Urban green development level comprehensive evaluation method based on digital technology

Abstract

The invention relates to the technical field of urban green development evaluation, and particularly discloses an urban green development level comprehensive evaluation method based on a digital technology, which comprises the following steps: constructing an urban green development level comprehensive evaluation index system and a digital technology development level comprehensive evaluation index system; urban index data are selected and preprocessed; a comprehensive evaluation model is constructed based on an entropy weight-TOPSIS method, the entropy weight method is used for determining the weight of each index, and the TOPSIS method is used for determining an evaluation index; analyzing a green development level index and a digital technology development level index of each city by using a comprehensive evaluation model; a spatial statistical analysis method, a random forest model and a VAR model are adopted to analyze a space-time effect and an interaction mechanism of digital technology development and urban green development. According to the method adopted by the invention, the limitation of traditional static evaluation is broken through, multi-dimensional coupling analysis of the digital technology and urban green development is realized, and scientific support can be provided for formulating a regional differentiation green development strategy.

Description

Technical Field

[0001] This invention relates to the field of urban green development evaluation technology, and in particular to a comprehensive evaluation method for urban green development level based on digital technology. Background Technology

[0003] Digital technology, as a key force driving the transformation and upgrading of modern industries, is triggering profound changes globally. Emerging technologies such as artificial intelligence, cloud computing, blockchain, and big data are profoundly reshaping the economic and social development landscape, comprehensively permeating all areas of production, life, and ecological governance. In urban green development practices, digital technology, by empowering key scenarios such as smart grids, intelligent transportation, and precision agriculture, has not only promoted the green transformation of traditional infrastructure and the transformation and upgrading of energy-intensive industries, but also realized the digital monitoring and intelligent management of the ecological environment, providing strong technical support for accelerating the construction of green and smart cities and enhancing the sustainable development capabilities of cities. Current research on digital technology and urban green development has made significant progress, and a preliminary theoretical framework for digital technology empowering green development has been established.

[0004] However, there are still several limitations in the existing framework that need to be addressed: First, most studies adopt a static analysis paradigm, which fails to fully analyze the complex dynamic coupling mechanism between digital technology and green development; second, although the green development evaluation system is becoming more and more complete, quantitative research on the empowerment path of digital technology is still weak; finally, urban green development research focuses on connotation interpretation and case analysis, lacks cross-regional and long-term dynamic tracking, and is difficult to reveal its spatiotemporal heterogeneity characteristics and evolutionary laws.

[0005] Therefore, there is an urgent need to build a technical system that can accurately characterize spatiotemporal effects, identify key driving factors, and predict co-evolutionary trends, so as to break through the bottlenecks of the static and localized nature of existing evaluation methods and provide scientific support for the formulation of regionally differentiated green development strategies. Summary of the Invention

[0006] The purpose of this invention is to provide a comprehensive evaluation method for urban green development based on digital technology. This method addresses the problems of insufficient dynamic monitoring, weak regional coordination, and lack of quantitative analysis of influencing mechanisms in existing urban green development evaluation and digital technology empowerment paths. Through multi-dimensional data fusion, spatiotemporal effect analysis, and intelligent prediction, it achieves precise matching and synergistic optimization between digital technology and urban green development.

[0007] To achieve the above objectives, this invention provides a comprehensive evaluation method for urban green development level based on digital technology, comprising the following steps:

[0008] S1. Construct an indicator system, including a comprehensive evaluation indicator system for the level of urban green development and a comprehensive evaluation indicator system for the level of digital technology development;

[0009] S2. Select city indicator data and preprocess it;

[0010] S3. Construct a comprehensive evaluation model based on the entropy weight-TOPSIS method, where the entropy weight method is used to determine the weight of each indicator, and the TOPSIS method is used to determine the evaluation index.

[0011] S4. Use a comprehensive evaluation model to analyze the green development level index and digital technology development level index of each city;

[0012] S5. Using spatial statistical analysis, random forest model and VAR model, we analyze the spatiotemporal effects and interaction mechanism between digital technology development and urban green development.

[0013] Preferably, in S1, a comprehensive evaluation index system for green development level is constructed, including four primary indicators: economic development, social development, environmental development, and resource development.

[0014] Economic development includes three secondary indicators: per capita GDP, number of registered unemployed urban residents, and the proportion of the tertiary sector's added value to GDP.

[0015] Social development includes three secondary indicators: education expenditure, per capita urban road area, and the number of employees enrolled in basic medical insurance.

[0016] Environmental development includes three secondary indicators: the rate of harmless treatment of domestic waste, the green coverage rate of built-up areas, and the sewage treatment rate.

[0017] Resource development includes two secondary indicators: per capita daily domestic water consumption and the number of industrial enterprises above a designated size.

[0018] Preferably, in S1, a comprehensive evaluation index system for the development level of digital technology is constructed, including three primary indicators: digital infrastructure and mobile applications, digital industry investment, and digital technology development.

[0019] Digital infrastructure and mobile applications include three secondary indicators: the number of international internet users, the number of mobile phone users at the end of the year, and the annual electricity consumption.

[0020] Digital industry investment includes two secondary indicators: science expenditure and internal R&D expenditure.

[0021] The development of digital technology includes three secondary indicators: the number of patents granted, the number of employees in information transmission, computer services and software industries, and telecommunications revenue.

[0022] Preferably, in S3, the entropy weight method is used to determine the weights of each indicator, and the steps are as follows:

[0023] Positive indicators:

[0024]

[0025] Negative indicators:

[0026]

[0027] Where i represents a city, i = 1, 2, ..., m, j represents a measurement indicator, j = 1, 2, ..., n; x ij X represents the raw data for the j-th indicator in the i-th city. ij x represents the standardized data of the j-th indicator for the i-th city; min(j) x max(j) Let represent the minimum and maximum values ​​of the j-th indicator, respectively.

[0028] Calculate the specific gravity p ij :

[0029]

[0030] Calculate the information entropy e of the j-th measurement index. j :

[0031]

[0032] Finally, calculate the entropy weight w j for:

[0033]

[0034] The TOPSIS method determines the evaluation index as follows:

[0035] Construct the weighted matrix T ij :

[0036]

[0037] Determine the positive and negative ideal solutions:

[0038]

[0039] Calculate the Euclidean distance between the index and the positive and negative ideal solutions:

[0040]

[0041] Calculate the relative closeness C of the optimal solution. i :

[0042]

[0043] C iThe higher the value, the higher the city's overall evaluation score.

