Method for constructing sustainable city update evaluation index system
By optimizing the scoring function weights for the expansion and gradient descent of the original data set, the problem of insufficient robustness and generalization of the existing evaluation system is solved, and the objective and accuracy of sustainable urban renewal evaluation is achieved.
Patent Information
- Application Number
- CN202410982302.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-15
AI Technical Summary
Due to strong subjectivity and insufficient data, the existing sustainable urban renewal evaluation system has poor robustness and generalization, and lacks objectivity and accuracy.
By extending the original data set, the machine learning and gradient descent method are used to optimize the scoring function weights, and the target evaluation index system is constructed to improve the robustness and generalization of the evaluation system.
An objective, simple and accurate evaluation method for the evaluation of sustainable urban renewal is realized, and an evaluation index system with rich data volume and optimized weight is provided.
Smart Images

Figure CN118966820A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of sustainable urban renewal evaluation, and in particular to a method for constructing a sustainable urban renewal evaluation index system. Background Art
[0002] For a long time, urban renewal has become an important part of urban research because of its commitment to solving urban decay problems, improving urban economy, and enhancing urban reputation. However, urban renewal has been criticized for always focusing on short-term economic goals, excessive pursuit of profits, and "large-scale demolition and construction" leading to increased social exclusion and weakened social identity. Different from the traditional renewal model, sustainable urban renewal is a cost-intensive, multifunctional, and eco-friendly urban renewal model that promotes the coordinated development of resources, environment, economy, society, and residents' lives in accordance with the "people-oriented" idea and the optimization of urban ecological environment. It is becoming a modern governance model that conforms to global sustainable development. Sustainable urban renewal decision-making emphasizes a series of behaviors in which relevant actors use scientific technologies and methods to make decisions on the mode selection, timing design, spatial layout, funding arrangement, and benefit distribution of urban renewal in the stages of renewal planning, scheme design, and mid- and late-stage implementation adjustments.
[0003] However, the existing sustainable urban renewal evaluation systems are mostly artificially formulated and highly subjective. At the same time, due to the lack of relevant data on sustainable urban renewal, the construction of the evaluation system is random, with poor robustness and generalization. Summary of the invention
[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a method for constructing a sustainable urban renewal evaluation index system, by expanding the original data set and enriching the data volume to improve the robustness and generalization of the evaluation system; through the gradient descent method, the weight coefficient of the scoring function is optimized to provide an objective and simple evaluation method for sustainable urban renewal evaluation.
[0005] To achieve the above object, the present invention provides the following solutions:
[0006] A method for constructing a sustainable urban renewal evaluation index system, comprising:
[0007] Preset at least one unweighted indicator;
[0008] Constructing a scoring function based on all the unweighted indicators;
[0009] Collecting the indicator data and sustainable city scores corresponding to all the unweighted indicators of multiple sustainable urban renewals to obtain an original data set;
[0010] According to a preset data volume standard, the original data set is expanded by using machine learning to obtain an expanded data set;
[0011] Inputting the indicator data in the extended data set into the scoring function to obtain a scoring result;
[0012] Calculating the loss value of the sustainable city score corresponding to the score result and the indicator data according to a preset loss function;
[0013] According to the loss value, the weights in the scoring function are optimized using the gradient descent method to obtain an iterative function, and the process returns to the step of "inputting the indicator data in the extended data set into the scoring function to obtain a scoring result";
[0014] When the loss value belongs to a preset threshold range, the iterative function of the last iteration is output to obtain a target evaluation index system.
[0015] Preferably, it also includes:
[0016] Calculating the confusion matrix of the target evaluation index system using a preset validation set;
[0017] Calculating precision using the confusion matrix;
[0018] When the accuracy reaches a preset output standard, the target evaluation index system is output.
[0019] Preferably, a scoring function is constructed according to all the unweighted indicators, including:
[0020] The unweighted index is multiplied and accumulated with the preset initial weight to obtain an initial function; the calculation formula of the initial function is: F = a 1 x 1 +a 2 x 2 +...+a n x n ; F is the calculation result of the initial function; x 1 is the first unweighted indicator; a 1 is the weight coefficient of the first unweighted indicator; n is the total number of unweighted indicators;
[0021] The weight of the initial function is set to 1, and the unweighted indicator construction scoring function is obtained; the calculation formula of the unweighted indicator construction scoring function is: F = x 1 +x 2 +...+x n .
