A method and system for identifying urban creative districts based on an ST-DBSCAN spatio-temporal clustering algorithm

Abstract

The application provides a kind of city creative area identification method and system based on ST-DBSCAN space-time clustering algorithm, comprising: step 1, collecting target enterprise data of target city and preprocessing;Step 2, generate city creative enterprise electronic map based on the enterprise address of the target enterprise;Step 3, on the city creative enterprise electronic map, the target enterprise is clustered and partitioned for the first time based on the DBSCAN algorithm of spatial distance, and the DBSCAN candidate area is preliminarily identified;Step 4, based on the target enterprise data, the target enterprise calculation index is screened, the weight of each calculation index of the target enterprise is determined by entropy method, weighted summation, and the enterprise creativity is calculated;Step 5, in the DBSCAN candidate area, the kernel density is calculated with the enterprise creativity as the weight, and the density partition is formed;different grades of creative areas are divided according to the different density levels.The application provides a kind of city creative area identification and method and system based on DBSCAN-KDE joint clustering algorithm, to realize the accurate identification of city creative area.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of big data analysis, and particularly relates to a city creative district identification method and system based on an ST-DBSCAN spatiotemporal clustering algorithm. BACKGROUND

[0002] Current city creative district identification methods mainly rely on traditional clustering algorithms (such as K-means and hierarchical clustering) and single-index evaluation systems. However, the existing technology has limitations: first, traditional clustering algorithms are sensitive to noise and cannot effectively identify clusters with uneven density (such as core areas and edge areas), resulting in rough partitioning results. Second, existing methods use single indicators (such as the number of patents) to evaluate city functional district distribution, ignoring the synergistic effect of multi-dimensional attributes (such as employee size and revenue generation capacity), resulting in one-sided and lack of objectivity. Third, traditional kernel density estimation (KDE) does not combine enterprise attribute weights and relies only on spatial point density to divide regions, making it difficult to distinguish the clustering effect of high-innovation enterprises and unable to accurately identify core creative districts. Fourth, existing question-and-answer systems are mostly based on general knowledge bases and lack deep integration with professional field data (such as enterprise attributes and creative district characteristics), making it difficult to provide targeted decision-making suggestions. SUMMARY

[0003] To address the above technical conditions and limitations, the present application provides a city creative district identification method and system based on a DBSCAN-KDE joint clustering algorithm, to achieve accurate identification of city creative districts and automatically identify user question types through intelligent question-and-answer functions, providing convenient information acquisition and decision support.

[0004] To solve the above technical problems, the present application provides the following technical solutions:

[0005] A city creative district identification method based on an ST-DBSCAN spatiotemporal clustering algorithm, comprising the following steps:

[0006] Step 1: Collect target enterprise data of a target city and perform preprocessing;

[0007] Step 2: Generate a city creative enterprise electronic map based on the enterprise addresses of the target enterprises;

[0008] Step 3: On the city creative enterprise electronic map, perform a first clustering partitioning of the target enterprises based on the DBSCAN algorithm of spatial distance, and preliminarily identify DBSCAN candidate areas;

[0009] Step 4: Calculate indicators of target enterprises based on the target enterprise data, determine the weights of each calculation indicator of the target enterprises through the entropy method, and calculate the enterprise creativity through weighted summation;

[0010] Step 5, in the DBSCAN candidate area, the kernel density is calculated by taking the enterprise creativity as the weight to form the density partition; and different creativity areas of different grades are divided according to the different density grades.

[0011] Further, the target enterprise in the step 1 is a cultural, sports and entertainment enterprise.

[0012] The target enterprise data includes: enterprise name, registered address, industry classification, operating income, number of employees, company size.

[0013] Further, the step 2 includes the following sub-steps:

[0014] Step 2.1, obtaining the enterprise address information of the target enterprise, and converting the enterprise address information into enterprise geographic coordinates;

[0015] Step 2.2, importing the city basic map in the GIS platform, and associating the target enterprise data to the belonging area according to the enterprise geographic coordinates through spatial connection to generate the city creative enterprise electronic map.

[0016] Further, in the step 3, based on the city creative enterprise electronic map, the DBSCAN candidate area is preliminarily identified by setting the field radius and the minimum K value parameters.