[0044] Preferably, in S5, the spatial statistical analysis method includes:

[0045] The global Moran index is used to analyze the spatial relationships of urban green development levels using a spatial adjacency matrix, and the local Moran index map is used to identify specific clustering patterns of green development.

[0046] The dynamic evolution of urban green development level is analyzed using a nonparametric kernel density estimation method based on Gaussian function.

[0047] A spatial Markov chain model is used to analyze the state transition probability and spatial spillover effect of urban green development level.

[0048] Preferably, in S5, an ensemble learning method is used to construct a random forest model containing 500 decision trees, the minimum number of leaf node samples is set to 5, and out-of-bag prediction and out-of-bag prediction variable importance assessment are enabled.

[0049] Partial dependency plots illustrate the marginal effect of a single or multiple independent variables on the dependent variable. The calculation process for a partial dependency plot of a single independent variable is as follows:

[0050] By fixing the values ​​of other independent variables, traversing the value space of the target independent variable, and calculating the average predicted value of the model for each value, the relationship curve between the average predicted value and the value of the target independent variable is the partial dependency graph.

[0051] Friedman's statistic is used to measure the feature interaction strength H, and its calculation formula is as follows:

[0052]

[0053] Among them, PDP i'j' This represents the partial dependency value of the combination of features i′ and j′, while PDP i' and PDP j' H represents the partial dependency values ​​of features i′ and j′ individually, respectively. The larger the H value, the stronger the interaction between features i′ and j′.

[0054] The study selected digital technology level, per capita GDP, the proportion of tertiary industry added value in GDP, per capita urban road area, and average annual population as variables affecting the green development level of urban agglomerations for analysis.

[0055] Preferably, in S5, a VAR model is used to analyze the dynamic interaction between digital technology and the level of green development in urban agglomerations. The steps are as follows:

[0056] First, the green development level index and digital technology development level index obtained from S4 are subjected to the ADF unit root test to test the stationarity of the original time series.

[0057] If the original time series is a non-stationary series, first-order differencing or logarithmic differencing is used until the stationarity test is passed. If both variables are integrated series of the same order, the Johansen cointegration test is further used to determine whether there is a long-term equilibrium relationship between the green development level index and the digital technology development level index.

[0058] Next, the order of the VAR model is determined based on the AIC, BIC and HQIC indices, the VAR model is established, and the stability of the VAR model is judged by the eigenvalue plot.

[0059] Finally, impulse response analysis is used to demonstrate the specific paths and time-delay effects of the impact, while variance decomposition is used to quantify the contribution of each variable to the system fluctuations, and the trend of digital technology and urban green development level is predicted.

[0060] Therefore, the present invention adopts the above-mentioned comprehensive evaluation method for urban green development level based on digital technology, and the beneficial effects are as follows:

[0061] (1) This invention breaks through the limitations of traditional single-dimensional evaluation. Through the four-dimensional green development framework and three-dimensional digital technology system, it achieves a comprehensive depiction of the sustainable development status of cities and significantly improves the scientificity and objectivity of the evaluation results.

[0062] (2) This invention combines spatial statistical analysis with spatiotemporal dynamic modeling to intuitively present the regional gradient differences and dynamic evolution patterns of green development levels, providing quantitative basis for identifying “high-value agglomeration areas” and “lagging areas”, and helping regional collaborative governance.

[0063] (3) This invention uses the random forest model to reveal the interaction between digital technology and factors such as economy and infrastructure, providing data support for targeted policy implementation and avoiding policy blindness.

[0064] (4) This invention uses the two-way dynamic analysis and trend prediction of the VAR model to predict the synergistic evolution path of digital technology and green development in advance. The decision support module can output customized solutions for different regions, thereby improving the efficiency of policy implementation.

[0065] (5) This invention enables closed-loop management of the entire process of green development empowered by digital technology, from monitoring, evaluation, analysis to prediction and decision-making, promoting the deep integration of "digitalization and greening", helping cities break through resource and environmental constraints and achieve high-quality and sustainable development.

[0066] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0067] Figure 1 This is an overall flowchart of an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention;

[0068] Figure 2 This is a partial Moran's Index map for 2011, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0069] Figure 3 This is a partial Moran's Index map for 2015, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0070] Figure 4 This is a partial Moran's Index map for 2019, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0071] Figure 5 This is a partial Moran's Index map for 2023, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0072] Figure 6 This is a kernel density estimation map of the Yangtze River Delta urban agglomeration, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0073] Figure 7 This is a two-dimensional kernel density estimation curve of Jiangsu Province, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0074] Figure 8 This is a two-dimensional kernel density estimation curve of Zhejiang Province, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0075] Figure 9 This is a two-dimensional kernel density estimation curve of Anhui Province, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0076] Figure 10 This is a schematic diagram illustrating the feature importance of an embodiment of the comprehensive evaluation method for urban green development level based on digital technology according to the present invention;

[0077] Figure 11This is a partial dependency graph of a single feature in an embodiment of a comprehensive evaluation method for urban green development level based on digital technology according to the present invention, wherein (a) is a partial dependency graph of digital technology, (b) is a partial dependency graph of per capita GDP, (c) is a partial dependency graph of the proportion of tertiary industry added value to GDP, (d) is a partial dependency graph of per capita urban road area, and (e) is a partial dependency graph of annual average population.

[0078] Figure 12 This is a diagram illustrating the interaction effect between digital technology and per capita GDP in an embodiment of the comprehensive evaluation method for urban green development level based on digital technology according to the present invention.

[0079] Figure 13 This is a diagram illustrating the interaction effect between digital technology and the proportion of GDP in the tertiary industry, according to an embodiment of the comprehensive evaluation method for urban green development based on digital technology of the present invention.

[0080] Figure 14 This is a diagram illustrating the interaction effect between digital technology and per capita urban road area in an embodiment of the present invention, which is a comprehensive evaluation method for urban green development level based on digital technology.

[0081] Figure 15 This is a diagram illustrating the interaction effect between the tertiary industry's added value as a percentage of GDP and per capita GDP, according to an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0082] Figure 16 This is a diagram illustrating the interaction effect between per capita GDP and per capita urban road area in an embodiment of a comprehensive evaluation method for urban green development level based on digital technology according to the present invention.

[0083] Figure 17 This is a diagram illustrating the interaction effect between the tertiary industry's added value as a percentage of GDP and per capita urban road area, according to an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0084] Figure 18 This is a feature interaction intensity heatmap of an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention;

[0085] Figure 19 This is a feature root diagram of an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention;

[0086] Figure 20 This is an impulse response curve of digital technology to urban green development, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0087] Figure 21This is an impulse response curve of urban green development to digital technology, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention.

