Method and device for quantifying stress effect of city expansion on green space

Abstract

The embodiment of the invention provides a method and device for quantifying the stress effect of city expansion on a green space. The method comprises the following steps: acquiring city expansion data and green space data of a plurality of city regions; the city expansion data comprises a plurality of city expansion features, and the green space data comprises a plurality of green space features; clustering the plurality of city regions according to various green space features in the green space data of the plurality of city regions; for any clustering cluster, training a machine learning model by using the city expansion data and the green space data of the city region in the clustering cluster to obtain a green space data prediction model corresponding to the clustering cluster; and explaining the prediction model by using an interpreter model, and determining the positive and negative values and the strength of the stress effect of various city expansion features on the green spatial features under the clustering cluster. In this way, the stress effect of city expansion on the green space can be effectively quantified through the combination of machine learning and the interpreter.

Description

Technical Field

[0001] The present invention relates to the field of environmental protection technology, and in particular to a method and device for quantifying the coercive effect of urban expansion on green space. Background Art

[0002] Although rapid urbanization around the world has brought opportunities for economic and social development, it has also brought major challenges to sustainable development. At present, most cities are experiencing rapid urban expansion, and the most prominent manifestation of this in space is the expansion of urban construction land, that is, the green space in urban areas, including farmland, forests, grasslands, wetlands, etc., is converted into impermeable surfaces (e.g., roads, houses, buildings). This conversion is often irreversible. Urban green space is the carrier of urban ecosystem maintenance. The generation of urban green space has a variety of environmental effects, psychological and economic benefits, such as reducing air pollution and urban heat island effect, increasing terrestrial carbon storage, reducing stress and promoting health. The negative impact of urban expansion on urban green space is called "ecological pressure" or "coercion effect", which will lead to the degradation of urban green space functions and further affect the function and pattern of urban ecosystems. In order for urbanization to proceed in a more sustainable way, how to quantify the "coercion effect" caused by urban expansion in urban planning and use this indicator to guide regional urban planning is an urgent issue. Summary of the invention

[0003] In a first aspect, an embodiment of the present invention provides a method for quantifying the coercive effect of urban expansion on green space, the method comprising:

[0004] Obtaining urban expansion data and green space data for a plurality of urban areas; wherein the urban expansion data includes a plurality of urban expansion characteristics, and the green space data includes a plurality of green space characteristics;

[0005] According to various green space characteristics in the green space data of the multiple urban areas, the multiple urban areas are clustered to obtain multiple clusters;

[0006] For any cluster, the urban expansion data of the urban areas in the cluster is used as samples and the corresponding green space data is used as labels to train the machine learning model to obtain the green space data prediction model corresponding to the cluster. The explainer model is used to interpret the green space data prediction model corresponding to the cluster to determine the positive and negative and strength of the coercive effects of various urban expansion characteristics on the green space characteristics under the cluster.

[0007] In some implementable methods of the first aspect, the multiple urban expansion characteristics are changes in urban construction land area, changes in development intensity, changes in the number of buildings, changes in the number of high-rise buildings, changes in average building height, area-weighted fractal dimension index, changes in maximum patch index, landscape shape index, changes in population density, and changes in night light index; the multiple green space characteristics are green space area, green space pattern, and green space function.

[0008] In some implementations of the first aspect, clustering the multiple urban areas according to various green space features in the green space data of the multiple urban areas includes:

[0009] According to various green space characteristics in the green space data of multiple urban areas, the K-Medoids algorithm is used to cluster multiple urban areas.

[0010] In some implementations of the first aspect, the machine learning model is a GBRT model; and the interpreter model is a SHAP model.

[0011] In some implementations of the first aspect, the green space data prediction model corresponding to the cluster is interpreted using an interpreter model to determine the positive and negative and strong coercive effects of various urban expansion characteristics on green space characteristics under the cluster, including:

[0012] The SHAP model is used to explain the green space data prediction model corresponding to the clusters, and the Shapley values ​​of various urban expansion characteristics on green space characteristics under the clusters are calculated. The positive and negative and strength of the coercive effects of various urban expansion characteristics on green space characteristics under the clusters are determined according to the positive and negative and size of the Shapley values.

[0013] In some implementations of the first aspect, the method further includes:

[0014] Based on the dependency diagram of the SHAP model, the trend and threshold of the coercive effect of various urban expansion characteristics on green space characteristics under clustering clusters are determined.

