Building environment index measurement method for remarkably improving activity amount of residents in park

Abstract

A built-up environment index measuring method for remarkably improving activity amount of residents in a park comprises the steps of sampling annual built-up environment indexes and natural environment indexes xi in an L range around a sample park, collecting resident park activity amount data y in the same period, adopting spss software, applying a linear regression model, taking y as a dependent variable and xi as an independent variable, and calculating the activity amount of the residents in the park. Analyzing an independent variable xi having significant positive correlation with the dependent variable y to obtain a built environment index xi capable of significantly improving the activity amount y of the resident park; and writing R language codes according to a machine learning gradient boosting decision tree GBDT model to obtain a threshold interval of a built environment index xi capable of remarkably improving the activity amount y of the resident park. Aiming at the influence of the built environment around the park on the activity amount of residents, a set of built environment index measurement analysis method for remarkably improving the activity amount of the residents in the park is formed, an influence curve of the built environment around the park on the activity amount of the residents is formed, and an optimization scheme of the optimal built environment around the park is obtained.

Description

Technical Field

[0001] The present invention relates to the field of urban planning, and in particular to a built environment index measurement method that significantly improves the amount of residents' activities in parks. Background Art

[0002] Urban parks are the first choice for residents to exercise. Differences in the built environment around them will directly lead to changes in the amount of activity of residents who come to the park, such as the average building height, building density, public transportation station density, sanitation facility coverage, and greening rate near the park. At the same time, in addition to built environment factors, natural environmental factors such as temperature, precipitation, and slope will also affect the amount of activity of residents in the park. After comprehensively considering the above-mentioned parks and surrounding built environments and other control factors, the analysis was carried out using park card swiping data and remote sensing data, and finally a built environment indicator measurement and layout optimization method that can significantly improve the amount of activity of residents in the park was achieved.

[0003] By searching the keywords "built environment" and "activity" on the invention patent platform of the China Patent Office and sorting out and comparing invention patents related to built environment optimization, we found that the existing patents have the following shortcomings:

[0004] (1) There are methods to measure the built environment, but they do not take into account the nonlinear impact of the built environment on residents’ activity levels.

[0005] (2) No consideration was given to exploring built environment indicators that could promote residents’ activity levels.

[0006] (3) The influence of control variables such as weather, topography and other natural environmental factors is not considered, which makes the research on built environment indicators unrealistic.

[0007] (4) Only the linear method is used to analyze the impact coefficient of the built environment, which makes it impossible to quantitatively analyze and evaluate the impact of the built environment and propose targeted optimization strategies. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a built environment index measurement method that significantly improves the amount of residents' activities in the park. This method aims at the impact of the built environment around the park on the amount of residents' activities, and programs a set of built environment index measurement and analysis methods that significantly improve the amount of residents' activities in the park.

[0009] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0010] A built environment indicator measurement method that significantly improves residents' activities in parks includes the following steps:

[0011] Step 1) Annual built environment index and natural environment index x within the range L around the sample park i Sampling is performed, where i is a natural number; and the data y of residents' park activity volume in the same period is collected, where the residents' park activity volume is the product of the residents' park activity duration and activity frequency;

[0012] Step 2) Use SPSS software to apply a linear regression model, with y as the dependent variable and x as the i As the independent variable, analyze the independent variable x that is significantly positively correlated with the dependent variable y. i , we can obtain the built environment index x that can significantly increase the amount of residents’ park activities y i ;

[0013] Step 3) Write R code based on the machine learning gradient boosting decision tree GBDT model, taking y as the dependent variable and the built environment indicator x obtained in step 2) that can significantly increase the amount of residents' park activities y i is the independent variable, according to the built environment index x i The nonlinear influence curve of residents’ park activity volume y is used to obtain the built environment index x that can significantly increase residents’ park activity volume y. i The threshold interval.

[0014] Furthermore, the built environment indicators include floor area ratio, green space ratio, distance to rail stations, bus station density, service facility density, average house price, building density and average building height; the natural environment indicators include temperature, humidity, precipitation, wind speed, air pressure, sunshine duration, terrain elevation and terrain slope.