[0088] Figure 22 This is a prediction chart of the digital technology development level and the urban green development level, based on an embodiment of the comprehensive evaluation method for urban green development level based on digital technology of the present invention. Detailed Implementation

[0089] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0090] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0091] like Figure 1 As shown, the present invention provides a comprehensive evaluation method for urban green development level based on digital technology, comprising the following steps:

[0092] S1. Based on the three-system framework of economy-society-environment, this invention introduces the resource dimension at the theoretical level and constructs a four-dimensional collaborative indicator system of "economy-society-environment-resource". This system emphasizes the dynamic correlation and synergistic effect of the four dimensions, and the degree of interaction and coordination determines the quality, efficiency and sustainability of urban green development.

[0093] In terms of indicator system construction, this invention combines the regional characteristics of the Yangtze River Delta urban agglomeration as an important engine of economic development to construct a systematic indicator system, including a comprehensive evaluation indicator system for urban green development level and a comprehensive evaluation indicator system for digital technology development level.

[0094] The comprehensive evaluation index system for green development includes four primary indicators: economic development, social development, environmental development, and resource development, with 11 representative secondary indicators as follows:

[0095] Economic development includes three secondary indicators: per capita GDP, number of registered unemployed urban residents, and the proportion of the tertiary sector's added value to GDP.

[0096] Social development includes three secondary indicators: education expenditure, per capita urban road area, and the number of employees enrolled in basic medical insurance.

[0097] Environmental development includes three secondary indicators: the rate of harmless treatment of domestic waste, the green coverage rate of built-up areas, and the sewage treatment rate.

[0098] Resource development includes two secondary indicators: per capita daily domestic water consumption and the number of industrial enterprises above a designated size.

[0099] The specific evaluation index system is shown in Table 1.

[0100] Table 1 Comprehensive Evaluation Index System for Urban Green Development Level

[0101]

[0102] To evaluate the level of digital technology development, this invention constructs a comprehensive evaluation index system for the level of digital technology development, comprising three primary indicators: digital infrastructure and mobile applications, digital industry investment, and digital technology development. This system covers a total of eight secondary indicators, as follows:

[0103] Digital infrastructure and mobile applications include three secondary indicators: the number of international internet users, the number of mobile phone users at the end of the year, and the annual electricity consumption.

[0104] Digital industry investment includes two secondary indicators: science expenditure and internal R&D expenditure.

[0105] The development of digital technology includes three secondary indicators: the number of patents granted, the number of employees in information transmission, computer services and software industries, and telecommunications revenue.

[0106] The specific evaluation index system is shown in Table 2:

[0107] Table 2. Comprehensive Evaluation Index System for Digital Technology Level

[0108]

[0109] S2. Select city indicator data and preprocess it;

[0110] Based on the "Development Plan for the Yangtze River Delta Urban Agglomeration", this invention selects urban indicator data of 26 cities in the Yangtze River Delta as the research object. The research period is from 2011 to 2023. All data are from the "China Urban Statistical Yearbook", "China Statistical Yearbook" and "China Science and Technology Statistical Yearbook". Some cities have missing values ​​for indicators. The KNNimputer and GM(1,1) models are used to fill in the missing values.

[0111] S3. Construct a comprehensive evaluation model based on the entropy weight-TOPSIS method, where the entropy weight method is used to determine the weight of each indicator, and the TOPSIS method is used to determine the evaluation index.

[0112] The specific steps of the entropy weight-TOPSIS method are as follows:

[0113] (1) The entropy weight method is used to determine the weights of each indicator. The steps are as follows:

[0114] Positive indicators:

[0115]

[0116] Negative indicators:

[0117]

[0118] Where i represents a city, i = 1, 2, ..., m, j represents a measurement indicator, j = 1, 2, ..., n; x ij X represents the raw data of the j-th indicator in the i-th city. If the indicator data is a positive indicator, it is calculated using formula (1); if the indicator data is a negative indicator, it is calculated using formula (2). ij x represents the standardized data of the j-th indicator for the i-th city; min(j) x max(j) These represent the minimum and maximum values ​​of the j-th indicator, respectively.

[0119] Calculate the specific gravity p ij :

[0120]

[0121] Calculate the information entropy e of the j-th measurement index. j :

[0122]

[0123] Finally, calculate the entropy weight w j for:

[0124]

[0125] (2) The TOPSIS method is used to determine the evaluation index, specifically:

[0126] Construct the weighted matrix T ij :

[0127]

[0128] Determine the positive and negative ideal solutions:

[0129]

[0130] Calculate the Euclidean distance between the index and the positive and negative ideal solutions:

[0131]

[0132] Calculate the relative closeness C of the optimal solution. i :

[0133]

[0134] C i The higher the value, the higher the city's overall evaluation score.

[0135] S4. The green development level index and digital technology development level index of cities in the Yangtze River Delta were analyzed using a comprehensive evaluation model. The results are shown in Tables 3 and 4.

[0136] Table 3 Urban Green Development Level Index of the Yangtze River Delta Urban Agglomeration

[0137] years 2011 2014 2016 2021 2022 2023 Shanghai 0.498 0.534 0.556 0.587 0.591 0.626 Suzhou 0.514 0.529 0.541 0.566 0.572 0.583 Wuxi City 0.439 0.440 0.451 0.461 0.468 0.493 Nanjing 0.425 0.407 0.415 0.454 0.457 0.453 Changzhou 0.364 0.383 0.407 0.425 0.413 0.412 Nantong City 0.444 0.451 0.454 0.438 0.414 0.425 Yangzhou 0.336 0.339 0.343 0.360 0.361 0.370 Zhenjiang City 0.387 0.379 0.386 0.401 0.390 0.379 Taizhou City 0.403 0.365 0.384 0.417 0.411 0.411 Yancheng City 0.363 0.325 0.348 0.386 0.380 0.377 Hangzhou 0.435 0.389 0.382 0.423 0.432 0.461 Ningbo City 0.406 0.359 0.340 0.395 0.395 0.420 Jiaxing City 0.355 0.313 0.306 0.399 0.401 0.390 Huzhou City 0.384 0.363 0.374 0.392 0.391 0.450 Shaoxing City 0.374 0.326 0.314 0.354 0.414 0.441 Taizhou City 0.281 0.368 0.388 0.405 0.406 0.419 Zhoushan City 0.330 0.313 0.311 0.331 0.338 0.339 Jinhua City 0.353 0.380 0.386 0.405 0.414 0.430 Hefei City 0.269 0.283 0.288 0.312 0.319 0.344 Wuhu City 0.366 0.349 0.353 0.359 0.357 0.349 Ma'anshan City 0.286 0.255 0.268 0.335 0.346 0.341 Chuzhou City 0.391 0.468 0.461 0.459 0.451 0.444 Xuancheng City 0.348 0.352 0.379 0.411 0.415 0.417 Tongling City 0.260 0.221 0.197 0.308 0.315 0.310 Chizhou City 0.326 0.332 0.334 0.343 0.345 0.359 Anqing City 0.258 0.254 0.271 0.418 0.413 0.413

