are completely matched with the statistical output, thus achieving the organic unity of the two elements of "total conservation - spatial heterogeneity". Finally, the accuracy was verified using county-level statistical raw coal production data independent of the modeling data: the coefficient of determination (R²) between the MII spatial integral value and the corresponding statistical production in each county was calculated. 2 The system evaluates the reliability of the index using statistical indicators such as root mean square error (RMSE) and mean relative error. Step 2: Integrate nighttime light remote sensing data with GDP statistics to construct a socio-economic development index. The specific method is as follows: Nighttime light remote sensing data, as a crucial form of space-based Earth observation information, can directly detect traces of human activity on the Earth's surface from satellite platforms, providing innovative research perspectives for fields such as urban boundary extraction, population estimation, urbanization monitoring, and disaster assessment. In recent years, this data has been adopted by the field of economics to characterize the intensity of regional economic activity; however, nighttime light data itself has limited ability to identify the economic added value of low-light-intensity industries such as agriculture, forestry, and mining. On the other hand, although Gross Domestic Product (GDP) is the core statistical indicator for measuring socio-economic development, it has inherent limitations in depicting regional development imbalances and differences in the distribution of actual well-being among the population. To overcome the shortcomings of these single data sources, this invention proposes to integrate nighttime light remote sensing data with GDP statistics to construct a comprehensive index called SEDI, thereby achieving a more comprehensive and accurate assessment of the level of socio-economic development. Specifically, this invention integrates two types of nighttime light remote sensing data, DMSP-OLS and NPP-VIIRS, to generate improved nighttime light intensity time-series data that has been corrected and fused. Then, it combines this improved nighttime light intensity time-series data with GDP statistics for various regions to construct a socio-economic development index as follows: Computational model: In the formula, This represents the total nighttime light intensity of each region. The normalized equations represent the period of study. Step 3: Integrate four key ecological factors—greenness, humidity, aridity, and temperature—to construct a comprehensive index for assessing the quality of the regional ecological environment. The specific method is as follows: Comprehensive indicators The RSEI (Remote Sensing Index) is a comprehensive indicator based on remote sensing technology, capable of rapidly and macroscopically assessing the quality of the regional ecological environment. This index systematically characterizes the impact of human activities on the ecological environment by integrating four key ecological factors: greenness, humidity, aridity, and heat. The calculation methods for these factors are mature and reliable, and have been widely applied and validated in relevant studies, ensuring the operability of the index construction process and the comparability of the results. This invention systematically extracts the aforementioned four key ecological factors based on remote sensing imagery. Given the low surface vegetation cover in mining areas, the traditional Normalized Difference Vegetation Index (NDVI) is highly sensitive to soil background, easily leading to a significant underestimation of greenness in such scenarios. To overcome this technical limitation and improve the accuracy and applicability of the RSEI in mining areas, this invention makes a key improvement: replacing NDVI with the Soil-Adjusted Vegetation Index (SAVI) as the core characterization variable for greenness. SAVI, by introducing an adaptive soil adjustment coefficient, can effectively compensate for and reduce spectral interference from the soil background under low vegetation cover conditions, thus more realistically reflecting the surface vegetation status. This improvement significantly enhances the stability of the RSEI index in sparsely vegetated areas and its sensitivity to ecological changes.

