Urban ecosystem service flow analysis system and method based on neural network

Abstract

The invention relates to an urban ecosystem service flow analysis system and method based on a neural network. The system comprises a geographic information module, a data processing module and a neural network model. The geographic information module is used for calibrating geographic grid units and establishing a mapping relation between the target data and the geographic grid units; the geographic grid unit comprises a supply unit and a demand unit; the data processing module is used for preprocessing the source data to obtain target data of each geographic grid unit; the target data comprises population data, economic data, geographic space data and page view data; the neural network model is used for training according to a training sample constructed by the target data, and analyzing the ecosystem service flow through the trained neural network model. The method has the advantages that the dynamic relation among the supply, the demand and the service flow can be accurately described, and the analysis accuracy of the service flow of the urban ecosystem is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent measurement technology for urban ecosystems, and in particular to a system and method for analyzing urban ecosystem service flows based on neural networks. Background Technology

[0002] Urban ecosystem services, serving as a link between natural ecosystems and human socio-economic systems, are a crucial foundation for sustainable urban development. The supply capacity and service delivery efficiency of urban ecosystems directly determine the quality of the urban ecological environment and the quality of life for residents. Urban ecosystem services are also vital for ensuring urban resilience and promoting ecological civilization. Therefore, the quantitative assessment and optimized regulation of urban ecosystem services are important directions in urban planning research and crucial practical needs for advancing new urbanization.

[0003] The realization of ecosystem service value is jointly influenced by supply-side characteristics, demand-side characteristics, and service flow characteristics. However, a unified and operational analytical framework has not yet been established, making it difficult to systematically characterize the dynamic relationship among supply, demand, and service flow. Traditional urban ecosystem assessments primarily focus on the supply side of ecosystem services, neglecting the transmission process from supply to demand, actual accessibility, and supply-demand matching. Research on the spatial transmission characteristics of ecosystem service flows and the demand-side generation mechanism is insufficient, making it difficult to support refined assessment and analysis. Furthermore, traditional urban ecosystem service assessments suffer from relatively limited data foundations and insufficient spatial coverage, failing to accurately reflect the service flow characteristics of urban ecosystems. This leads to imbalances in the supply and demand of ecosystem services, insufficient reliability and practicality of assessment results, and an inability to meet the needs of refined and scientific decision-making support in urban planning and ecological management. Summary of the Invention

[0004] The technical problem to be solved by this invention is: addressing the technical problems existing in the prior art, this invention provides a neural network-based urban ecosystem service flow analysis system and method that can accurately characterize the dynamic relationship between supply, demand and service flow, and improve the accuracy of urban ecosystem service flow analysis.

[0005] To solve the above-mentioned technical problems, the technical solution proposed by this invention is: a city ecosystem service flow analysis system based on neural networks, including a geographic information module, a data processing module, and a neural network model; The geographic information module is used to calibrate geographic grid units and establish a mapping relationship between target data and geographic grid units; the geographic grid unit includes supply units and demand units; The data processing module is used to preprocess the source data to obtain the target data for each geographic grid unit; the target data includes population data, economic data, geospatial data, and access data. The neural network model is used to train on training samples constructed from the target data, and the trained neural network model is used to analyze ecosystem service flows.

[0006] Furthermore, the neural network model uses the population data, economic data, and geospatial data to construct vectors as input layers, and the access volume data as output layers.

[0007] Furthermore, the neural network contains at least one hidden layer, the input function of the neural network is a weighted summation function, the transfer function is a sigmoid function, and the learning rule is a backpropagation rule.

[0008] Furthermore, it also includes a supply analysis module, used to construct a supply transfer function based on the access volume and the supply volume of the supply unit, determine the potential supply volume of the demand unit based on the supply transfer function, and evaluate the supply and demand status of the demand unit based on the potential supply volume and the access volume.

[0009] Furthermore, the economic data includes housing price indicators and production value indicators; the geospatial data includes area, grid type, straight-line distance, traffic distance, and traffic duration; the geographic grid units are divided according to geographic boundaries; and the area of ​​the geographic grid unit is less than a preset area threshold.

[0010] A method for analyzing urban ecosystem service flows based on neural networks includes the following steps: Step S1. In a geographic information system, geographic grid cells are labeled, wherein the geographic grid cells include supply cells and demand cells; and a mapping relationship between target data and the geographic grid cells is established. Step S2. Preprocess the source data to obtain the target data for each geographic grid cell; the target data includes population data, economic data, geospatial data, and access data; Step S3. Construct a neural network model, train the neural network model by constructing training samples based on the target data, and analyze the ecosystem service flow through the trained neural network model.