[0138] Analysis of the spatiotemporal evolution characteristics in Table 3 reveals that the green development level of the 26 cities in the Yangtze River Delta region showed a continuous upward trend from 2011 to 2023. Among them, Shanghai and Suzhou had significantly higher comprehensive green development indices than the overall average, reflecting the particularly outstanding achievements of these two cities in coordinated development across dimensions such as economic transformation, social progress, environmental protection, and resource utilization.

[0139] Table 4. Digital Technology Development Level Index of the Yangtze River Delta Urban Agglomeration

[0140]

[0141]

[0142] Analysis of the spatiotemporal evolution trends in Table 4 reveals a continuous upward trend in the development level of digital technology in the Yangtze River Delta urban agglomeration. Notably, the digital technology development levels of core cities such as Shanghai, Hangzhou, Hefei, and Nanjing are significantly higher than the overall average level of the Yangtze River Delta urban agglomeration. This spatial differentiation is closely related to key factors such as the geographical advantages, population agglomeration effect, resource allocation capacity, and economic development level of each city.

[0143] S5. Using spatial statistical analysis, random forest model and VAR model, we analyze the spatiotemporal effects and interaction mechanism between digital technology development and urban green development.

[0144] Spatial statistical analysis methods include Moran's index, kernel density estimation methods, and spatial Markov chain models;

[0145] The Moran index is an important indicator for detecting spatial autocorrelation. By comparing the similarity of observations in adjacent regions, it determines whether there is a significant spatial dependence between research variables. Based on the difference in research scale, the Moran index can be divided into two categories: the global Moran index and the local Moran index. The former reflects the overall spatial correlation pattern, while the latter further identifies local spatial heterogeneity patterns.

[0146] The green development levels of various urban agglomerations are classified into four types: high-high (HH) clustering, low-low (LL) clustering, high-low (HL) outlier, and low-high (LH) outlier. Global Moran Index I and local Moran Index I... i The formula is as follows:

[0147]

[0148] in, Let m be the sample variance, m be the total number of cities, and t be the total number of cities. i Let t represent the green development level of the i-th city. k W represents the green development level of the k-th city. ik The spatial weight matrix is ​​used in this invention; the spatial adjacency weight matrix is ​​used instead.

[0149] This invention employs a nonparametric kernel density estimation method based on Gaussian functions to empirically analyze the dynamic evolution trend of green development level in the Yangtze River Delta urban agglomeration. It assumes there are m observations (i.e., the total number of cities), and t... i The probability density function of the random variable t, representing the green development level of each city, is estimated as follows:

[0150]

[0151] in, Let represent the average value, and h be the bandwidth parameter, which acts as a smoothing coefficient to control the smoothness of the density curve. This invention determines the optimal bandwidth using the Silverman criterion and employs the Gaussian kernel function for estimation to analyze the evolution trend of the green development level of the Yangtze River Delta urban agglomeration. The specific expression is as follows:

[0152]

[0153] This invention uses a spatial Markov chain model to analyze the state transition probability and spatial spillover effect of urban green development level.

[0154] A traditional Markov chain is a stochastic process with Markov properties, used to dynamically adjust the green development levels of different cities in different years, so that the state transition probability of the system at a certain moment depends only on the current state.

[0155] This invention reveals the dynamic evolution trend of different urban development levels by constructing an M×M transition matrix, as shown in the following formula:

[0156]

[0157] Among them, P ds Let n be the probability of transitioning from state d at time r to state s at time r+1. ds Let n be the number of cities that transition from state type d to state type s during the period from r to r+1. d This represents the number of cities in state d throughout the entire research period.

[0158] This invention's spatial Markov chain, based on the traditional Markov chain, combines spatial spillover effects to observe the trend evolution of spatial spillover effects on urban green development levels. The M×M transition probability matrix is ​​decomposed into an M×M×M transition probability matrix using a spatial weight matrix, as shown in the following formula:

[0159]

[0160] Where y represents the spatial lag factor, t i The observation value of spatial unit i, w ik The elements in the spatial weight matrix W represent the weights of the spatial lag factor; here, the spatial adjacency weight matrix is ​​still used. This formula shows that the spatial lag value of each spatial unit is a weighted sum of its neighboring units, with the weights determined by spatial relationships.

[0161] First, spatial autocorrelation analysis:

[0162] (1) Global Moran index;

[0163] This invention employs the global Moran's index and utilizes a spatial adjacency matrix to analyze the spatial relationships among cities' green development levels. Table 5 shows that the global Moran's index for green development levels in the Yangtze River Delta urban agglomeration is positive for all cities except 2014, and all passed the 5% significance test, indicating a significant spatial positive correlation and clustering effect. Specifically, cities with similar green development levels exhibit a clustering phenomenon of high and low levels. The presence of localized low-to-low clustering areas in 2014 indicates the existence of green development "lowlands," necessitating cross-regional governance mechanisms to break spatial barriers and promote balanced and coordinated development across the entire region.