[0015] In some implementations of the first aspect, the method further includes:

[0016] The positive and negative and strong and weak coercive effects of various urban expansion characteristics on green space characteristics under each cluster are spatially mapped to obtain the spatial distribution differences of the coercive effects of various urban expansion characteristics on green space characteristics.

[0017] In a second aspect, an embodiment of the present invention provides a device for quantifying the coercive effect of urban expansion on green space, the device comprising:

[0018] An acquisition module, used to acquire urban expansion data and green space data of multiple urban areas; wherein the urban expansion data includes multiple urban expansion characteristics, and the green space data includes multiple green space characteristics;

[0019] A clustering module, used for clustering the multiple urban areas according to various green space characteristics in the green space data of the multiple urban areas to obtain multiple clusters;

[0020] The explanation module is used to train the machine learning model for any cluster, using the urban expansion data of the urban areas in the cluster as samples and the corresponding green space data as labels to obtain the green space data prediction model corresponding to the cluster; the explainer model is used to interpret the green space data prediction model corresponding to the cluster to determine the positive and negative and strength of the coercive effects of various urban expansion characteristics on green space characteristics under the cluster.

[0021] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described above.

[0022] In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium storing computer instructions, where the computer instructions are used to enable a computer to execute the method described above.

[0023] In an embodiment of the present invention, the coercive effect of urban expansion on green space can be effectively quantified through the combination of machine learning and interpreter, providing a scientific basis for formulating effective urban planning strategies and green space protection measures.

[0024] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are used to better understand the present invention and do not constitute a limitation of the present invention. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0026] Figure 1 A flow chart of a method for quantifying the coercive effect of urban expansion on green space provided by an embodiment of the present invention;

[0027] Figure 2 Schematic diagram of clustering results;

[0028] Figure 3 It is a schematic diagram showing the trend of the coercive effect of some urban expansion characteristics on green space characteristics under some clusters;

[0029] Figure 4 Schematic diagram of the threshold value of the coercive effect of urban expansion characteristics on green space characteristics under clustering;

[0030] Figure 5 Schematic diagram of the spatial distribution differences of the coercive effects of some urban expansion characteristics on green space characteristics;

[0031] Figure 6 A structural diagram of a device for quantifying the coercive effect of urban expansion on green space provided by an embodiment of the present invention;

[0032] Figure 7 The figure is a structural diagram of an exemplary electronic device capable of implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] In addition, the term "and / or" in the present invention is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects before and after are in an "or" relationship.

[0035] In order to solve the technical problems arising from the background technology, the embodiments of the present invention provide a method, device, equipment and storage medium for quantifying the coercive effect of urban expansion on green space. Specifically, urban expansion data and green space data of multiple urban areas are obtained; the urban expansion data include multiple urban expansion features, and the green space data include multiple green space features; the multiple urban areas are clustered according to the various green space features in the green space data of the multiple urban areas; for any cluster cluster, the machine learning model is trained with the urban expansion data and green space data of the urban areas in the cluster cluster to obtain a green space data prediction model corresponding to the cluster cluster; the prediction model is interpreted using an interpreter model to determine the positive and negative and strength of the coercive effect of various urban expansion features on the green space features under the cluster cluster.

[0036] In this way, the combination of machine learning and interpreters can be used to effectively quantify the coercive effect of urban expansion on green space, providing a scientific basis for formulating effective urban planning strategies and green space protection measures.

[0037] In conjunction with the accompanying drawings, a method, device, equipment and storage medium for quantifying the coercive effect of urban expansion on green space provided by an embodiment of the present invention are described in detail through specific embodiments.

[0038] Figure 1 A flow chart of a method for quantifying the coercive effect of urban expansion on green space provided by an embodiment of the present invention, such as Figure 1 As shown, the method 100 for quantifying the coercive effect of urban expansion on green space may include:

[0039] S110, obtaining urban expansion data and green space data of multiple urban areas.

[0040] Among them, the urban area is a unit area of ​​500m*500m, the urban expansion data includes a variety of urban expansion characteristics, and the green space data includes a variety of green space characteristics. Specifically, the various urban expansion characteristics are the change in urban construction land area (LC), development intensity change (DI), change in the number of buildings (BN), change in the number of high-rise buildings (HB), change in average building height (MHB), area-weighted fractal dimension index (FRAC), change in maximum patch index (LPI), landscape shape index (LSI), change in population density (POP), and change in night light index (LD); the various green space characteristics are green space area (AREA), green space pattern (GPCA), and green space function (NPP). It is worth noting that the above characteristics are all calculated based on the change in the starting year and the ending year data. Due to the different value ranges and units of different characteristics, for better comparison and modeling, the data is regularized from -100 to 100 (without changing the positive and negative value range of the indicator) through the constructed model.