[0011] Furthermore, the neural network model uses the population data, economic data, and geospatial data to construct vectors as input layers, and the access volume data as output layers.

[0012] Furthermore, the neural network contains at least one hidden layer, the input function of the neural network is a weighted summation function, the transfer function is a sigmoid function, and the learning rule is a backpropagation rule.

[0013] Furthermore, it also includes step S4, constructing a supply transfer function based on the number of visits and the supply quantity of the supply unit, determining the potential supply quantity of the demand unit based on the supply transfer function, and evaluating the supply and demand status of the demand unit based on the potential supply quantity and the number of visits.

[0014] Furthermore, the economic data includes housing price indicators and production value indicators; the geospatial data includes area, grid type, straight-line distance, traffic distance, and traffic duration; the geographic grid units are divided according to geographic boundaries; and the area of ​​the geographic grid unit is less than a preset area threshold.

[0015] Compared with the prior art, the advantages of the present invention are as follows: 1. This invention constructs a neural network model, trains the neural network model with real data, and analyzes urban ecosystem service flows using the trained neural network model. It can realistically reflect the impact of different types of factors as input layers of the neural network model on urban ecosystem service flows, thereby comprehensively, accurately reflecting the impact of various factors on urban ecosystem service flows and improving the accuracy of urban ecosystem service flow analysis.

[0016] 2. This invention uses multiple types of data, such as population data, economic data, geospatial data, and access volume, as the data foundation for urban ecosystem service flow analysis. By fusing various types of data through a neural network model, it can more comprehensively and holistically reflect the impact of various factors on urban ecosystem service flow, making the analysis system more representative and reducing bias caused by data samples. Attached Figure Description

[0017] Figure 1 This is a structural block diagram of an urban ecosystem service flow analysis model according to a specific embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the neural network model structure according to a specific embodiment of the present invention.

[0019] Figure 3 This is an error graph of training a neural network model in a specific embodiment of the present invention.

[0020] Figure 4 This is a flowchart of a specific embodiment of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0022] This embodiment presents a neural network-based urban ecosystem service flow analysis system, such as... Figure 1 As shown, it includes a geographic information module, a data processing module, and a neural network model. The geographic information module is used to calibrate geographic grid units and establish a mapping relationship between target data and geographic grid units. Geographic grid units include supply units and demand units. The data processing module is used to preprocess the source data to obtain the target data for each geographic grid unit. The target data includes population data, economic data, geospatial data, and access data. The neural network model is used to train based on training samples constructed from the target data and to analyze ecosystem service flows through the trained neural network model.

[0023] In this embodiment, parks and green spaces are used as representatives of typical urban ecosystems, and a geographic information module is built based on GIS (Geography Information System). In GIS, parks are marked as supply units according to their geographical boundaries, and other non-park areas are divided into demand units according to a preset grid division rule. In this embodiment, the grid division rule for demand units can be selected according to the input vector parameters of the neural network model, such as dividing demand units in the GIS system using rectangular division, or dividing geographic grid units according to geographic boundaries. In this implementation, census data and other population statistics can be selected as the population data in the geographic grid units, i.e., it is preferable to divide demand units according to building boundaries (building boundaries are artificial geographic boundaries) to ensure the accuracy of the obtained population data. If geographic grid units are divided according to rectangular division or other methods, when there are obstructive geographic boundaries such as walls or rivers in the grid, it is preferable to further divide the grid into smaller geographic grid units according to these geographic boundaries. Further preferably, the area of ​​the divided geographic grid units is less than a preset area threshold, and the area threshold is preferably a value within the range of [10, 20] hectares.

[0024] In this embodiment, the access data of the demand unit to the supply unit can be obtained through questionnaires and sampling statistics. Preferably, the number of park visitors and the source of the visitors (i.e., which demand unit the visitors come from) can be obtained through mobile phone signaling.