[0164] Table 5. Moran Index of Urban Green Development Level

[0165] years MoranI Expect E(I) Standard deviation sd(I) Z-value p-value 2011 0.286 -0.04 0.133 2.445 0.014 2012 0.263 -0.04 0.133 2.27 0.023 2013 0.235 -0.04 0.133 2.062 0.039 2014 0.192 -0.04 0.133 1.737 0.082 2015 0.234 -0.04 0.133 2.056 0.04 2016 0.23 -0.04 0.133 2.028 0.043 2017 0.32 -0.04 0.133 2.703 0.007 2018 0.38 -0.04 0.133 3.15 0.002 2019 0.426 -0.04 0.133 3.495 0.000 2020 0.363 -0.04 0.133 3.023 0.003 2021 0.275 -0.04 0.133 2.362 0.018 2022 0.309 -0.04 0.133 2.615 0.009 2023 0.338 -0.04 0.133 2.838 0.005

[0166] (2) Local Moran index;

[0167] This invention further identifies specific clustering patterns of green development through local Moran's index maps. For example... Figures 2-5 As shown, the green development of the Yangtze River Delta urban agglomeration exhibits significant spatial heterogeneity: the Moran Index peaked at 0.426 in 2019, indicating the most pronounced high-value and low-value clustering characteristics in the region that year; while the Moran Index declined somewhat in 2023, it remained at a relatively high level, indicating that spatial clustering within the region remained strong; cities such as Shanghai, Suzhou, and Wuxi have long been located in the first quadrant, i.e., the high-high (HH) clustering zone, demonstrating stable high-value clustering characteristics and radiating influence on surrounding cities; whereas cities such as Anqing and Tongling have consistently been located in the third quadrant, reflecting the persistence of low-value clustering. Overall, there is a significant imbalance in regional development, and balanced development can be further promoted through policy regulation and regional cooperation in the future.

[0168] Second, analysis of the dynamic evolution of urban green development level based on kernel density function:

[0169] To analyze the spatial differentiation and dynamic evolution of green development levels in the Yangtze River Delta urban agglomeration, this invention employs a nonparametric kernel density estimation method based on Gaussian kernel functions to plot kernel density curves for the overall region and for the three provinces of Jiangsu, Zhejiang, and Anhui, as shown below. Figures 6-9 As shown.

[0170] from Figure 6 It can be observed that the overall level of green development in cities within the Yangtze River Delta region shows a rightward shift, indicating that the level of green development in various regions is gradually improving. Analyzing the distribution characteristics and morphology, the height of the main peak is rising amidst fluctuations, suggesting a decreasing trend in the breadth of urban-rural integration and a narrowing of spatial disparities. Meanwhile... Figure 6 The absence of a significant trailing phenomenon indicates that the level of urban green development is relatively concentrated across different years.

[0171] analyze Figures 7-9 The distribution characteristics of the kernel density curves reveal that the overall green development level of cities in the Yangtze River Delta region exhibits a single-peak clustering trend. However, in the early stages, there were differences within the three provinces: Jiangsu and Anhui provinces showed weak secondary peaks on the right side of their kernel density curves, while Zhejiang province had a more pronounced auxiliary peak on the left side. This phenomenon indicates that the green development level of cities in the three provinces initially had a multi-center distribution characteristic. As time progressed, the secondary peaks gradually weakened, reflecting the evolution of the regional development model from a multi-center dispersion to a single-center concentration. Some cities within the three provinces formed significant development advantages through policy support or resource agglomeration.

[0172] Third, analyzing the evolution of urban green development levels based on spatial Markov chain models:

[0173] This invention uses the Markov chain model to further analyze the dynamic evolution characteristics of urban green development level. Based on the distribution characteristics of sample data, the quartile method is used to divide the urban green development level into four levels: low level (Class I), low-to-medium level (Class II), medium-to-high level (Class III), and high level (Class IV). The corresponding threshold parameters are 0.339, 0.383, and 0.415, respectively. The results of the traditional Markov chain transition probability matrix are shown in Table 6.

[0174] Table 6 Transition Probability Matrix of Traditional Markov Chain

[0175] t / t+1 Ⅰ Ⅱ Ⅲ Ⅳ Ⅰ 0.831 0.157 0.012 0 Ⅱ 0.091 0.740 0.169 0 Ⅲ 0 0.088 0.738 0.175 Ⅳ 0 0 0.097 0.903

[0176] Table 6 shows that the probabilities of each type of city maintaining its current level of green development are 83.1%, 74.0%, 73.8%, and 90.3%, respectively. The probabilities on the diagonal lines are all higher than those off-diagonal, indicating that the green development levels of various cities in the Yangtze River Delta region are stable. Specifically, cities with high-level development are more likely to maintain their current state than those with low-level development, suggesting a trend of "group convergence" in the green development levels of cities in the Yangtze River Delta region, with cities of lower green development levels having a relatively greater chance of improving to higher levels.

[0177] Furthermore, during the observation period, the level of green development in cities in the Yangtze River Delta region showed a progressive trend. Specifically, the probabilities of cities of type I, II, and III moving up to the next higher level were 15.7%, 16.9%, and 17.5%, respectively, all higher than their probabilities of moving down, indicating that cities with lower levels of green development are more likely to improve to higher levels.

[0178] However, traditional Markov transition probability matrices only calculate probabilities based on their own state evolution, ignoring the influence of spatial geographic factors on system state changes. To overcome this limitation, this invention introduces spatial lag variables to extend the traditional model, ultimately constructing a spatial Markov transition probability matrix, as shown in Table 7.

[0179] Table 7 Spatial Markov Chain Transition Probability Matrix

[0180]

[0181] As shown in Table 7, on the one hand, under different spatial lag states, the off-diagonal elements of the spatial Markov transition matrix all have non-zero values, indicating that the level of green development between cities has both the potential for upward transfer and the risk of downward transfer. From the perspective of the transfer path, the level leap of urban green development mostly occurs between adjacent types.

[0182] On the other hand, optimizing the green development level of geographically adjacent areas has a positive spatial spillover effect on the region's green transformation. Regions with lower green development levels are more likely to migrate upwards when adjacent to regions with higher development levels. Data shows that for Type I, the probability of upward migration increases with increasing adjacency type: 8.1%, 14.3%, 33.3%, and 40.0%, respectively. This indicates that regions with medium and high development levels have a positive spatial spillover effect on the dynamic convergence process of development levels in their adjacent regions.

[0183] Analyzing the factors influencing green development is an important research topic for promoting the coordinated development of the economy and the environment. The rapid development of digital technology, the transformation of economic growth models, the optimization of industrial structures, and the advancement of urbanization can all have a profound impact on green development. Random forest models, as powerful machine learning algorithms, can effectively handle complex relationships among multiple variables and have unique advantages in analyzing the factors influencing green development.

[0184] This invention employs an ensemble learning method to construct a random forest model containing 500 decision trees. The main parameter settings are as follows: the minimum number of leaf node samples is set to 5 to prevent overfitting of the decision trees; out-of-bag prediction and out-of-bag prediction variable importance assessment are enabled, and out-of-bag data is used to evaluate the model and perform feature importance analysis.