[0022] Preferably, according to a preset data volume standard, the original data set is expanded using machine learning to obtain an expanded data set, including:
[0023] Build a basic GAN model; the basic GAN model includes: a generator and a discriminator;
[0024] Constructing the objective function of the basic GAN model;
[0025] Setting a first loss function of the generator and a second loss function of the discriminator;
[0026] According to the original data set, the generator and the discriminator are alternately trained using the first loss function and the second loss function to obtain a trained generation model;
[0027] According to the data volume standard, the original data set is expanded using the generation model to obtain the extended data set.
[0028] Preferably, according to the loss value, the weights in the scoring function are optimized using a gradient descent method to obtain an iterative function, including:
[0029] Calculating the gradient of the loss function;
[0030] The weights in the scoring function are updated using the gradient; the updated calculation formula is: ω (t+1) is the weight in the scoring function optimized for the tth time; α is the learning rate; is the loss function with respect to ω (t) The gradient of
[0031] The original weights in the scoring function are replaced by the updated weights in the scoring function to obtain the iterated function.
[0032] Preferably, the method for obtaining the sustainable city score adopts online voting; the sustainable city score is divided into 1 to 5 levels, with a total of 5 levels.
[0033] Preferably, it also includes:
[0034] Using sampling with replacement, a group of data is removed from the extended data set each time to obtain a group of extended data subsets, and all data in the extended data set are traversed to obtain multiple groups of the extended data subsets;
[0035] Iterating the iterative function using each group of the extended data subsets to obtain multiple intermediate evaluation index systems;
[0036] Using a preset validation set to validate all of the intermediate evaluation index systems to obtain multiple precision rates;
[0037] The intermediate evaluation index system corresponding to the highest precision rate is constructed as the target evaluation index system.
[0038] Preferably, the loss function is mean square error.
[0039] The present invention discloses the following technical effects:
[0040] The present invention provides a method for constructing a sustainable urban renewal evaluation index system. By expanding the original data set, the problem of too low data volume is solved, and the data volume is enriched as well as the robustness and generalization of the evaluation system are improved. By using the gradient descent method, the problem of finding the optimal weight is solved, and the weight coefficient of the scoring function is optimized. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0042] Figure 1 A schematic flow chart of a method for constructing a sustainable urban renewal evaluation index system provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.
[0044] The purpose of the present invention is to provide a method for constructing a sustainable urban renewal evaluation index system, by expanding the original data set and enriching the data volume to improve the robustness and generalization of the evaluation system; by using the gradient descent method, the weight coefficient of the scoring function is optimized to provide an objective and simple evaluation method for sustainable urban renewal evaluation.
[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0046] Figure 1A flow chart of a method for constructing a sustainable urban renewal evaluation index system provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the present invention provides a method for constructing a sustainable urban renewal evaluation index system, comprising:
[0047] Step 100: Preset at least one unweighted indicator;
[0048] Step 200: construct a scoring function based on all unweighted indicators;
[0049] Step 300: Collect indicator data and sustainable city scores corresponding to all non-weighted indicators of multiple sustainable urban renewals to obtain an original data set;
[0050] Step 400: according to a preset data volume standard, the original data set is expanded using machine learning to obtain an expanded data set;
[0051] Step 500: input the index data in the extended data set into the scoring function to obtain a scoring result;
[0052] Step 600: Calculate the loss value of the sustainable city score corresponding to the score result and the indicator data according to a preset loss function;
[0053] Step 700: According to the loss value, the weights in the scoring function are optimized using the gradient descent method to obtain an iterative function, and the process returns to the step of “inputting the indicator data in the extended data set into the scoring function to obtain a scoring result”;
[0054] Step 800: When the loss value belongs to the preset threshold range, the iterative function of the last iteration is output to obtain the target evaluation index system.