[0017] Further, the step 4 includes the following sub-steps:

[0018] Step 4.1, based on the target enterprise quantity and the target enterprise calculation index, an original data matrix is constructed, including positive and negative indexes, and the target enterprise calculation index is standardized and normalized;

[0019] Step 4.2, the information entropy and weight of each target enterprise calculation index are calculated, and the enterprise creativity is determined by weighted summation;

[0020] Step 4.3, the enterprise creativity is imported into the city creative enterprise electronic map.

[0021] Further, the target enterprise calculation index includes: registered capital, operating income, number of employees.

[0022] Further, the information entropy and weight of the target enterprise calculation index are respectively:

[0023]

[0024]

[0025] Among them, , Reflecting the dispersion degree and importance of the first calculation index, the target enterprise the information entropy of the target enterprise the target enterprise the target enterprise the target enterprise the weight of the target enterprise the information utility value of the target enterprise k the target enterprise the target enterprise

[0026] the target enterprise the creativity of the target enterprise

[0027] the target enterprise

[0028] Further, the step 5 comprises:

[0029] Step 5.1, calculate the kernel density in the DBSCAN candidate area according to the creativity, output the kernel density result and optimize it through cross-validation method;

[0030] Step 5.2, divide the kernel density result into high, medium and low value areas, respectively corresponding to the core creativity area, the general creativity area and the potential creativity area;

[0031] Step 5.3, convert the corresponding core creativity area into a polygon.

[0032] Further, it further comprises:

[0033] Step 6, collect relevant knowledge documents, and combine the results of step 1 and step 4, upload document data based on Coze platform, and automatically build an intelligent agent question and answer device;

[0034] Step 7, based on Coze, build the workflow of the question and answer device, respond to the user's request for asking questions, identify the user's question information, get the problem type, and the problem type includes one of the enterprise information class, the creativity area information and the feature class, the creativity area development countermeasure class problem, based on the problem type, select the corresponding workflow path for execution;

[0035] Step 8, install the running environment and dependencies on the sealos cloud server, configure network connection and security group rules, receive the secret tower AI calling request from the question and answer device, and return the result to the user after integrating the secret tower AI returned result and the knowledge base calling result through the internal large model of the question and answer device.

[0036] ​​In another aspect, the present application also provides a city creative district identification system based on the ST-DBSCAN spatio-temporal clustering algorithm, comprising:

[0037] A data collection module is configured to collect target enterprise data of a target city and perform preprocessing;

[0038] An electronic map generation module is configured to generate a city creative enterprise electronic map based on enterprise addresses of the target enterprises;

[0039] A DBSCAN candidate area identification module is configured to perform a first clustering partition on the target enterprises based on a DBSCAN algorithm of spatial distance on the city creative enterprise electronic map, and preliminarily identify a DBSCAN candidate area;

[0040] An enterprise creativity calculation module is configured to calculate indexes of target enterprises based on the target enterprise data, determine weights of each calculation index of the target enterprises by an entropy method, and calculate enterprise creativity by weighted summation;

[0041] A creative district grade division module is configured to calculate kernel density by taking enterprise creativity as a weight in the DBSCAN candidate area, form a density partition, and divide creative districts of different grades according to different density grades.

[0042] Compared with the prior art, the present application has the following beneficial effects:

[0043] The present application uses DBSCAN and k-nearest neighbor algorithm joint clustering to identify high-density areas through neighborhood radius (ε) and minimum sample size (MinPts), effectively process noise data and find clusters of arbitrary shape, and solve the problem of rough partition results of traditional algorithms. In addition, by objectively calculating the weights of enterprise attributes (registered capital, revenue, number of employees, number of patents, etc.) through information entropy, subjective weighting bias is eliminated, multi-dimensional features of enterprise creativity are comprehensively reflected, and the scientificity of creative district evaluation is improved. In addition, combined with the DBSCAN candidate area result, kernel density analysis is performed by enterprise creativity, the spatial distribution and attribute characteristics are deeply integrated, and creative districts of different degrees can be accurately divided. Finally, combined with the knowledge base and large model technology, automatic identification of user problem types (enterprise information, creative district characteristics, development countermeasures, etc.) is realized, accurate answers are generated, and dynamic decision support is provided. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0045] Figure 1 The figure is a general flowchart of the present application.