[0185] First, a partial dependency plot is used to illustrate the marginal effect of a single or multiple independent variables on the dependent variable. The calculation process for a partial dependency plot of a single independent variable is as follows:

[0186] By fixing the values ​​of other independent variables, traversing the value space of the target independent variable, and calculating the average predicted value of the model for each value, the relationship curve between the average predicted value and the value of the target independent variable is the partial dependency plot. The partial dependency plot can intuitively present the trend of the influence of the independent variable on the dependent variable, and help to analyze the mechanism of each factor in the model.

[0187] Feature interaction analysis aims to explore the impact of interactions between different independent variables on the dependent variable. By calculating the partial dependencies of different feature combinations and comparing them with the partial dependencies of individual features, the strength of interactions between features can be assessed.

[0188] Then, this invention uses Friedman's statistic to measure the strength of feature interactions, the calculation formula of which is:

[0189]

[0190] Among them, PDP i'j' This represents the partial dependency value of the combination of features i′ and j′, while PDP i' and PDP j' H represents the individual partial dependency values ​​of features i′ and j′, respectively. The larger the H value, the stronger the interaction between features i′ and j′.

[0191] This invention selects digital technology level, per capita GDP, the proportion of tertiary industry value added to GDP, per capita urban road area, and average annual population as key variables to systematically analyze their impact mechanism on the green development level of urban agglomerations. Through quantitative research, it reveals the intensity and interaction of each factor, providing a scientific basis for promoting sustainable urban development.

[0192] The importance scores of each variable were calculated using out-of-bag data from a random forest model, and the results are as follows: Figure 10 As shown. From Figure 10 As can be seen from the data, among the many factors influencing green development, digital technology scores relatively high in importance, indicating that digital technology has a significant impact on green development. The importance of per capita urban road area and per capita GDP is next, while the proportion of tertiary industry value added to GDP and average annual population are relatively less important, but still have a certain impact on green development.

[0193] (1) Draw a partial dependency graph for a single feature, the result of which is... Figure 11 As shown, Figure 11 As shown in (a), with the improvement of digital technology, the level of green development first rises and then tends to stabilize. This indicates that digital technology can promote green development to a certain extent, but its marginal contribution to green development may gradually decrease after digital technology develops to a certain stage.

[0194] like Figure 11 As shown in (b), there is a positive correlation between per capita GDP and green development; that is, the higher the per capita GDP, the higher the level of green development. This indicates that economic development provides a solid material foundation and technological guarantee for green development.

[0195] like Figure 11 As can be seen from (c), there is a relatively complex relationship between the proportion of the tertiary industry's added value to GDP and green development. When the proportion of the tertiary industry's added value to GDP is low, the level of green development rises with economic growth. However, when the proportion of the tertiary industry's added value to GDP exceeds a certain threshold, the level of green development may fluctuate or even decline.

[0196] like Figure 11 As shown in (d) of the graph, the partial dependency plot of per capita urban road area indicates that, within a certain range, an increase in per capita urban road area has a positive impact on green development, but this impact weakens after exceeding a certain value. This may be because moderate road construction can improve urban traffic efficiency and reduce energy consumption and environmental pollution, while excessive construction may lead to problems such as resource waste.

[0197] like Figure 11 As can be seen from (e), the relationship between average annual population and green development is relatively weak, but overall, the increase in population size may put some pressure on green development.

[0198] (2) Draw a bivariate partial dependency graph, the result of which is as follows: Figure 12-17As shown, there are significant interaction effects among the variables. Specifically, digital technology and per capita GDP exhibit a synergistic enhancement characteristic; once per capita GDP reaches a certain threshold, the promoting effect of digital technology on green development significantly increases. Digital technology and the proportion of tertiary industry value added in GDP show a mutually reinforcing relationship, jointly promoting the improvement of green development levels. In terms of infrastructure, the synergistic effect of digital technology and per capita urban road area effectively optimizes the transportation system and reduces pollution emissions. The interaction analysis among economic indicators shows that the growth of per capita GDP drives the increase of the proportion of the tertiary industry, forming a virtuous cycle; at the same time, per capita GDP and road construction mutually promote each other, and the improvement of infrastructure further supports economic development. In addition, there is an optimal synergistic range between per capita urban road area and the proportion of tertiary industry value added in GDP; appropriate matching can maximize the benefits of green development.

[0199] (3) Draw the interaction intensity matrix diagram, the result is as follows: Figure 18 As shown in the image, observing the feature interaction intensity matrix heatmap reveals significant differences in the interaction intensity between different features. Digital technology exhibits a strong correlation, showing high interaction intensity with per capita GDP, the proportion of tertiary industry value added in GDP, and per capita urban road area. In contrast, the interaction intensity between average annual population and other features is generally lower.

[0200] From the perspective of economic development, the interaction between per capita GDP and the proportion of tertiary industry value added in GDP is relatively strong, reflecting a close intrinsic link between regional economic development level and industrial structure. In the process of economic development, the increase in per capita GDP is often accompanied by adjustments in industrial structure, and changes in the proportion of the tertiary industry will, in turn, affect the quality and speed of regional economic development. The two may have a dynamic relationship of coordinated development or mutual constraint. Furthermore, there is also a certain degree of interaction between per capita urban road area and characteristics such as per capita GDP and the proportion of tertiary industry value added in GDP. This fully demonstrates that urban road construction is not an isolated phenomenon, but rather intertwined and mutually influential with the evolution of regional economic development level and industrial structure.

[0201] Random forest models have identified digital technology as a significant influencing factor on the level of urban green development. Therefore, this invention employs a VAR model to analyze the dynamic interaction between digital technology and green development levels in urban agglomerations, revealing the influence paths and time-lag effects through impulse response analysis, and predicting trends over the next six years.

[0202] The VAR model is an econometric model used to analyze the dynamic relationships of multivariate time series. Its core lies in characterizing the mutual influences and time lag effects among variables. This model can capture the complex interaction mechanisms between variables and quantify the lasting impact of shocks through impulse response functions. This invention utilizes this model to reveal the dynamic relationship between digital technology and green development, predict their synergistic evolution trend, and assess the contribution of digital technology through variance decomposition, providing a time-series basis for differentiated policy formulation.

[0203] In this embodiment, a VAR model is used to analyze the dynamic interaction between digital technology and urban green development levels in the Yangtze River Delta urban agglomeration from 2011 to 2023, and to predict the relationship between digital technology and urban green development levels. The steps are as follows:

[0204] First, the ADF unit root test was performed on the green development level index and digital technology development level index of the city obtained from S4 from 2011 to 2023 to test the stationarity of the original time series. The root test results are shown in Tables 8 and 9.

