TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities

Abstract

The present invention discloses a TOPSIS-guided neural network evaluation method for the livability of urban residential communities. The method includes the following steps: S1. Select appropriate factors and construct an evaluation index system for the livability of urban residential communities; S2. Determine the weights of each element according to the entropy method, and use the TOPSIS method to preliminarily evaluate the livability score and ranking of urban residential communities; S3. According to the livability score ranking of urban residential communities, select the first and last two parts of the communities ranked in the top n% and the bottom n%; S4. Take the livability value range of the top n% of the communities as the high livability value range, and take the livability value range of the bottom n% of the communities as the low livability value range; S5. Use the samples to train the neural network model. The present invention uses TOPSIS to obtain samples of urban residential communities with high and low livability, and to a certain extent overcomes the problem that the determination of factor weights in the existing evaluation method is relatively subjective.

Description

Technical Field

[0001] The present invention relates to the fields of territorial space planning, ecological environment protection, etc., and particularly relates to a TOPSIS-guided neural network evaluation method for the livability of urban residential communities. Background Art

[0002] More and more people live in cities, and cities have become the main places for human production and life. The service level and urban environmental level provided by cities have gradually become important indicators for measuring the quality of life of people in a country or region. With the continuous improvement of living standards, people have put forward higher requirements for living conditions and living environments, and "livable" has become a popular term in urban development and construction. As the most direct habitat for urban residents, the evaluation of the livability of urban residential communities is not only beneficial for the government to understand the current situation of each community and guide the later planning and construction, but also can be used as a reference for community selection, which is of great significance.

[0003] There is no unified definition of a livable community in the academic circle yet. However, in the consensus, a livable community should cover aspects such as a good living environment, a strong cultural and social environment, a comfortable ecological and natural environment, and a convenient living and service environment. There have been some achievements in the research on the evaluation of the livability of urban residential communities, which are divided into two types: subjective evaluation and objective evaluation. Among them, the subjective evaluation methods mainly include questionnaires and analytic hierarchy process, while the objective evaluation methods mainly include principal component analysis, multi-objective decision-making, and clustering methods. In 2020, Liang et al. published "Assessment of the impact of climate change on cities livability in China" in "Science of The Total Environment", using the analytic hierarchy process to evaluate the livability of 288 cities from 2006 to 2016 and quantifying the impact of climate change on the livability of cities. In 2019, Zhang et al. published "Community scale livability evaluation integrating remote sensing, surface observation and geospatial big data" in "International Journal of Applied Earth Observation and Geoinformation", evaluating the livability of several urban residential communities by using TOPSIS and Monte Carlo tests and analyzing the evaluation results from two aspects of sensitivity and uncertainty. In 2018, Paul et al. published "Livability assessment within a metropolis based on the impact of integrated urban geographic factors (IUGFs) on clustering urban centers of Kolkata" in "Cities". When evaluating the livability of cities, the clustering method was used to cluster urban agglomerations into five categories, and different categories represent different livability levels.

[0004] However, there are certain deficiencies in the existing livability evaluation methods: subjective evaluation methods such as the analytic hierarchy process and questionnaires are too subjective, and the results are not easily convincing; in objective evaluation methods such as principal component analysis and multi-objective decision-making, the determination of the weights of factors still requires human intervention and there is a certain degree of subjectivity. How to overcome the problem that the determination of factor weights in the existing livability evaluation methods is relatively subjective is worthy of further study. Summary of the Invention

[0005] Regarding the problems in the related art: The determination of factor weights in the livability evaluation method has a certain degree of subjectivity. When using the index system to evaluate livability, generally more influencing factors will be considered. When evaluating, it is first necessary to determine the weights of each factor and then evaluate the livability, and the subjectivity of factor weights will affect the result of livability evaluation. The present invention proposes a TOPSIS-guided neural network evaluation method for the livability of urban residential communities to overcome the above-mentioned technical problems existing in the existing related technologies.

[0006] To this end, the specific technical solution adopted by the present invention is as follows: This method includes the following steps:

[0007] S1. Select appropriate factor factors, construct an evaluation index system for the livability of urban residential communities, and quantify the evaluation indexes;

[0008] S2. Determine the weights of each element according to the entropy method, and use the TOPSIS method to preliminarily evaluate the livability score and ranking of urban residential communities;

[0009] S3. According to the livability score ranking of urban residential communities, select the first and last parts of the communities ranked in the top n% and the bottom n%, and establish a training sample set for the neural network model;

[0010] S4. Take the livability value range of the top n% of the communities as the high livability value range, and take the livability value range of the bottom n% of the communities as the low livability value range. Establish a loss function according to whether the sample evaluation result falls within the corresponding high livability and low livability value ranges;

[0011] S5. Use the samples to train the neural network model, optimize the model parameters based on the above loss function, construct a neural network evaluation model for the livability of urban residential communities, and obtain the model evaluation result.