[0025] In this embodiment, preferred economic data includes housing price indicators and production value indicators, while geospatial data includes area, grid type, straight-line distance, traffic distance, and traffic duration. The neural network model uses population data, economic data, and geospatial data to construct vectors as the input layer and access volume data as the output layer. In this embodiment, the grid type is preferably distinguished according to the grid type of the demand unit, including residential grids, commercial grids, industrial grids, etc.; the straight-line distance is preferably the minimum straight-line distance between the demand unit and the entrance of the supply unit (park); the traffic distance is preferably the minimum traffic distance (driving distance) between the demand unit and the entrance of the supply unit (park). It should be noted that the parameters are not limited to those listed above, and other parameters can be selected to construct the input layer vector of the neural network model. For example, geospatial data can also include seasonal or monthly time data, so that the urban ecosystem service flow analysis can reflect time characteristics and reflect the changes in demand from demand units to supply units at different times or seasons. In this embodiment, data preprocessing includes screening and cleaning of raw data, removal of invalid and abnormal data, and data standardization, normalization, and spatialization processing to ensure data format and standardization, and high data accuracy.

[0026] In this embodiment, the neural network includes at least one hidden layer. The input function of the neural network is a weighted summation function, the transfer function is a sigmoid function, and the learning rule is a backpropagation rule. More preferably, the number of neurons in the hidden layer of the neural network model is twice or more the number of input feature parameters.

[0027] In this embodiment, as Figure 2 As shown, a fully connected neural network based on a multilayer perceptron was constructed. The input layer of the neural network contains 9 neurons, and there are 2 hidden layers (hidden layer 1 and hidden layer 2). Hidden layer 1 contains 17 neurons, hidden layer 2 contains 8 neurons, and the output layer contains 1 neuron. The input vector is constructed using population data, grid type, area, straight-line distance, traffic distance, traffic duration, housing price index, and production value index of the demand unit as input parameters, and the number of visits is used as the output parameter. In this embodiment, 3000 sample data were generated through a simulator, and the data was normalized and processed before being used as samples. 35% of these samples were randomly selected as training samples for the neural network model. The convergence condition of the neural network model was set to 0.0001, the learning rate to 0.2, and the momentum to 0.7. The neural network model was then trained. Figure 3The training error graph shown indicates that the neural network model converged after more than 500 iterations. After the neural network model was trained, the remaining 65% was used as analysis samples. The trained neural network model was then used to analyze these samples, and the total mean square error was 1.95E-4. This shows that the neural network model in this embodiment has a very small deviation between its predictions and the actual values, achieving a high level of fitting accuracy and reliable generalization ability. In other words, the trained neural network model in this embodiment can realistically and accurately reflect the impact of various factors on urban ecosystem service flows and precisely analyze the state of urban ecosystem service flows.

[0028] In this embodiment, a supply analysis module is also included, used to construct a supply transfer function based on the number of visits and the supply quantity of the supply unit, determine the potential supply quantity of the demand unit based on the supply transfer function, and evaluate the supply and demand status of the demand unit based on the potential supply quantity and the number of visits. In this embodiment, the transfer function is as follows: As shown, where, The potential supply of demand units. The supply quantity for the supply unit. This is the ratio of the number of visits to the demand unit to the actual supply of the corresponding supply unit. The potential supply of the supply unit to the demand unit can be calculated using this transfer function. Furthermore, the supply-demand matching status of the demand unit is determined by comparing the number of visits to the demand unit (i.e., the potential demand of the demand unit to the supply unit) obtained from the neural network model with the potential supply, as shown in equation (1): (1) in, This refers to the potential supply from the supply unit to the demand unit. This represents the potential demand from the supply unit to the demand unit. and This is a preset threshold. Wherein, The preferred value is 0.8. The preferred version is 1.2. After determining the supply and demand matching status of the demand units, different colors are assigned to each demand unit through the geographic information module, so as to intuitively display the supply and demand matching status of each demand unit in the geographic information system.

[0029] It should be noted that this embodiment uses only one park green space as the supply unit for research and analysis. When there are multiple parks green spaces in the area under study, each park green space can be used as a supply unit first. Service flow analysis can be performed through a neural network model to determine the supply and demand situation of each demand unit for each park green space. Then, the summation can be performed to obtain the total supply and demand situation and supply and demand matching status of the demand units.

[0030] The urban ecosystem service flow analysis method based on neural networks in this embodiment is a specific implementation method of the aforementioned service flow analysis model, such as... Figure 4 As shown, the specific steps include: Step S1. Marking geographic grid units in the geographic information system, whereby geographic grid units include supply units and demand units; and establishing a mapping relationship between target data and geographic grid units; Step S2. Preprocessing the source data to obtain target data for each geographic grid unit; the target data includes population data, economic data, geospatial data, and access data; Step S3. Constructing a neural network model, training the neural network model using training samples constructed based on the target data, and analyzing ecosystem service flows through the trained neural network model.