[0041] As an example, the above urban expansion characteristics and green space characteristics are selected through the following steps:

[0042] Urban expansion characteristics (also known as urban expansion indicators) were selected from three aspects: urban expansion intensity, urban expansion pattern, and socio-economic aspects of urban expansion. Urban expansion intensity aims to describe the characteristics of urban construction land in two and three dimensions, and is described by five characteristics: change in urban construction land area (LC), change in development intensity (DI), change in the number of buildings (BN), change in the number of high-rise buildings (HB), and change in average building height (MHB). Urban expansion pattern aims to describe the shape, structure and other characteristics of urban construction land in space. 16 characteristics were selected, including 6 area edge characteristics, 3 shape characteristics, and 7 aggregation characteristics. All characteristics were calculated using Fragstats4.2. The 16 characteristics were screened by PCA, and the characteristics were reduced from 16 dimensions to 5 dimensions while maintaining an explained variance ratio of 95%. Finally, five characteristics, including area-weighted fractal dimension index (FRAC), maximum patch index change (LPI), landscape shape index (LSI), area-weighted proximity index (CONTIG), and aggregation index change (AI), were selected as the description of the spatial pattern of urban expansion. At present, the types and density of POIs, GDP, tertiary industry income, industrial pollution level, population, night light index, etc. are mostly used to characterize economic and social development. Based on the consideration of time dimension and data accuracy, the two characteristics of population density change (POP) and night light index change (LD) are selected as the description of urban expansion in social and economic aspects.

[0043] The variance inflation coefficient was used to test the multicollinearity of the above 12 features, and the features with values ​​greater than 10 were deleted. Finally, the area-weighted proximity index (CONTIG) and the change in the aggregation index (AI) were screened out, and the remaining 10 features were selected as urban expansion features to describe the changes in urban expansion in three dimensions. The specific selected features can be shown in Table 1.

[0044] Table 1

[0045]

[0046]

[0047]

[0048] The description of green space here is mainly carried out from three aspects, including green space area (AREA), green space pattern (GPCA), and green space function (NPP), as shown in Table 2. AREA mainly selects the area of ​​5 land use types: woodland, grassland, wetland and cultivated land. The study found that the morphology and connectivity of GPCA are mainly affected by urban built-up areas, so the two features of area-weighted fractal dimension index (FRAC_AM) and area-weighted proximity index (CONTIG-AM) are mainly selected, and PCA is used to synthesize the two features. The meaning of the synthesized feature is that the closer the GPCA value is to 0, the worse the connectivity of the patch and the more fragmented the shape. On the contrary, it also indicates that the morphological regularity and connectivity of the patch are enhanced. The green space function is described by the change in vegetation net primary productivity (NPP).

[0049] Table 2

[0050]

[0051]

[0052] S120, clustering the multiple urban areas according to various green space features in the green space data of the multiple urban areas to obtain multiple clusters.

[0053] In some embodiments, the clustering algorithm used here is the K-Medoids algorithm. Specifically, the K-Medoids algorithm uses the object closest to the center as the center point. PAM is the most widely used method to implement the K-Medoids algorithm. Here, each urban area of ​​500m*500m is regarded as a sample, and the data set used for clustering contains N samples, each of which is defined by three green space features of AREA, GPCA, and NPP, that is, each sample has three dimensions. In the K-Medoids algorithm, the K value needs to be set in advance and repeatedly verified. According to the relevant technology, the extended variance method can be used to determine the K value. Here, K is expected to be less than an acceptable error, and its approximate range is determined based on the analysis of the landscape characteristics of the study area by urban planning and landscape planning professionals. In addition, the clustering results of different K values ​​are compared here, and the most suitable K value is found. In the clustering process, if the K value is too large, more clusters are obtained, and the similarity of some types is high. If the K value is too small, degraded types with different characteristics may be clustered into one class, thereby affecting subsequent analysis.