[0012] Furthermore, the evaluation index system includes an objective layer, a criterion layer and an element layer;

[0013] Among them, the objective layer is the livability of urban residential communities;

[0014] The criterion layer includes spatial structure, community environment, traffic conditions and service facilities;

[0015] The element layer includes 14 indexes, namely internal greening rate, surrounding greening rate, thermal comfort, population density, plot ratio, road network density, road connectivity, transportation facilities, educational facilities, life services, leisure and entertainment, sports and fitness, parks and shopping facilities.

[0016] Furthermore, the step of determining the weights of each element according to the entropy method and using the TOPSIS method to preliminarily evaluate the livability of urban residential communities, obtaining the livability ranking of urban residential communities and the method evaluation result includes the following steps:

[0017] S21. Unify the monotonicity and dimension of indicators;

[0018] S22. Calculate the weights of 14 indicators using the entropy weight calculation formula;

[0019] The expression of the entropy weight calculation formula is:

[0020]

[0021] In the formula, w j represents the weight of the j-th indicator;

[0022] n represents the number of indicators;

[0023] e j represents the entropy value of the j-th indicator.

[0024] S23. Use the TOPSIS method to calculate the degree of closeness of each urban residential community to the most livable and least livable communities;

[0025] S24. Sort according to the degree of closeness to obtain the livability ranking and method evaluation results of all urban residential communities.

[0026] Furthermore, the expression of the entropy weight calculation formula is:

[0027]

[0028] In the formula, w j represents the weight of the j-th indicator;

[0029] n represents the number of indicators;

[0030] e j represents the entropy value of the j-th indicator.

[0031] Furthermore, the operation expression for using the TOPSIS method to calculate the degree of closeness of each urban residential community to the most livable and least livable communities is:

[0032]

[0033] In the formula, E i represents the degree of closeness and takes values in (0, 1). The closer to 1, the better, and the closer to 0, the worse;

[0034] represents the distance between the i-th urban residential community indicator and the optimal value;

[0035] represents the distance between the i-th urban residential community indicator and the worst value.

[0036] Further, according to the ranking of the livability scores of urban residential communities, select the top n% and the bottom n% of the communities at both ends, and establish a training sample set for the neural network model, including the following steps:

[0037] S31. Statistically analyze the livability scores of urban residential communities evaluated by the method to obtain the corresponding histogram and normal distribution curve;

[0038] S32. Select the top and bottom n% of the communities on the normal curve as the alternative training sample set for the neural network, and adjust n% to obtain different training sample sets.

[0039] Further, the expression of the loss function is:

[0040]

[0041] CS = 1 if E i ≤H max and E i ≥H min |E i ≤L max and E i ≥L min

[0042] where n is the total number of samples;

[0043] represents the number of correctly classified samples. Use the above formula to judge each sample prediction result E i one by one, and the samples that meet the conditions are correctly classified samples;

[0044] H max represents the highest value of the most livable samples;

[0045] H min represents the lowest value of the most livable samples;

[0046] L max represents the highest value of the least livable samples;

[0047] L min represents the lowest value of the least livable samples.

[0048] Further, use the samples to train the neural network model, optimize the model parameters based on the above loss function, and construct a neural network evaluation model for the livability of urban residential communities to obtain the model evaluation result, including the following steps:

[0049] S51. Based on the loss function, use the Sigmoid function as the activation function to construct a three-layer neural network model to evaluate the livability of urban residential communities;

[0050] The expression of the Sigmoid function is:

[0051]

[0052] Wherein, g(x) represents the Sigmoid function;

[0053] x is the input data of the Sigmoid function;

[0054] e is the natural constant;

[0055] S52. Screen the training sample set, compare the classification error rates and goodness-of-fit of different n% of the training sample sets, and use the sample set with the highest goodness-of-fit as the training set;

[0056] S53. Optimize the model parameters, where the model parameters include the number of network layers, the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes, the activation function, and the learning rate;

[0057] S54. Output the evaluation result of the livability of urban residential communities.