[0015] Furthermore, L is 500 meters.

[0016] Furthermore, in the step 1), the buffer tool under the ArcGIS 10.7 toolbox is used to sample the built environment indicators and natural environment indicators in the L range around the sample park.

[0017] The beneficial effect of the present invention is that: in view of the impact of the built environment around the park on the amount of residents' activities, a set of built environment index measurement and analysis methods that significantly improve the amount of residents' activities in the park are programmed. The built environment and control factors are integrated to form an impact curve of the built environment around the park on the amount of residents' activities, and the optimal built environment optimization plan around the park is obtained. This method makes up for the lack of nonlinear relationship and threshold exploration in the existing methods on analyzing the built environment and the amount of residents' activities. This method can obtain built environment indicators and thresholds that significantly improve the amount of residents' activities in the park, and takes into account the impact of differences in natural environmental factors in the park and the surrounding area. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1is a gradient boosting decision tree model diagram of an embodiment;

[0019] Figure 2 is a nonlinear influence curve diagram of the floor area ratio index of the embodiment on the amount of residents' activities;

[0020] Figure 3 It is a nonlinear influence curve diagram of the green space ratio index of the embodiment on the residents' activity volume. DETAILED DESCRIPTION

[0021] The present invention will be further described in detail below in conjunction with the embodiments. Figure 1 As shown in Figure 1, a built environment indicator measurement method that significantly improves residents' activities in the park. By using this measurement method on the built environment within 500 meters around the park, the specific implementation process is as follows:

[0022] Step 1) Use the buffer tool under the ArcGIS 10.7 toolbox to create a 500-meter buffer zone within the boundary of the target park.

[0023] Step 2) Collect and preprocess the built environment data such as building density, volume ratio, green space ratio, etc. within 500 meters around the target park, the activity data such as the duration and frequency of residents' activities, and the natural environment data such as temperature, humidity, and precipitation. Download the average climate and vegetation data for 2023, the land use data for 2023, elevation data, and road vector data through platforms such as the Earth Science Data Center, Google Earth Engine, the Geographic Data Resource Center of the Chinese Academy of Sciences, and the National Meteorological Science Data Center. Collect the activity data such as the duration of residents' activities (entry-exit interval time) and the frequency of activities (number of entries) through the park card swiping data.

[0024] Step 3) Using SPSS software and a linear regression model, with the product of residents' activity duration and activity frequency as independent variables, and built environment and natural environment indicators as independent variables, we analyzed the independent variables that were significantly positively correlated with the dependent variables, that is, the built environment indicators that could significantly increase the amount of residents' activities in the park.

[0025] Step 4) According to the principle of the gradient boosting decision tree (GBDT) model, write R language code, use the machine product of residents' activity duration and activity frequency as independent variables, and the built environment indicators that can significantly increase residents' activities in the park obtained in step 3 as independent variables, and obtain the threshold range of built environment indicators that can significantly increase residents' activities in the park. The principle of the gradient boosting decision tree (GBDT) model is as follows:

[0026] This method sets x as a list of independent variables, including various built environment and natural environment control variables such as building density, volume ratio, and green space ratio; and sets y as an independent variable, namely the amount of residents' activities, including the duration of residents' activities, the frequency of activities, and their product. f(x) is a function that approximates y with x, and the spatial training set of w samples is set to W = {(x1, y1), (x2, y2), …, (x N ,y N )}, the idea of ​​GBDT algorithm is to learn weak learners by fitting residuals in each iteration. Since the activity volume of residents is statistical data, let the Poisson loss function be In addition, in order to minimize the squared error loss function, this study is based on the basis function b(x; θ n ) expands f(x) additively to f n (x), as follows:

[0027]

[0028] Where: f(x) is the final fitting model; a n is the nth decision tree b(x;θ n ) weight; x is the input independent variable; θ n In order to make the loss function in this round of iteration The minimum decision tree construction parameter. This study uses the Gaussian distribution loss function; N is the total number of decision trees.