[0046] Figure 2 The figure is a distribution diagram of enterprises in a certain city in an embodiment of the present application.

[0047] Figure 3 The figure is a DBSCAN initial clustering diagram in an embodiment of the present application.

[0048] Figure 4 The figure is a city creative space recognition and classification diagram in an embodiment of the present application.

[0049] Figure 5 The figure is a question and answer device workflow diagram in an embodiment of the present application. DETAILED DESCRIPTION

[0050] To make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0051] Embodiment 1

[0052] The present application will be further described below in combination with the drawings.

[0053] As shown in the figure, the present application provides a city creative district recognition method based on big data and an intelligent agent question and answer device, including the following steps: Figure 1 Step 1, collect target enterprise data of a target city and perform preprocessing; collect “culture, sports and entertainment industry” enterprise data of the target city from a public enterprise information platform (Qichacha), and the fields include: enterprise name, registered address, industry classification, operating income, number of employees, company size, etc. The specific implementation is divided into the following sub-steps,

[0054] Step 1.1, use the Scrapy framework of Python automatic tool to crawl data based on Qichacha API (interface permission needs to be applied) or simulate a browser;

[0055] Step 1.2, combine Baidu map address resolution API ( / geocoding / v3) and regular expression matching to convert unstructured addresses into four-level administrative division labels;

[0056] It should be noted that field cleaning may be needed in step 1 to filter invalid data (such as records with empty operating income);

[0057] It should be noted that field cleaning may be needed in step 1 to filter invalid data (such as records with empty operating income);

[0058] Step 2, generate a city creative enterprise electronic map based on the enterprise address of the target enterprise, as shown in Figure 2 The specific implementation sub-steps are as follows:

[0059] Step 2.1, use Baidu Map API to obtain the geographic coding of the enterprise address, and convert the address information into enterprise geographic coordinates; in the specific embodiment, Baidu Map API batch geographic coding (quota application is required) is used to return longitude and latitude coordinates (lng, lat);

[0060] Step 2.2, import the city basic map (including administrative division and road network) in the GIS platform (ArcGIS), and associate the enterprise point data (including longitude and latitude coordinates) to the corresponding area through spatial connection (Spatial Join).

[0061] Step 3, on the city creative enterprise electronic map, perform the first clustering partition on the target enterprise based on the DBSCAN algorithm of spatial distance, and preliminarily identify the DBSCAN candidate area; as shown in Figure 3 ;

[0062] The specific implementation is as follows:

[0063] Step 3.1, open the Python environment of ArcGIS Pro (through “Analysis” → “Python” → “Python window”), install scikit-learn and necessary dependencies;

[0064] Step 3.2, read the enterprise point data in the Python window, and set the neighborhood parameters (ε, MinPts);

[0065] Here, ε describes the neighborhood distance threshold of a sample, and MinPts describes the threshold of the number of samples in the neighborhood of a sample with a distance of ε, that is, the K value;

[0066] Step 4, calculate the target enterprise computing indicators based on the target enterprise data, determine the weight of each computing indicator of the target enterprise through the entropy method, and calculate the enterprise creativity through weighted summation;

[0067] The specific implementation is as follows:

[0068] Step 4.1, determine the number of enterprises , specific computing indicators (such as registered capital, operating income, and number of employees, which can be selected), construct the original data matrix, which is a positive and negative indicator, and perform standardization and normalization processing, and finally import the data into ArcGIS to form a new field “enterprise creativity”;

[0069] Positive index standardization:

[0070] Negative indicator standardization:

[0071] Where, represents all values of the column indicator.

[0072] Normalization processing:

[0073] Step 4.2, calculate information entropy and weight , determine the enterprise creativity ;

[0074] The information entropy of the th indicator is: The weight of the

[0075] th indicator is:

[0076] The creativity of the enterprise is the weighted sum of each indicator:

[0077] Step 4.3, import the enterprise creativity into the city creative enterprise electronic map.