[0058] The beneficial effects of the present invention are as follows: Aiming at the problem of subjectivity in determining factor weights in the existing livability evaluation methods, the present invention proposes a TOPSIS-guided neural network evaluation method for the livability of urban residential communities. By using TOPSIS, samples of urban residential communities with high and low livability are obtained, and the classification error rate is used as the loss function of the neural network, which overcomes the problem of relatively subjective determination of factor weights in the existing evaluation methods to a certain extent.

[0059] The method of the present invention has strong adaptability. It has been proved by practice that using the method of the present invention can comprehensively and scientifically evaluate the livability of urban residential communities, meet the needs of actual production, and has stronger practicability compared with the traditional evaluation methods for the livability of urban residential communities. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0061] Figure 1 is a flowchart of the TOPSIS-guided neural network evaluation method for the livability of urban residential communities according to an embodiment of the present invention;

[0062] Figure 2 is a schematic diagram of the TOPSIS-guided neural network evaluation model for the livability of urban residential communities according to an embodiment of the present invention;

[0063] Figure 3 It is a statistical histogram of TOPSIS evaluation results in the TOPSIS-guided neural network evaluation method for the livability of urban residential communities according to an embodiment of the present invention;

[0064] Figure 4 It is a topological structure diagram of a BP neural network in the TOPSIS-guided neural network evaluation method for the livability of urban residential communities according to an embodiment of the present invention;

[0065] Figure 5 It is a schematic diagram of cross-validation between a neural network and TOPSIS in the TOPSIS-guided neural network evaluation method for the livability of urban residential communities according to an embodiment of the present invention. Detailed implementation manners

[0066] To further illustrate the embodiments, the present invention provides accompanying drawings. These accompanying drawings are part of the disclosure of the present invention, mainly used to illustrate the embodiments, and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are usually used to represent similar components.

[0067] According to an embodiment of the present invention, a TOPSIS-guided neural network evaluation method for the livability of urban residential communities is provided.

[0068] Now, the present invention will be further described in conjunction with the accompanying drawings and specific implementation manners. As Figures 1-5 shown, according to an embodiment of the present invention, a TOPSIS-guided neural network evaluation method for the livability of urban residential communities is provided, and the method includes the following steps:

[0069] S1. Select appropriate factors, construct an evaluation index system for the livability of urban residential communities, and quantify the evaluation indexes;

[0070] There are many indexes affecting livability, and a scientific evaluation index system for livability is a crucial step. When constructing the evaluation system, the principles of scientificity, systematicness, pertinence, and operability should be followed.

[0071] The evaluation index system includes an objective layer, a criterion layer, and an element layer;

[0072] Among them, the objective layer is the livability of urban residential communities;

[0073] The criterion layer includes spatial structure, community environment, traffic conditions, and service facilities;

[0074] The element layer includes 14 indicators, namely internal greening rate, surrounding greening rate, thermal comfort, population density, plot ratio, road network density, road connectivity, transportation facilities, educational facilities, life services, leisure and entertainment, sports and fitness, parks and shopping facilities.

[0075] S2. Determine the weights of each element according to the entropy method, and use the TOPSIS method to preliminarily evaluate the livability score and ranking of urban residential communities;

[0076] To overcome the subjectivity in weight determination, the entropy method is used for weighting; when weighting, if there are negative numbers in the original data, the data needs to be made non - negative. For positive and negative indicators, the monotonicity and dimension of the indicators need to be unified, and then the weights are determined to obtain the evaluation results of the TOPSIS method.

[0077] Step S2 includes the following steps:

[0078] S21. Unify the monotonicity and dimension of the indicators;

[0079] S22. Calculate the weights of 14 indicators using the entropy method weight calculation formula;

[0080] Among them, the expression of the entropy method weight calculation formula is:

[0081]

[0082] In the formula, w j represents the weight of the j - th indicator;

[0083] n represents the number of indicators;

[0084] e j represents the entropy value of the j - th indicator.

[0085] S23. Use the TOPSIS method to calculate the degree of closeness of each urban residential community to the most livable and least livable communities, and its operation expression is:

[0086]

[0087] In the formula, E i represents the degree of closeness and takes values in (0, 1). The closer to 1, the better; the closer to 0, the worse;

[0088] represents the distance between the i - th urban residential community and the optimal value;

[0089] represents the distance between the i - th urban residential community and the worst value.

[0090] S24. Obtain the livability scores and rankings of all urban residential communities according to the proximity degree ranking.