[0205] Table 8 Digital Technology Level Score - ADF Test Table

[0206]

[0207] According to Table 8, the t-statistic of the original time series is greater than the value at the 10% confidence level, and the p-value is greater than 0.05, indicating that it is a non-stationary series. At the same time, the original time series is subjected to first-order differencing or logarithmic differencing. The t-statistic of the first-order differencing is less than the value at the 1% confidence level, and the p-value is much less than 0.05. Therefore, the first-order differencing series is a stationary series. Let d = 1, and the subsequent analysis will be based on the first-order differencing series.

[0208] Table 9. Urban Green Development Level Score - ADF Test Table

[0209]

[0210] According to Table 9, the t-statistic of the original urban green development level series is greater than the value at the 10% confidence level, and the p-value is greater than 0.05. Therefore, the series is non-stationary. Then, the series is first-differenced. Finally, the t-statistic of the first-difference is less than the value at the 1% confidence level, and the p-value is much less than 0.05. Therefore, the first-difference series is stationary. Subsequent analysis will be carried out based on the first-difference series.

[0211] Therefore, it can be seen that the original variable sequence is non-stationary at the 5% level, while the variable sequence after first difference is stationary at the 1% level, which satisfies the premise that each variable is stationary as required by the VAR model.

[0212] If both variables are integrated series of the same order, based on the unit root test, the Johansen test is further used to test whether there is a cointegration relationship between the integrated time series of digital technology and urban green development, in order to determine whether there is a long-term equilibrium relationship between the green development level index and the digital technology development level index. The results are shown in Table 10.

[0213] Table 10 Johansen Cointegration Test

[0214] Null hypothesis Characteristic roots trace 10% critical value 5% critical value 1% critical value None (No cointegration relationship) 0.954 30.574 13.429 15.494 19.935 At most 1 cointegration 0.280 2.652 2.705 3.841 6.635

[0215] The test results in Table 10 show that there is one cointegration relationship, indicating that a VAR model can be constructed.

[0216] Next, the VAR model is ordered. Generally, the smaller the lag order, the higher the available sample information and the simpler the model. This invention is based on the AIC, BIC and HQIC indices to determine the VAR model order, establish the VAR model, and judge the stability of the VAR model through the eigenvalue plot. As shown in Table 11, when the order is 3, the values ​​of the three indices all reach the minimum. Therefore, the lag order of the VAR model is set to 3.

[0217] Table 11 Order Determination Table

[0218] order AIC BIC HQIC 0 -18.154 -18.094 -18.221 1 -22.876 -22.694 -23.075 2 -22.471 -22.168 -22.802 3 -23.811 -23.388 -24.276

[0219] After constructing the VAR model, its stability is determined by analyzing the feature map. If all eigenvalues ​​are within the unit circle, the model is considered stable. If eigenvalues ​​are outside the unit circle, the model may not possess long-term stability. Figure 19 As can be seen, all the eigenvalues ​​are within the unit circle, which means that the constructed VAR model has long-term stability.

[0220] like Figure 20 , Figure 21 As shown, this invention uses impulse response diagrams to reflect the mutual impact of digital technology and urban green development. (Observation) Figure 20 The figure illustrates the dynamic impact of a one-unit positive shock on digital technology on itself and the level of urban green development. It shows that the self-response of digital technology peaks at 0.004 in the initial stage of the shock, then gradually declines, approaching zero in period 11. This indicates that policy or technological improvements in digital technology can significantly enhance its own level in the short term, but the effect weakens over time, requiring continuous investment to maintain long-term effects. Meanwhile, the shock response to the level of urban green development is generally positive, initially rising to a peak in period two before declining and stabilizing, indicating that digital technology has a promoting effect on urban green development, but its impact weakens over time.

[0221] Figure 21 This reflects the impact of shocks to the level of urban green development on both its own capabilities and its digital technology level. The study found that the response value of urban green development to its own capabilities gradually decreased, but remained generally positive. However, the response to digital technology exhibited a "lagging but continuous" characteristic, starting with a negative value, gradually rising to a peak of 0.001 in the second period, and then slowly declining. This indicates that there is a 1-2 year time lag between urban green development and digital technology, and once a synergistic effect is formed, its impact has strong sustainability.

[0222] Figure 20 and Figure 21 This study reveals an asymmetric interaction between digital technology and green development. The results show that digital technology has a significant and lasting impact on urban green development, while the self-reinforcing effect of urban green development is significant but its feedback effect on digital technology is relatively small. This finding verifies the long-term nature of "technology-driven green transformation," indicating that the strategic value of technological innovation may be underestimated due to time lag effects, and that a balance needs to be struck between short-term benefits and long-term technological planning in practice.

[0223] Next, variance decomposition was performed on the first-order difference variables to quantify the contribution of each variable to the system fluctuations. For the level of digital technology, as shown in Table 12, its fluctuations mainly stem from its own historical shocks, consistently accounting for over 93%, while urban green development has a weak explanatory power, indicating that digital technology development has strong endogenous inertia and is limited by short-term influences from external environmental factors. Conversely, as shown in Table 13, the fluctuations in the level of urban green development are more driven by digital technology, with its contribution rate rising from 30.119% in period 1 to 40.757% and remaining stable, while its own contribution rate decreased from 69.881% to 59.243%, highlighting the core supporting role of digital technology in green transformation.

[0224] Table 12 Diff1_Variance Decomposition Table of Digital Technology Level

[0225]

[0226]

[0227] Table 13 Diff1_Variance Decomposition Table of Urban Green Development

[0228] order Standard error Diff1 - Digital Technology Level (%) Diff1_Urban Green Development Level (%) 1 0.004 30.119 69.881 2 0.006 33.184 66.816 3 0.006 33.306 66.694 4 0.006 33.797 66.203 5 0.006 35.633 64.367 6 0.006 37.772 62.228 7 0.006 37.739 62.261 8 0.006 38.758 61.242 9 0.006 40.751 60.249 10 0.006 40.757 59.243

[0229] Finally, a trend forecast for the level of digital technology and the level of urban green development over the next six years was conducted, and the results are as follows: Figure 22According to forecasts and analysis, over the next six years, the development levels of urban digital technology and urban green development in the Yangtze River Delta urban agglomeration will maintain a coordinated upward trend. From an evolutionary perspective, the region has completed its initial accumulation and development phase, and currently, digital technology and urban green development exhibit a positive interaction. The forecast results indicate that the Yangtze River Delta urban agglomeration should focus on promoting the transformation of its industrial structure towards green and low-carbon directions, and continuously optimize the region's high-quality development pattern. This development trend will provide a useful reference for other urban agglomerations.