[0021] In the factor integration stage, this invention uses principal component analysis to reduce the dimensionality of the four standardized key factors and selects the first principal component with the largest amount of information as the final RSEI index value. Soil-regulated vegetation index The specific mathematical expression is: In the formula, and These represent the reflectance of the red and near-infrared bands of the remotely sensed image, respectively. Representing soil modifiers, although many studies typically set the "L" value to 0.5 for different regions, this value (L=0.5) does not produce optimal results in mining areas; therefore, in mining areas... Negative values ​​(such as L=-0.2) can effectively reduce soil background noise; Step 4: Construct a socio-ecological system analysis framework for mining cities to systematically quantify the interactions between socio-economic and ecological environmental components. The specific method is as follows: Because the interaction between SEDI and RSEI exhibits asymmetric characteristics, this invention constructs a socio-ecological system analysis framework for mining cities to systematically quantify the interaction relationship between its socio-economic and ecological environmental components (SEII) and to deeply analyze the threshold characteristics of MII's influence on this relationship, as well as the interaction relationships. The mathematical expression is: In the formula, Represents the coefficients of the explanatory variables, representing right The intensity of the influence; conversely, Then it means right The intensity of the impact; Specifically, in the quantification of the various components of the socio-ecological system of mining cities, mining activities are represented by the Mining Index (MII), the Ecological Environment Index (RSEI) reflects the ecological environment status, and nighttime light remote sensing data and socio-economic data are used for characterization. The Random Forest model is employed to capture the complex nonlinear relationships between components. As an ensemble learning method, the Random Forest model significantly reduces model variance and improves prediction accuracy by constructing a large number of decision trees and integrating their results, making it particularly suitable for handling high-dimensional and complex data. Due to its advantages in modeling nonlinear relationships, handling high-dimensional features, and suppressing overfitting, it has been widely applied in various research fields. The mathematical expression for using the Random Forest model to capture the complex nonlinear relationships between components is: In the formula, Input sample The final predicted value, i.e. the predicted socioeconomic development index. Comprehensive indicators for prediction , It is the first Tree samples The predicted value, Represents the total number of trees in the forest; I In the formula, I It is the first The importance score of each feature It is the first Error of trees on samples outside the bag, It is the first After the value of the first feature is randomly shuffled, the second feature... The error of each tree on the same sample, if the features are shuffled If the model error increases significantly, it means that this feature is very important, and its importance score will be high. Step 5: Analyze the mining intensity index using a panel quantile regression model. Interaction relationship The impact; This method relaxes the restrictions on the assumptions regarding the error distribution, which not only fully utilizes the large sample characteristics of panel data but also accurately represents the influence of independent variables on changes in the conditional distribution of covariates. This improves the explanatory power of the model, and its estimators exhibit better robustness and effectiveness. This method does not require the panel data to follow a normal distribution and can effectively eliminate outlier interference. It can effectively characterize the influence on the explained variable at different quantiles by controlling for the variability of explanatory variables. The panel quantile regression model expression is as follows: In the formula, The conditional quantile function representing the intensity of socio-ecological system interactions. Represents the social-ecosystem interaction, Represents a constant. Represents the explanatory variable matrix. Represents the research period, Represents the sample size of the study. represent Influence coefficient under quantile Represents the defined quantile; In the formula, Represents the influence coefficient. Represents the number of quantiles in the array. The quantile number of the quantiles Group, Represents the quantile loss function. Representing the The weighting coefficients of quantiles, Representing the The influence coefficient of quantiles.

[0022] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for identifying the socio-ecological impacts of mining intensity on mining cities, characterized in that, Includes the following steps: Step 1: Based on Landsat series satellite remote sensing images, a combination of computer-automated identification and expert-guided manual visual interpretation method is used to extract annual mining activity land use vector data of concentrated mining development areas and establish a spatial density field model. The mining activity land use vector data includes mining sites, solid waste, transfer sites, tailings ponds, and ecological restoration areas. Step 2: Integrate nighttime light remote sensing data with GDP statistics to construct a socio-economic development index. ; Step 3: Integrate four key ecological factors—greenness, humidity, aridity, and temperature—to construct a comprehensive index for assessing the quality of the regional ecological environment. ; Step 4: Construct a socio-ecological system analysis framework for mining cities to systematically quantify the interactions between socio-economic and ecological environmental components. ; Step 5: Analyze the mining intensity index using a panel quantile regression model. Interaction relationship The impact.

2. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 1, characterized in that, The step 1 of establishing the spatial density field model includes: S1.

1. Convert the vector data of mining activity land use into a set of centroid points representing a regular distribution of 100-meter grid. S1.2 Constructing a spatialized model based on kernel density estimation.

3. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 2, characterized in that, The specific method of S1.2 is as follows: Using the centroid points of the S1.1 grid as a set, each centroid point serves as the representative density calculation core. A search radius parameter of 2000 meters is set. Based on the quadratic kernel function, the cumulative density contribution value received by each output pixel from all mining points within the search range is calculated. Continuous mining density grid surface density field raster data is generated through a spatial weighted accumulation algorithm. The mathematical expression is as follows: In the formula, Representative location point Density field raster data at that location, This represents the number of all mining points within the search radius. Represents the kernel function. Representative location point To the The Euclidean distance between mining sites This represents the search radius.

4. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 3, characterized in that, In section S1.2, municipal-level administrative units represent statistical units. First, the average mining density within each unit is calculated. Then, a statistical relationship model between raw coal production and the average mining density is constructed using a multinomial regression method. The mathematical expression is as follows: In the formula, Represents the city-level raw coal production statistics. This represents the average mining density of the corresponding unit. , , These represent the regression coefficients obtained by fitting using the least squares method; The average mining density of a unit can be obtained by overlaying density field raster data with the boundary of a municipal administrative unit, extracting the density values ​​of all pixels in each unit, and calculating their arithmetic mean. Based on the established regression relationship, the density field raster data is recalibrated to represent the standardized mining intensity index.

5. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 4, characterized in that, In step S1.2, spatial correction is performed using a total quantity control method: First, calculate the mining intensity index within each municipal unit. The spatial integral of the initial value is then compared with the statistical raw coal production of the corresponding unit to calculate the relative error; a correction coefficient is then introduced. =Statistical output / spatial integral value, adjust the value of each pixel in the unit proportionally, repeat this process until the relative error of all city-level units is less than 5%.

6. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 1, characterized in that, In step 2, the socio-economic development index is constructed. The specific method is as follows: By integrating two types of nighttime light remote sensing data, DMSP-OLS and NPP-VIIRS, improved nighttime light intensity time-series data were generated after correction and fusion. Subsequently, this improved nighttime light intensity time-series data was combined with GDP statistics for various regions to construct a socio-economic development index as follows: Computational model: In the formula, This represents the total nighttime light intensity of each region. The normalized equation represents the period of study.

7. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 1, characterized in that, In step 3, a comprehensive index for assessing the quality of the regional ecological environment is constructed. The specific method is as follows: Soil-based vegetation index adjustment The core characteristic variable representing the greenness factor is expressed mathematically as follows: In the formula, and These represent the reflectance of the red and near-infrared bands of the remotely sensed image, respectively. Represents soil regulating factors; In the factor integration stage, principal component analysis was used to reduce the dimensionality of the four key factors—greenness, humidity, dryness, and heat—after standardization. The first principal component, which carries the most information, was then selected as the representative comprehensive indicator for the final analysis. .

8. The method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 1, characterized in that, In step 4, the interaction between socio-economic and ecological environment components is quantified. The specific method is as follows: Interaction relationship The mathematical expression is: In the formula, Represents the coefficients of the explanatory variables, representing right The intensity of the influence; conversely, Then it means right The intensity of the impact.

9. A method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 8, characterized in that, In step 4, a random forest model is used to capture the complex nonlinear relationships between the components. The mathematical expression is as follows: In the formula, Input sample The final predicted value, i.e. the predicted socioeconomic development index. Comprehensive indicators for prediction , It is the first Tree samples The predicted value, Represents the total number of trees in the forest; I In the formula, I It is the first The importance score of each feature It is the first Error of trees on samples outside the bag, It is the first After the value of the first feature is randomly shuffled, the second feature... Error of each tree on the same sample.

10. A method for identifying the impact of mining intensity on the socio-ecological system of mining cities according to claim 1, characterized in that, In step 5, the panel quantile regression model expression is as follows: In the formula, The conditional quantile function representing the intensity of socio-ecological system interactions. Represents the social-ecosystem interaction, Represents a constant. Represents the explanatory variable matrix. Represents the research period, Represents the sample size of the study. represent Influence coefficient under quantile Represents the defined quantile; In the formula, Represents the influence coefficient. Represents the number of quantiles in the array. The quantile number of the quantiles Group, Represents the quantile loss function. Representing the The weighting coefficients of quantiles, Representing the The influence coefficient of quantiles.

Images