[0091] S3. According to the livability score ranking of urban residential communities, select the first and last parts of the communities ranked in the top n% and the bottom n%, and establish a training sample set for the neural network model, as Figure 3 shown, including the following steps:

[0092] S31. Statistically analyze the livability scores of urban residential communities in the evaluation results of the method to obtain the corresponding histogram and normal distribution curve;

[0093] S32. Select the first and last n% of the communities on the normal curve as the alternative training sample set for the neural network. Adjust n% (such as setting n% to 20%, 10%, and 5% respectively) to obtain different training sample sets.

[0094] Using the characteristics of the most livable and least livable urban residential communities can reduce the dependence of the model on the TOPSIS evaluation results. Therefore, select the data at the beginning and end of the normal curve of the livability scores as the training set. Different numbers of training samples have a greater impact on the fitting results of the model, and it is necessary to select different sizes of training sets for comparative experiments.

[0095] S4. Take the range of livability values of the top n% of the communities as the high livability value range, and the range of livability values of the bottom n% of the communities as the low livability value range. Establish a loss function according to whether the sample evaluation results fall within the corresponding high and low livability value ranges. The expression of the loss function is:

[0096]

[0097] CS = 1 if E i ≤H max and E i ≥H min |E i ≤L max and E i ≥L min

[0098] where n is the total number of samples; represents the number of correctly classified samples. Use the above formula to judge each sample prediction result E i one by one. Those that meet the conditions are correctly classified samples. H max represents the highest value of the most livable samples; H min represents the lowest value of the most livable samples; L max represents the highest value of the least livable samples; L min represents the lowest value of the least livable samples.

[0099] S5. Use the samples to train a neural network model, optimize the model parameters based on the above loss function, construct a neural network evaluation model for the livability of urban residential communities, and obtain the model evaluation results, as Figure 2 、 Figure 4 shown, including the following steps:

[0100] S51. Based on the loss function, use the Sigmoid function as the activation function to construct a three-layer neural network model to evaluate the livability of urban residential communities.

[0101] The expression of the Sigmoid function is:

[0102]

[0103] In the formula, g(x) represents the Sigmoid function, x is the input data of the Sigmoid function, and e is the natural constant (approximately equal to 2.718281828459).

[0104] S52. Screen the training sample set, compare the classification error rates and goodness-of-fit of different n% of the training sample sets, and use the sample set with the highest goodness-of-fit as the training set;

[0105] S53. Optimize the model parameters, where the model parameters include the number of network layers, the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes, the activation function, and the learning rate;

[0106] S54. Output the evaluation results of the livability of urban residential communities.

[0107] Different parameters will affect the evaluation results of the neural network. Set up a comparative experiment to determine the optimal size of the training set, and determine other appropriate parameters according to empirical formulas, experimental characteristics, and comparative experiments, etc. Construct a neural network evaluation model for the livability of urban residential communities to obtain the evaluation results, reducing the subjective influence existing in the evaluation process.

[0108] In addition, the experimental data in the present invention includes vector data: administrative divisions, land use plans, transportation road networks, and the third land survey data; image data: Landsat, high-resolution, MODIS, and night light remote sensing data; Internet data: traffic facilities, education, shopping, parks, public service facilities, entertainment, etc. interest points and community housing prices, etc. crawled from Baidu Maps and Gaode Maps; statistical data: township-level population data, all of which have been preprocessed.

[0109] In summary, by means of the above technical solutions of the present invention, the present invention addresses the problem of subjectivity in determining factor weights in existing livability evaluation methods, and proposes a TOPSIS-guided neural network evaluation method for the livability of urban residential communities. By using TOPSIS, samples of urban residential communities with high and low livability are obtained, and the classification error rate is used as the loss function of the neural network, which overcomes to a certain extent the problem of relatively subjective determination of factor weights in existing evaluation methods. The method of the present invention has strong adaptability. It has been proven by practice that using the method of the present invention can comprehensively and scientifically evaluate the livability of urban residential communities, meet the needs of actual production, and has stronger practicability compared with traditional methods for evaluating the livability of urban residential communities.

[0110] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0111] The above-described embodiments merely represent several implementation manners of the present invention, and the description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the invention patent should be subject to the appended claims.