[0054] As an example, the K-Medoids algorithm is implemented using the kmedoids package in Python. The specific steps are as follows:

[0055] 1) Randomly select K points from all samples as the center points, denoted as OK , and then calculate the remaining points one by one (denoted as M, M = NO K ) to the center point, denoted as D (O,M) , the calculation formula of Euclidean distance is as follows:

[0056]

[0057] x M =(x 1M ,x 2M ,…,x iM ) and x O =(x 1O ,x 2O ,…,x iO );

[0058] 2) According to the value D(O,M), point M is assigned to the nearest center point O K The cluster where it is located and the initial cluster is obtained.

[0059] 3) Replace the center point O with the remaining points M K , and evaluate the cost of this behavior, that is, replacing the midpoint M of the cluster before and after to the center point O K The Euclidean distance difference is denoted by C MK . Calculate the total cost T MK , to determine whether the center point O can be replaced by the point OM K , total cost T MK The calculation formula is as follows:

[0060]

[0061] Among them, T MK <0 means that after the above replacement behavior, the total deviation of clustering becomes smaller, point O M Replace the center point O K Become the new center point.

[0062] 4) Repeat steps 2) and 3) until T MK No longer changes.

[0063] It should be noted that the benefits of clustering here are:

[0064] The composition and development characteristics of green space in different urban areas are different. If a unified model is constructed to fit the relationship between urban expansion and green space, the explanatory power of the model will be weakened, and it will not be able to accurately explain the coercive effect of urban expansion on green space. Here, the K-Medoids algorithm is used to divide urban areas into several clusters with similar green space characteristics. It can be seen that the urban expansion characteristics within the same cluster have similar effects on green space coercion. Subsequently, each cluster can be modeled and analyzed separately, so as to accurately capture the coercive effect of urban expansion on green space.

[0065] S130, for any cluster, the urban expansion data of the urban area in the cluster is used as a sample, and the corresponding green space data is used as a label to train the machine learning model to obtain the green space data prediction model corresponding to the cluster; the green space data prediction model corresponding to the cluster is interpreted using the interpreter model to determine the positive and negative and strength of the coercive effects of various urban expansion characteristics on the green space characteristics under the cluster.

[0066] Specifically, for any cluster, the urban expansion data of the urban area in the cluster is used as a sample, and the corresponding green space data is used as a label to train the GBRT model to obtain the green space data prediction model corresponding to the cluster; the green space data prediction model corresponding to the cluster is interpreted using the SHAP model, and the Shapley value of various urban expansion characteristics on green space characteristics under the cluster is calculated. According to the positive and negative and size of the Shapley value, the positive and negative and strength of the coercive effect of various urban expansion characteristics on green space characteristics under the cluster are determined, that is, if the Shapley value is positive, it is a positive contribution to green space coercion; if the Shapley value is negative, it is a negative contribution to green space coercion, thereby quantifying the coercive effect of urban expansion on green space and summarizing the key urban expansion characteristics that affect urban green space in each dimension. On this basis, the trend and threshold of the coercive effect of various urban expansion characteristics on green space characteristics under the cluster can also be determined based on the dependency plot of the SHAP model. At the same time, the positive and negative and strength of the coercive effects of various urban expansion characteristics on green space characteristics under each cluster can be spatially mapped to obtain the spatial distribution differences of the coercive effects of various urban expansion characteristics on green space characteristics and visualize them.

[0067] As an example, two machine learning methods, the GBRT model and the SHAP model, are used here to explore the relative importance of various characteristics of urban expansion to green space characteristics, as well as the marginal effects between the two, to determine the types of coercion and the differences in contribution. The GBRT model is a machine learning model based on a decision tree regression model and a gradient boosting model. The GBRT model is based on the gradient boosting method. By building many decision trees, these decision trees are combined into a "strong learner" in an iterative manner to improve the prediction accuracy. The main advantage of the GBRT model is that the residual of the previous iteration is reduced through each iteration calculation. A new regression tree model is established according to the gradient direction of the residual reduction. These iterations will gradually build a more complete regression tree model. Finally, the prediction result is obtained by integrating the prediction results of the regression tree model.

[0068] Machine learning models can often achieve the expected accuracy, but their application is often limited because the features of the model cannot be reasonably explained. Therefore, it is difficult to obtain instructive suggestions for its application. To solve this problem, the SHAP model was proposed to explain the output of machine learning models. The explanation of the SHAP model is to attribute the output value to the Shapley value of each feature. The SHAP model provides a method to estimate the contribution of each feature based on game theory and local explanation. In the SHAP model, the contribution of each feature to the model output is allocated according to its marginal contribution. In other words, the SHAP model can calculate the Shapley value of each feature to measure its impact on the final output value.