[0230] Therefore, this invention adopts the aforementioned comprehensive evaluation method for urban green development level based on digital technology. At the theoretical level, by constructing a two-dimensional evaluation system of digital technology and urban green development, it reveals the inherent mechanism of urban green development from the perspective of digital empowerment. At the methodological level, it integrates spatial statistics and machine learning algorithms to achieve high-precision simulation and attribution analysis of spatiotemporal evolution. At the practical level, based on a dynamic prediction model, it calculates the long-term co-evolution trend of future urban green development level and digital technology, providing quantitative decision support for regional differentiated policy design.

[0231] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A comprehensive evaluation method for urban green development level based on digital technology, characterized in that, Includes the following steps: S1. Construct an indicator system, including a comprehensive evaluation indicator system for the level of urban green development and a comprehensive evaluation indicator system for the level of digital technology development; S2. Select city indicator data and preprocess it; S3. Construct a comprehensive evaluation model based on the entropy weight-TOPSIS method, where the entropy weight method is used to determine the weight of each indicator, and the TOPSIS method is used to determine the evaluation index. S4. Use a comprehensive evaluation model to analyze the green development level index and digital technology development level index of each city; S5. Using spatial statistical analysis, random forest model and VAR model, we analyze the spatiotemporal effects and interaction mechanism between digital technology development and urban green development.

2. The comprehensive evaluation method for urban green development level based on digital technology according to claim 1, characterized in that, In S1, a comprehensive evaluation index system for the level of green development is constructed, including four primary indicators: economic development, social development, environmental development, and resource development. Economic development includes three secondary indicators: per capita GDP, number of registered unemployed urban residents, and the proportion of the tertiary sector's added value to GDP. Social development includes three secondary indicators: education expenditure, per capita urban road area, and the number of employees enrolled in basic medical insurance. Environmental development includes three secondary indicators: the rate of harmless treatment of domestic waste, the green coverage rate of built-up areas, and the sewage treatment rate. Resource development includes two secondary indicators: per capita daily domestic water consumption and the number of industrial enterprises above a designated size.

3. The comprehensive evaluation method for urban green development level based on digital technology according to claim 1, characterized in that, In S1, a comprehensive evaluation index system for the development level of digital technology is constructed, including three primary indicators: digital infrastructure and mobile applications, digital industry investment, and digital technology development. Digital infrastructure and mobile applications include three secondary indicators: the number of international internet users, the number of mobile phone users at the end of the year, and the annual electricity consumption. Digital industry investment includes two secondary indicators: science expenditure and internal R&D expenditure. The development of digital technology includes three secondary indicators: the number of patents granted, the number of employees in information transmission, computer services and software industries, and telecommunications revenue.

4. The comprehensive evaluation method for urban green development level based on digital technology according to claim 1, characterized in that, In S3, the entropy weight method determines the weights of each indicator, and the steps are as follows: Positive indicators: Negative indicators: Where i represents a city, i = 1, 2, ..., m, j represents a measurement indicator, j = 1, 2, ..., n; x ij X represents the raw data for the j-th indicator in the i-th city. ij x represents the standardized data of the j-th indicator for the i-th city; min(j) x max(j) Let represent the minimum and maximum values ​​of the j-th indicator, respectively. Calculate the specific gravity p ij : Calculate the information entropy e of the j-th measurement index. j : Finally, calculate the entropy weight w j for: The TOPSIS method determines the evaluation index as follows: Construct the weighted matrix T ij : Determine the positive and negative ideal solutions: Calculate the Euclidean distance between the index and the positive and negative ideal solutions: Calculate the relative closeness C of the optimal solution. i : C i The higher the value, the higher the city's overall evaluation score.

5. The comprehensive evaluation method for urban green development level based on digital technology according to claim 1, characterized in that, In S5, spatial statistical analysis methods include: The global Moran index is used to analyze the spatial relationships of urban green development levels using a spatial adjacency matrix, and the local Moran index map is used to identify specific clustering patterns of green development. The dynamic evolution of urban green development level is analyzed using a nonparametric kernel density estimation method based on Gaussian function. A spatial Markov chain model is used to analyze the state transition probability and spatial spillover effect of urban green development level.

6. The comprehensive evaluation method for urban green development level based on digital technology according to claim 1, characterized in that, In S5, an ensemble learning method is used to construct a random forest model containing 500 decision trees, the minimum number of leaf node samples is set to 5, and out-of-bag prediction and out-of-bag prediction variable importance evaluation are enabled. Partial dependency plots illustrate the marginal effect of a single or multiple independent variables on the dependent variable. The calculation process for a partial dependency plot of a single independent variable is as follows: By fixing the values ​​of other independent variables, traversing the value space of the target independent variable, and calculating the average predicted value of the model for each value, the relationship curve between the average predicted value and the value of the target independent variable is the partial dependency graph. Friedman's statistic is used to measure the feature interaction strength H, and its calculation formula is as follows: Among them, PDP i'j' This represents the partial dependency value of the combination of features i′ and j′, while PDP i' and PDP j' H represents the partial dependency values ​​of features i′ and j′ individually, respectively. The larger the H value, the stronger the interaction between features i′ and j′. The study selected digital technology level, per capita GDP, the proportion of tertiary industry added value in GDP, per capita urban road area, and average annual population as variables affecting the green development level of urban agglomerations for analysis.

7. The comprehensive evaluation method for urban green development level based on digital technology according to claim 1, characterized in that, In S5, the VAR model is used to analyze the dynamic interaction between digital technology and the level of green development in urban agglomerations. The steps are as follows: First, the green development level index and digital technology development level index obtained from S4 are subjected to the ADF unit root test to test the stationarity of the original time series. If the original time series is a non-stationary series, first-order differencing or logarithmic differencing is used until the stationarity test is passed. If both variables are integrated series of the same order, the Johansen cointegration test is further used to determine whether there is a long-term equilibrium relationship between the green development level index and the digital technology development level index. Next, the order of the VAR model is determined based on the AIC, BIC and HQIC indices, the VAR model is established, and the stability of the VAR model is judged by the eigenvalue plot. Finally, impulse response analysis is used to demonstrate the specific paths and time-delay effects of the impact, while variance decomposition is used to quantify the contribution of each variable to the system fluctuations, and the trend of digital technology and urban green development level is predicted.

Images