[0078] Step 5, within the DBSCAN candidate area, calculate the kernel density with the enterprise creativity as the weight to form a density partition; according to the different density levels, different levels of creative areas are divided. As shown in Figure 4 ;

[0079] The specific implementation steps are as follows:

[0080] Step 5.1, open ArcGIS, open the "spatial analysis" toolbox → "density analysis" → "kernel density analysis", parameter setting, input point features: enterprise points → field: Creativity (creativity as weight) → output raster: set path (such as KernelDensity.tif) → search radius (bandwidth): adjust according to data distribution (such as 500 meters), which can be optimized through cross-validation;

[0081] The cross-validation method specifically adopts least squares cross-validation (LSCV), which does not require distribution assumption and is suitable for complex big data, and the formula is: ;

[0082] Through traversing the candidate bandwidth , the optimal bandwidth ​​​, calculate the leave-one-out density estimate, select the minimum LSCV .

[0083] Step 5.2, use the "reclassification" tool (Reclassify) to divide the kernel density result into high, medium and low value areas. The distribution corresponds to the core creative area, the general creative area and the potential creative area;

[0084] Step 5.3, use the "raster to polygon" tool (Raster to Polygon) to convert the high-density area into a polygon;

[0085] As shown in Figure 5 ; Step 6, collect relevant knowledge documents, and combine the results of steps 1 and 4, upload document data based on the Coze platform, and automatically build an intelligent agent question and answer device;

[0086] Step 7, based on Coze, build the workflow of the question and answer device, respond to user query requests, identify user query information, get the problem type, and the problem type includes one of enterprise information class, creative area information and feature class, and creative area development countermeasure class problem, based on the problem type, select the corresponding work flow path to execute;

[0087] Step 8, install the running environment and dependencies on the sealos cloud server, configure the network connection and security group rules, receive the secret tower AI calling request from the question and answer device, and return the result to the user after the secret tower AI returns the result and the knowledge base calling result are integrated through the internal large model of the question and answer device.

[0088] Embodiment 2

[0089] The embodiment provides a city creative area identification system based on an ST-DBSCAN spatiotemporal clustering algorithm, characterized in that it comprises:

[0090] A data acquisition module is configured to acquire target enterprise data of a target city and perform preprocessing;

[0091] An electronic map generation module is configured to generate a city creative enterprise electronic map based on enterprise addresses of the target enterprises;

[0092] A DBSCAN candidate area identification module is configured to perform a first clustering partition on the target enterprises based on a DBSCAN algorithm of spatial distance on the city creative enterprise electronic map, and preliminarily identify a DBSCAN candidate area;

[0093] An enterprise creativity calculation module is configured to calculate indexes of target enterprises based on the target enterprise data, determine weights of the target enterprise indexes by an entropy method, perform weighted summation, and calculate enterprise creativity;

[0094] The creative area grading module is used to calculate the core density in the DBSCAN candidate area, form density partitions, and divide the creative areas into different grades according to the different density grades.

[0095] It should be understood that the parts not elaborated in the specification are all prior art.

[0096] It should be understood that the above description of the preferred embodiments is more detailed and is not considered as a limitation on the scope of patent protection of the present application. It is not necessary or possible to enumerate all the embodiments here. Those skilled in the art can make substitutions or modifications without departing from the scope of the claims of the present application. The scope of protection of the present application should be subject to the appended claims.