[0069] Therefore, the results of the GBRT model are input into the SHAP model to explain the coercive relationship between urban expansion and green space. The SHAP model uses the shap package in python. The GBRT model uses the scikit-learn package in python. The 10 selected urban expansion characteristics are used as independent variables, and AREA, GPCA, and NPP are used as dependent variables. The model introduces some randomness to reduce overfitting, and a ten-fold cross validation is used to ensure its stability. Since the characteristics of each cluster are different, the parameters (such as tree complexity and learning rate) need to be set separately and adjusted according to the goodness of fit of the model.

[0070] GBRT Model:

[0071] Training samples T = ((x1, y1), (x2, y2), ..., (xn, yn)), differentiable loss function L(y, F(x)) and number of iterations M. The specific training steps of the GBRT model are as follows:

[0072] 1) Initialize the model

[0073]

[0074] 2) For m=1 to M:

[0075] a. Calculate the so-called pseudo residuals:

[0076]

[0077] b. Match the base learner (such as a tree learner) hm(x) with the pseudo residual, i.e., train it using the training set.

[0078] c. Calculate the multiplier γ by solving the following one-dimensional optimization problem m :

[0079]

[0080] d. Update the model:

[0081] F m (x) = F m―1 (x)+γh m (x);

[0082] 3) The iteration ends and the final model is output: F(x)=FM(x).

[0083] SHAP Model:

[0084]

[0085] Where g is the explained model; M is the number of input features; Z represents the presence of the corresponding feature (1 or 0); is the attributed value (Shapely value) of each feature, is a constant. All possible feature combinations in the tree model (including different orders) must calculate the Shapley value. Then the weighted summation method is used to calculate the sum. The calculation formula is as follows:

[0086]

[0087] in, is the contribution rate of feature k; K is the number of features; x represents a data point, |z'| is the number of non-zero entries in z'; f(x) is the output of the original model, that is, the value of the three features of the green space. Its formula is:

[0088]

[0089] Among them, α0 is the benchmark value, which represents the output when all input parameters are missing, usually the average value of the target variable in all samples; g(x') is the explanatory model.

[0090] When analyzing the impact of urban expansion on green space, it is often helpful to quantify the relative importance of each independent variable. Relative importance can be calculated by taking the average of the absolute values ​​of all Shapley values ​​corresponding to each feature:

[0091]

[0092] Among them, I k represents the relative importance of the feature; N represents the number of samples.

[0093] In summary, the present invention achieves at least the following technical effects:

[0094] 1) The present invention uses an innovative method to conduct an in-depth analysis of the impact of urban expansion and green space. It not only deconstructs urban expansion and green space into three dimensions respectively, but also provides a more detailed and comprehensive description of these two areas by selecting multiple indicators. This multi-dimensional analysis method can more accurately capture the changes in green space during urban development, thereby providing more accurate data support for urban planners. Through this method, the specific impact of urban expansion on green space can be more clearly understood, including the reduction of green space area, the degradation of ecological functions, and the loss of biodiversity. In addition, this detailed description also helps to identify and evaluate potential risks and opportunities in the process of urban expansion, and provides a scientific basis for the formulation of effective urban planning strategies and green space protection measures. In short, the application of the present invention will greatly promote the sustainable development of cities and the rational use of green space.

[0095] 2) Unlike other methods for exploring the impact of urban expansion on green space that use simple correlation techniques, the method provided by the present invention is based on the actual situation of the city. By calculating the changes in long-term data, it focuses on exploring the inherent connections between multivariate indicators, nonlinear change trends, and specific thresholds of impact on green space, and can provide more accurate data references for urban planning.

[0096] 3) The method of combining the GBRT model and the SHAP model introduced in the present invention is an advanced data analysis technology that aims to improve the interpretability and prediction accuracy of the model. Not only can a powerful prediction model be obtained, but also the decision-making process of the model can be deeply understood. The GBRT model provides powerful prediction capabilities, while the SHAP model provides an in-depth explanation of these predictions. This combination can identify the factors that have the greatest impact on urban expansion and green space, and how these factors interact with each other. In this way, the technical solution can provide urban planners with a clearer perspective and help them develop more effective strategies to balance urban development and environmental protection.