[0031] In this embodiment, the neural network model uses population data, economic data, and geospatial data to construct vectors as the input layer, and access volume data as the output layer. The neural network contains at least one hidden layer. The input function of the neural network is a weighted summation function, the transfer function is a sigmoid function, and the learning rule is the backpropagation rule.

[0032] In this embodiment, step S4 is also included: constructing a supply transfer function based on the number of visits and the supply quantity of the supply unit; determining the potential supply quantity of the demand unit based on the supply transfer function; and evaluating the supply and demand status of the demand unit based on the potential supply quantity and the number of visits.

[0033] In this embodiment, economic data includes housing price indicators and production value indicators, and geospatial data includes area, grid type, straight-line distance, traffic distance, and traffic duration; the geographic grid unit is divided according to geographic boundaries; the area of ​​the geographic grid unit is less than a preset area threshold.

[0034] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention should fall within the protection scope of the present invention.

Claims

1. A neural network-based urban ecosystem service flow analysis system, characterized in that: It includes a geographic information module, a data processing module, and a neural network model; The geographic information module is used to calibrate geographic grid units and establish a mapping relationship between target data and geographic grid units; the geographic grid unit includes supply units and demand units; The data processing module is used to preprocess the source data to obtain the target data for each geographic grid unit; the target data includes population data, economic data, geospatial data, and access data. The neural network model is used to train on training samples constructed from the target data, and the trained neural network model is used to analyze ecosystem service flows.

2. The urban ecosystem service flow analysis system based on neural networks according to claim 1, characterized in that: The neural network model uses the population data, economic data, and geospatial data to construct vectors as the input layer, and the access volume data as the output layer.

3. The urban ecosystem service flow analysis system based on neural networks according to claim 2, characterized in that: The neural network contains at least one hidden layer. The input function of the neural network is a weighted summation function, the transfer function is a sigmoid function, and the learning rule is a backpropagation rule.

4. The urban ecosystem service flow analysis system based on neural networks according to claim 3, characterized in that: It also includes a supply analysis module, which is used to construct a supply transfer function based on the number of visits and the supply quantity of the supply unit, determine the potential supply quantity of the demand unit based on the supply transfer function, and evaluate the supply and demand status of the demand unit based on the potential supply quantity and the number of visits.

5. The urban ecosystem service flow analysis system based on neural networks according to any one of claims 1 to 4, characterized in that: The economic data includes housing price indicators and production value indicators; the geospatial data includes area, grid type, straight-line distance, travel distance, and travel time; the geographic grid units are divided according to geographic boundaries; the area of ​​the geographic grid unit is less than a preset area threshold.

6. A method for analyzing urban ecosystem service flows based on neural networks, characterized in that: The steps include: Step S1. In the geographic information system, geographic grid cells are labeled, wherein the geographic grid cells include supply cells and demand cells; and a mapping relationship between target data and the geographic grid cells is established. Step S2. Preprocess the source data to obtain the target data for each geographic grid cell; the target data includes population data, economic data, geospatial data, and access data; Step S3. Construct a neural network model, train the neural network model by constructing training samples based on the target data, and analyze the ecosystem service flow through the trained neural network model.

7. The method for analyzing urban ecosystem service flows based on neural networks according to claim 6, characterized in that: The neural network model uses the population data, economic data, and geospatial data to construct vectors as the input layer, and the access volume data as the output layer.

8. The method for analyzing urban ecosystem service flows based on neural networks according to claim 7, characterized in that: The neural network contains at least one hidden layer. The input function of the neural network is a weighted summation function, the transfer function is a sigmoid function, and the learning rule is a backpropagation rule.

9. The method for analyzing urban ecosystem service flows based on neural networks according to claim 8, characterized in that: It also includes step S4, which involves constructing a supply transfer function based on the number of visits and the supply quantity of the supply unit, determining the potential supply quantity of the demand unit based on the supply transfer function, and evaluating the supply and demand status of the demand unit based on the potential supply quantity and the number of visits.

10. The method for analyzing urban ecosystem service flows based on neural networks according to any one of claims 6 to 9, characterized in that: The economic data includes housing price indicators and production value indicators; the geospatial data includes area, grid type, straight-line distance, travel distance, and travel time; the geographic grid units are divided according to geographic boundaries; the area of ​​the geographic grid unit is less than a preset area threshold.

Images