Claims

1. A method for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm, characterized in that, Includes the following steps: Step 1: Collect data on target companies in the target city and perform preprocessing; Step 2: Generate an electronic map of creative enterprises in the city based on the enterprise address of the target enterprise; Step 3: On the electronic map of creative enterprises in the city, the DBSCAN algorithm based on spatial distance is used to perform the first clustering and partitioning of the target enterprises to initially identify DBSCAN candidate areas; Step 4: Based on the target company data, filter the target company's calculated indicators, determine the weight of each calculated indicator of the target company using the entropy method, and calculate the company's creativity score by weighted summation; including the following sub-steps: Step 4.1: Based on the number of target enterprises and the calculated indicators of target enterprises, construct an original data matrix, including positive and negative indicators, and perform standardization and normalization processing on the calculated indicators of target enterprises; Step 4.2: Calculate the information entropy and weights of the calculated indicators for each target enterprise, and perform a weighted sum to determine the enterprise's creativity level; the information entropy and weights of the calculated indicators for each target enterprise are as follows: in, , Reflecting the The dispersion and importance of each calculated indicator For the target company The information entropy of the calculated index, The target number of companies To standardize and normalize the target enterprise's calculated indicators; Target Company The weight of each calculation indicator, For the first k The information utility value of each calculated indicator Calculate the number of indicators for enterprises. Target companies Creativity for: in, Calculate metrics for the target companies undergoing standardized processing; Step 4.3: Import the enterprise creativity score into the city's electronic map of creative enterprises; Step 5: Within the DBSCAN candidate area, calculate the kernel density using enterprise creativity as the weight to form density partitions; divide the creative areas into different levels according to the density level; including: Step 5.1: Within the DBSCAN candidate region, calculate the kernel density based on creativity, output the kernel density results, and optimize them using cross-validation. Step 5.2: Divide the kernel density results into high, medium and low value areas, which correspond to the core creative area, general creative area and potential creative area, respectively; Step 5.3: Convert the corresponding core creative area into a polygon.

2. The method for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm according to claim 1, characterized in that, The target enterprise in step 1 is a cultural, sports, and entertainment enterprise; The target company data includes: company name, registered address, industry classification, operating revenue, number of employees, and company size.

3. The method for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm according to claim 1, characterized in that, Step 2 includes the following sub-steps: Step 2.1: Obtain the enterprise address information of the target enterprise and convert the enterprise address information into enterprise geographic coordinates; Step 2.2: Import the city base map into the GIS platform, and use spatial connections to associate the target enterprise data with the region to which it belongs based on the enterprise's geographical coordinates, thereby generating an electronic map of creative enterprises in the city.

4. The method for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm according to claim 1, characterized in that, In step 3, based on the electronic map of creative enterprises in the city, candidate areas of DBSCAN are initially identified by setting the domain radius and minimum K value parameters.

5. The method for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm according to claim 1, characterized in that, The target company's calculation indicators include: registered capital, operating revenue, and number of employees.

6. The method for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm according to claim 1, characterized in that, Also includes: Step 6: Collect relevant knowledge documents and combine the results of Step 1 and Step 4. Upload the document data to the Coze platform to automatically construct an intelligent agent question-answering device. Step 7: Based on Coze, construct the workflow of the question-and-answer device, respond to user question requests, identify user question information, and obtain the question type. The question type includes one of the following: enterprise information, creative zone information and characteristics, and creative zone development strategy questions. Based on the question type, select the corresponding workflow path to execute. Step 8: Install the runtime environment and dependencies on the Sealos Cloud server, configure network connection and security group rules, receive the Secret AI call request from the question-answering device, and return the result returned by Secret AI and the knowledge base call result to the user through the large model inside the question-answering device.

7. A system for identifying urban creative zones based on the ST-DBSCAN spatiotemporal clustering algorithm, characterized in that, include: The data acquisition module is used to collect and preprocess data from target companies in the target city. An electronic map generation module is used to generate a city-based creative enterprise electronic map based on the enterprise address of the target enterprise. The DBSCAN candidate area identification module is used to perform the first clustering and partitioning of the target enterprises on the electronic map of creative enterprises in the city based on the DBSCAN algorithm of spatial distance, and to initially identify DBSCAN candidate areas. The enterprise creativity calculation module is used to filter the target enterprise calculation indicators based on the target enterprise data, determine the weight of each calculation indicator of the target enterprise through the entropy method, and calculate the enterprise creativity by weighted summation. The Creative Zone Classification Module is used to calculate kernel density within the DBSCAN candidate area, with enterprise creativity as the weight, to form density partitions; and to divide creative zones into different levels according to different density levels. The urban creative zone identification system based on the ST-DBSCAN spatiotemporal clustering algorithm is used to perform the steps in the urban creative zone identification method based on the ST-DBSCAN spatiotemporal clustering algorithm as described in any one of claims 1-6.

Images