[0097] 4) Different from other qualitative research effects techniques, this paper classifies and grades the results of the impact of urban expansion on green space, and uses a combination of the GBRT model and the SHAP model to quantify and visualize the "coercion effect", thereby achieving a comprehensive quantitative evaluation of the impact of multiple factors and measuring the degree of influence of different factors.

[0098] 5) The method provided by the present invention is suitable for quantitative evaluation of "stress effect" in any region within a specific period of time. Moreover, the method has low requirements for the amount and type of data, and only requires obtaining land use data at two time points (starting year and ending year); all data can be obtained from remote sensing images.

[0099] Taking a certain city (denoted as A) as an example, the method 100 for quantifying the coercive effect of urban expansion on green space provided by the embodiment of the present invention is described in detail as follows:

[0100] (1) Obtain urban expansion data and green space data for multiple urban areas in City A.

[0101] Among them, the urban area is a unit area of ​​500m*500m, the urban expansion data includes a variety of urban expansion characteristics, and the green space data includes a variety of green space characteristics. Specifically, the various urban expansion characteristics are the change in urban construction land area (LC), development intensity change (DI), change in the number of buildings (BN), change in the number of high-rise buildings (HB), change in average building height (MHB), area-weighted fractal dimension index (FRAC), change in maximum patch index (LPI), landscape shape index (LSI), change in population density (POP), and change in night light index (LD); the various green space characteristics are green space area (AREA), green space pattern (GPCA), and green space function (NPP). It is worth noting that the above characteristics are all calculated based on the change in the starting year and the ending year data. Due to the different value ranges and units of different characteristics, for better comparison and modeling, the data is regularized from -100 to 100 (without changing the positive and negative value range of the indicator) through the constructed model.

[0102] (2) Based on the various green space characteristics in the green space data of multiple urban areas, the K-Medoids algorithm is used to cluster the multiple urban areas to obtain multiple clusters. The clustering results can be as follows: Figure 2 shown.

[0103] (3) For any cluster, the urban expansion data of the urban area in the cluster is used as samples, and the corresponding green space data is used as labels to train the machine learning model to obtain the green space data prediction model corresponding to the cluster; the green space data prediction model corresponding to the cluster is interpreted using the interpreter model to determine the positive and negative and strong coercive effects of various urban expansion characteristics on green space characteristics under the cluster. On this basis, the trend and threshold of the coercive effects of various urban expansion characteristics on green space characteristics under the cluster can also be determined based on the dependency plot of the SHAP model, as follows: Figure 3 , Figure 4 At the same time, the positive and negative and strong and weak coercive effects of various urban expansion characteristics on green space characteristics under each cluster can be spatially mapped to obtain the spatial distribution differences of the coercive effects of various urban expansion characteristics on green space characteristics and visualize them, as shown in the following figure. Figure 5 shown.

[0104] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0105] The above is an introduction to a method embodiment. The following is a further explanation of the solution of the present invention through an apparatus embodiment.

[0106] Figure 6 A structural diagram of a device for quantifying the coercive effect of urban expansion on green space provided by an embodiment of the present invention, such as Figure 6 As shown, the device 600 for quantifying the coercive effect of urban expansion on green space may include:

[0107] The acquisition module 610 is used to acquire urban expansion data and green space data of multiple urban areas; wherein the urban expansion data includes multiple urban expansion characteristics, and the green space data includes multiple green space characteristics.

[0108] The clustering module 620 is used to cluster the multiple urban areas according to various green space features in the green space data of the multiple urban areas to obtain multiple clusters.

[0109] The explanation module 630 is used to train the machine learning model for any cluster, using the urban expansion data of the urban areas in the cluster as samples and the corresponding green space data as labels to obtain the green space data prediction model corresponding to the cluster; use the explainer model to interpret the green space data prediction model corresponding to the cluster, and determine the positive and negative and strength of the coercive effects of various urban expansion characteristics on the green space characteristics under the cluster.

[0110] Understandably, Figure 6 The various modules / units in the device 600 for quantifying the coercive effect of urban expansion on green space have the following functions: Figure 1 The functions of the various steps in the method 100 for quantifying the coercive effect of urban expansion on green space shown in the figure, and the ability to achieve their corresponding technical effects, are not described in detail here for the sake of brevity.