Claims

1. TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities, Characterized in that, This method includes the following steps: S1. Select appropriate factors and construct an evaluation index system for the livability of urban residential communities, and quantify the evaluation indexes; S2. Determine the weights of each element according to the entropy method, and use the TOPSIS method to preliminarily evaluate the livability score and ranking of urban residential communities; S3. According to the livability score ranking of urban residential communities, select the first and last parts of the top n% and bottom n% of the communities to establish a training sample set for the neural network model; S4. Take the livability value range of the top n% of the communities as the high livability value range, and the livability value range of the bottom n% of the communities as the low livability value range, and establish a loss function according to whether the sample evaluation results fall within the corresponding high and low livability value ranges; S5. Use the samples to train the neural network model, optimize the model parameters based on the above loss function, construct a neural network evaluation model for the livability of urban residential communities, and obtain the model evaluation results.

2. The TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities according to Claim 1, Characterized in that, The evaluation index system includes an objective layer, a criterion layer and an element layer; Among them, the objective layer is the livability degree of urban residential communities; The criterion layer includes spatial structure, community environment, traffic conditions and service facilities; The element layer includes 14 indexes, namely internal greening rate, surrounding greening rate, thermal comfort, population density, plot ratio, road network density, road connectivity, traffic facilities, educational facilities, life services, leisure and entertainment, sports and fitness, parks and shopping facilities.

3. The TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities according to Claim 2, Characterized in that, The step of determining the weights of each element according to the entropy method and using the TOPSIS method to preliminarily evaluate the livability score and ranking of urban residential communities includes the following steps: S21. Unify the index monotonicity and dimension; S22. Calculate the weights of 14 indexes by using the entropy method weight calculation formula; S23. Use the TOPSIS method to calculate the proximity of each urban residential community to the most livable and least livable communities; S24. Sort according to the proximity degree to obtain the livability score and ranking of all urban residential communities.

4. The TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities according to Claim 3, Characterized in that, The expression of the entropy method weight calculation formula is: where w j represents the weight of the j-th index; n represents the number of indexes; e j represents the entropy value of the j-th index.

5. The TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities according to Claim 3, Characterized in that, The operation expression for using the TOPSIS method to calculate the proximity of each urban residential community to the most livable and least livable communities is: where E i represents the proximity of the i-th urban residential area and takes values in the range (0, 1). A value closer to 1 indicates better quality, while a value closer to 0 indicates worse quality; Indicates the distance between the \(i\)-th urban residential area and the optimal value; Indicates the distance between the \(i\)-th urban residential area and the worst value.

6. The TOPSIS-guided Neural Network Evaluation Method for the Livability of Urban Residential Communities according to Claim 5, Characterized in that, According to the ranking of the livability scores of urban residential communities, select the first and last two parts of the communities ranked in the top n% and the bottom n%, and establish a training sample set for the neural network model, including the following steps: S31. Statistically analyze the livability scores of urban residential communities in the evaluation results of the method to obtain the corresponding histogram and normal distribution curve; S32. Select the first and last n% of the communities in the normal curve as the alternative training sample set for the neural network, and adjust n% to obtain different training sample sets.

7. The TOPSIS-guided neural network evaluation method for the livability of urban residential communities according to claim 6, characterized in that, The expression of the loss function is: CS = 1 if E i ≤H max and E i ≥H min |E i ≤L max and E i ≥L min where n is the total number of samples; Denote the number of correctly classified samples. Use the above formula for the sample prediction result E i Judge one by one, and those that meet the conditions are correctly classified samples; H max Indicates the highest value of the most livable sample; H min Indicates the lowest value of the most livable sample; L max Indicates the highest value of the least habitable sample; L min Indicates the lowest value of the least habitable sample.

8. The TOPSIS-guided neural network evaluation method for the livability of urban residential communities according to claim 7, characterized in that, Using the samples to train the neural network model, optimizing the model parameters based on the above loss function, constructing a neural network evaluation model for the livability of urban residential communities, and obtaining the model evaluation results, including the following steps: S51. Based on the loss function, use the Sigmoid function as the activation function to construct a three-layer neural network model to evaluate the livability of urban residential communities; The expression of the Sigmoid function is: In the formula, g(x) represents the Sigmoid function; x is the input data of the Sigmoid function; e is the natural constant; S52. Screen the training sample set, compare the classification error rates and goodness-of-fit of different training sample sets with n%, and use the sample set with the highest goodness-of-fit as the training set; S53. Optimize the model parameters, and the model parameters include the number of network layers, the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes, the activation function, and the learning rate; S54. Output the evaluation results of the livability of urban residential communities.

Images