[0055] Furthermore, it also includes:
[0056] The confusion matrix of the target evaluation index system is calculated using the preset validation set; the calculation formula of the confusion matrix is:
[0057]
[0058] Among them, a15 is the number of true scores of 1 and predicted scores of 5.
[0059] The accuracy is calculated using the confusion matrix; the accuracy calculation formula is:
[0060]
[0061] Among them, a mn is the number of cases where the true rating is m and the predicted rating is n.
[0062] When the accuracy reaches 0.98, the target evaluation index system is output.
[0063] Optionally, a scoring function is constructed based on all unweighted indicators, including:
[0064] The unweighted index is multiplied and added with the preset initial weight to obtain the initial function; the calculation formula of the initial function is: F = a 1 x 1 +a 2 x 2 +...+a n x n ; F is the calculation result of the initial function; x 1 is the first unweighted indicator; a 1 is the weight coefficient of the first unweighted indicator; n is the total number of unweighted indicators;
[0065] The weight of the initial function is set to 1, and the scoring function constructed by the unweighted indicator is obtained; the calculation formula of the scoring function constructed by the unweighted indicator is: F = x 1 +x 2 +...+x n .
[0066] Furthermore, in order to distinguish the influence of different indicators, a degree coefficient can be added to each item in actual operation, for example That is, to increase the first degree coefficient α to the first indicator 1 ; Since there is no good reference data for the degree coefficient, the random forest machine learning algorithm can be used in actual training to optimize the degree coefficient and the weight coefficient together.
[0067] Specifically, according to the preset data volume standard, machine learning is used to expand the original data set to obtain an extended data set, including:
[0068] Build a basic GAN model; the basic GAN model includes: generator and discriminator;
[0069] Construct the objective function of the basic GAN model;
[0070] Set the first loss function of the generator and the second loss function of the discriminator;
[0071] According to the original data set, the generator and the discriminator are alternately trained using the first loss function and the second loss function to obtain a trained generation model;
[0072] According to the data volume standard, the generative model is used to expand the original data set to obtain the extended data set.
[0073] Furthermore, according to the loss value, the weights in the scoring function are optimized using the gradient descent method to obtain an iterative function, including:
[0074] Calculate the gradient of the loss function;
[0075] The gradient is used to update the weights in the scoring function; the updated calculation formula is: ω(t+1) is the weight in the scoring function optimized for the tth time; α is the learning rate; is the loss function with respect to ω (t) The gradient of
[0076] The original weights in the scoring function are replaced with the updated weights in the scoring function to obtain an iterative function.
[0077] Specifically, the sustainable city score is obtained through online voting; the sustainable city score is divided into 5 levels from 1 to 5, with a total of 5 levels.
[0078] Furthermore, it also includes:
[0079] Using sampling with replacement, a group of data is removed from the extended data set each time to obtain a group of extended data subsets, and all data in the extended data set are traversed to obtain multiple groups of extended data subsets;
[0080] Iterate the iterative function using each group of extended data subsets to obtain multiple intermediate evaluation index systems;
[0081] Use the preset validation set to validate all intermediate evaluation index systems and obtain multiple precision rates;
[0082] The intermediate evaluation index system corresponding to the highest accuracy is constructed as the target evaluation index system.
[0083] Optionally, the loss function is mean square error.
[0084] Optionally, the unweighted indicators of this embodiment include: proportion of informal settlements, road network density, citizen participation, PM2.5 concentration and change rate, proportion of open area, population density change rate, regional warming sensitivity index, chlorophyll a concentration, number of new energy charging piles, night light intensity, per capita mobile data traffic, urban green space carbon fixation and oxygen release, etc. The number of unweighted indicators will affect the final effect of the evaluation system. Too few indicators will lead to a stronger randomness of the evaluation system, and more indicators will lead to a large amount of manpower and material resources in the data collection stage. Therefore, it is more reasonable to control the number of indicators between 10-100.
[0085] Furthermore, in order to increase the richness of the data, when collecting the unweighted indicators of cities, cities are selected evenly and randomly to avoid indirect influences of geography, economy, etc. on the construction of the evaluation system.