[0111] Figure 7 700 is a block diagram of an exemplary electronic device capable of implementing an embodiment of the present invention. Electronic device 700 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 700 may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown in the present invention, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required in the present invention.

[0112] like Figure 7 As shown, the electronic device 700 may include a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 may also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0113] Multiple components in the electronic device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the electronic device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0114] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as method 100. For example, in some embodiments, the method 100 may be implemented as a computer program product, including a computer program, which is tangibly contained in a computer-readable medium, such as a storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 700 via ROM 702 and / or a communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the method 100 described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the method 100 in any other appropriate manner (e.g., by means of firmware).

[0115] The various embodiments described above in the present invention can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs, which may be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general programmable processor, which may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0116] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0117] In the context of the present invention, a computer-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a computer-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0118] It should be noted that the present invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute method 100 and achieve the corresponding technical effect achieved by executing the method in an embodiment of the present invention. For the sake of concise description, they will not be repeated here.

[0119] In addition, the present invention also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method 100 is implemented.

[0120] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and the present invention is not limited here.

[0121] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for quantifying the coercive effect of urban expansion on green space, characterized in that: The method comprises: Acquire urban expansion data and green space data of a plurality of urban areas; wherein the urban expansion data includes a plurality of urban expansion characteristics, and the green space data includes a plurality of green space characteristics; According to various green space characteristics in the green space data of the multiple urban areas, the multiple urban areas are clustered to obtain multiple clusters; For any cluster, the urban expansion data of the urban area in the cluster is used as a sample, and the corresponding green space data is used as a label to train the machine learning model to obtain the green space data prediction model corresponding to the cluster; the green space data prediction model corresponding to the cluster is interpreted using the interpreter model to determine the positive and negative and strength of the coercive effects of various urban expansion characteristics on green space characteristics under the cluster.

2. The method according to claim 1, characterized in that The various urban expansion characteristics are respectively changes in urban construction land area, changes in development intensity, changes in the number of buildings, changes in the number of high-rise buildings, changes in average building height, area-weighted fractal dimension index, changes in the maximum patch index, landscape shape index, changes in population density, and changes in night light index; The various green space characteristics are green space area, green space pattern, and green space function.

3. The method according to claim 1, characterized in that The clustering of the multiple urban areas according to various green space features in the green space data of the multiple urban areas includes: According to various green space characteristics in the green space data of multiple urban areas, the K-Medoids algorithm is used to cluster multiple urban areas.

4. The method according to claim 1, characterized in that: The machine learning model is a GBRT model; the interpreter model is a SHAP model.

5. The method according to claim 4, characterized in that The use of the interpreter model to interpret the green space data prediction model corresponding to the cluster cluster to determine the positive and negative and strong coercive effects of various urban expansion characteristics on green space characteristics under the cluster cluster include: The SHAP model is used to interpret the green space data prediction model corresponding to the cluster, and the Shapley value of various urban expansion characteristics on the green space characteristics under the cluster is calculated. The positive and negative and strength of the coercive effect of various urban expansion characteristics on the green space characteristics under the cluster are determined according to the positive and negative and size of the Shapley value.

6. The method according to claim 5, characterized in that The method further comprises: Based on the dependency graph of the SHAP model, the trend and threshold of the coercive effect of various urban expansion characteristics on green space characteristics under the cluster are determined.

7. The method according to claim 5, characterized in that The method further comprises: The positive and negative and strong and weak coercive effects of various urban expansion characteristics on green space characteristics under each cluster are spatially mapped to obtain the spatial distribution differences of the coercive effects of various urban expansion characteristics on green space characteristics.

8. A device for quantifying the coercive effect of urban expansion on green space, characterized in that: The device comprises: An acquisition module, used to acquire urban expansion data and green space data of multiple urban areas; wherein the urban expansion data includes multiple urban expansion characteristics, and the green space data includes multiple green space characteristics; A clustering module, used for clustering the multiple urban areas according to various green space characteristics in the green space data of the multiple urban areas to obtain multiple clusters; The explanation module is used to train a machine learning model for any cluster, using the urban expansion data of the urban area in the cluster as samples and the corresponding green space data as labels to obtain a green space data prediction model corresponding to the cluster; use the explainer model to interpret the green space data prediction model corresponding to the cluster, and determine the positive and negative and strength of the coercive effects of various urban expansion characteristics on green space characteristics under the cluster.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images