[0029] To prevent overfitting, the learning rate parameter is used to adjust the step size of negative gradient descent, where the range of the learning rate parameter is [0,1]. The final model completed by the iteration of n weak classifiers is:

[0030] f n (x) = f n-1 (x)+ξa n b(x;θ n ) (2)

[0031] Where: ξ is the learning rate parameter; n is the number of iterations; f n-1 (x) is the model result obtained at the n-1th iteration.

[0032] GBDT provides a variable contribution ranking as the importance of each independent variable to the best fitting model, which is expressed as:

[0033]

[0034] Where: It represents the sum of the changes in the sum of squared errors for each leaf node classification after adding the ith independent variable in the iteration of the nth tree, which represents the contribution of the variable to reducing the loss function in this round of fitting; t represents the non-leaf node in the decision tree, J is the number of all non-leaf nodes in the decision tree; v(t) represents the feature selected by the tth non-leaf node; It indicates the reduction of the model square error loss when the feature selected by the tth non-leaf node is the i-th independent variable.

[0035] The final contribution expression of the i-th independent variable is:

[0036]

[0037] Where: Contribute to each round after n iterations The average value of ; N is the total number of iterations.

[0038] Step 5) Result output includes the following three parts.

[0039] 1. The built environment indicators that significantly improve residents’ activities in parks are shown in Table 1.

[0040] The model results show that among the built environment indicators, except for the average building height and building density, the P values ​​of other built environment indicators are significant, indicating that increasing the values ​​of indicators such as rail station distance, bus station density, service facility density, floor area ratio, average housing price, and green space ratio can significantly increase residents' activities in the park.

[0041] Table 1 Results of linear regression model

[0042]

[0043]

[0044] Note: *, **, *** indicate significance at 10%, 5%, and 1% levels, respectively.

[0045] 2. Nonlinear impact curves of various built environment indicators on residents’ activities, such as Figure 2 , Figure 3 shown.

[0046] 3. Each built environment indicator can significantly increase the threshold range of residents' activity volume.

[0047] According to the nonlinear curve, it can be concluded that when the volume is ≤2.0, the residents' activities can be maintained for the longest time; when the green space ratio is between 0.5-0.6, the residents' activities can be maintained for the longest time.

[0048] The above contents are only used to illustrate the technical solution of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention made by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. A built environment index measurement method that significantly improves residents' activities in parks, characterized by: The steps include: Step 1) Annual built environment index and natural environment index x within the range L around the sample park i Sampling is performed, where i is a natural number; and collecting the park activity data y of residents during the same period, wherein the park activity data of residents is the product of the activity duration and the activity frequency of residents in the park; Step 2) Use SPSS software to apply a linear regression model, with y as the dependent variable and x as the i As the independent variable, analyze the independent variable x that is significantly positively correlated with the dependent variable y. i , we can obtain the built environment index x that can significantly increase the amount of residents’ park activities y i ; Step 3) Write R code based on the machine learning gradient boosting decision tree GBDT model, taking y as the dependent variable and the built environment indicator x obtained in step 2) that can significantly increase the amount of residents' park activities y i is the independent variable, according to the built environment index x i The nonlinear influence curve of residents’ park activity volume y is used to obtain the built environment index x that can significantly increase residents’ park activity volume y. i The threshold interval.

2. The built environment index measurement method for significantly improving residents' activities in parks according to claim 1 is characterized by: The built environment indicators include floor area ratio, green space ratio, distance to rail stations, bus station density, service facility density, average house price, building density and average building height; the natural environment indicators include temperature, humidity, precipitation, wind speed, air pressure, sunshine duration, terrain elevation and terrain slope.

3. The built environment index measurement method for significantly improving residents' activities in parks according to claim 1 is characterized by: The L is 500 meters.

4. The built environment index measurement method for significantly improving residents' activities in parks according to claim 1, characterized in that: In the step 1), the buffer tool under the ArcGIS 10.7 toolbox is used to sample the built environment indicators and natural environment indicators in the L range around the sample park.

Images