[0086] The beneficial effects of the present invention are as follows:
[0087] The present invention enriches the data volume and improves the robustness and generalization of the evaluation system by expanding the original data set; through the gradient descent method, the weight coefficient of the scoring function is optimized, and the obtained target evaluation index system provides an objective and simple evaluation method for sustainable urban renewal evaluation.
[0088] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0089] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of the present invention. At the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A method for constructing a sustainable urban renewal evaluation index system, characterized in that: include: Preset at least one unweighted indicator; Constructing a scoring function based on all the unweighted indicators; Collecting the indicator data and sustainable city scores corresponding to all the unweighted indicators of multiple sustainable urban renewals to obtain an original data set; According to a preset data volume standard, the original data set is expanded by using machine learning to obtain an expanded data set; Inputting the indicator data in the extended data set into the scoring function to obtain a scoring result; Calculating the loss value of the sustainable city score corresponding to the score result and the indicator data according to a preset loss function; According to the loss value, the weights in the scoring function are optimized using the gradient descent method to obtain an iterative function, and the process returns to the step of "inputting the indicator data in the extended data set into the scoring function to obtain a scoring result"; When the loss value belongs to a preset threshold range, the iterative function of the last iteration is output to obtain a target evaluation index system.
2. A method for constructing a sustainable urban renewal evaluation index system according to claim 1, characterized in that: Also includes: Calculating the confusion matrix of the target evaluation index system using a preset validation set; Calculating precision using the confusion matrix; When the accuracy reaches a preset output standard, the target evaluation index system is output.
3. A method for constructing a sustainable urban renewal evaluation index system according to claim 1, characterized in that: The scoring function is constructed based on all the unweighted indicators, including: The unweighted index is multiplied and added with the preset initial weight to obtain an initial function; the calculation formula of the initial function is: F = a1x1 + a2x2 + ... + a n x n ; F is the calculation result of the initial function; x1 is the first unweighted indicator; a1 is the weight coefficient of the first unweighted indicator; n is the total number of unweighted indicators; The weight of the initial function is set to 1 to obtain the unweighted indicator construction scoring function; the calculation formula of the unweighted indicator construction scoring function is: F = x1 + x2 + ... + x n .
4. A method for constructing a sustainable urban renewal evaluation index system according to claim 1, characterized in that: According to the preset data volume standard, the original data set is expanded by using machine learning to obtain an expanded data set, including: Build a basic GAN model; the basic GAN model includes: a generator and a discriminator; Constructing the objective function of the basic GAN model; Setting a first loss function of the generator and a second loss function of the discriminator; According to the original data set, the generator and the discriminator are alternately trained using the first loss function and the second loss function to obtain a trained generation model; According to the data volume standard, the original data set is expanded using the generation model to obtain the extended data set.
5. A method for constructing a sustainable urban renewal evaluation index system according to claim 1, characterized in that: According to the loss value, the weights in the scoring function are optimized using the gradient descent method to obtain an iterative function, including: Calculating the gradient of the loss function; The weights in the scoring function are updated using the gradient; the updated calculation formula is: ω (t+1) is the weight in the scoring function optimized for the tth time; α is the learning rate; is the loss function with respect to ω (t) The gradient of The original weights in the scoring function are replaced by the updated weights in the scoring function to obtain the iterated function.
6. A method for constructing a sustainable urban renewal evaluation index system according to claim 1, characterized in that: The sustainable city score is obtained by online voting; the sustainable city score is divided into 5 levels from 1 to 5, with a total of 5 levels.
7. A method for constructing a sustainable urban renewal evaluation index system according to claim 4, characterized in that: Also includes: Using sampling with replacement, a group of data is removed from the extended data set each time to obtain a group of extended data subsets, and all data in the extended data set are traversed to obtain multiple groups of the extended data subsets; Iterating the iterative function using each group of the extended data subsets to obtain multiple intermediate evaluation index systems; Using a preset validation set to validate all of the intermediate evaluation index systems to obtain multiple precision rates; The intermediate evaluation index system corresponding to the highest precision rate is constructed as the target evaluation index system.
8. A method for constructing a sustainable urban renewal evaluation index system according to claim 5, characterized in that: The loss function is the mean square error.
