Park intelligent room state diagram generation system and method based on multi-source data fusion

Abstract

The invention discloses a park intelligent room state diagram generation system and method based on multi-source data fusion. The system comprises a data fusion module, a demand prediction module and a visualization module. Enterprise multi-source data is obtained and preprocessed, labeling is carried out, and a label data source is generated; the demand prediction module obtains a label data source, inputs the label data source into an area demand model in combination with enterprise historical expansion data and an industry average growth rate, and outputs an enterprise demand area; upstream and downstream relationships among enterprises are obtained, the enterprises, suppliers and customers are associated in a triple form, and an investment attraction recommendation list is generated; and the visualization module superposes the label data source of the enterprise and the house resource state information, constructs a visual house resource state diagram, renders the visual house resource state diagram, superposes the visual house resource state diagram in layers, generates a thermodynamic diagram, obtains the risk degree and the number of vacant days, and realizes investment attraction and efficient management. According to the method, resource allocation is optimized, and collaborative development of the park industry is promoted.

Description

Technical Field

[0001] The present invention relates to the field of intelligent decision-making technology for smart parks, and in particular to a system and method for generating an intelligent room status map for a park based on multi-source data fusion. Background Art

[0002] With intensifying industrial competition in modern industrial parks, they need to accurately grasp business needs and quickly match housing resources with enterprises to enhance their overall competitiveness. However, traditional industrial park management models struggle to meet these demands. On the one hand, business development dynamics and market conditions are constantly changing, and demand for housing is shifting accordingly. Traditional management methods are unable to timely capture and respond to these changes. On the other hand, the industrial park is home to numerous enterprises and complex industrial chain relationships. Failure to effectively organize and leverage these relationships can lead to a lack of targeted investment promotion efforts and a significant waste of resources.

[0003] Traditional park property management relies primarily on manual record-keeping and simple property registration systems. Staff manually record basic property information, such as building, floor, room number, and lease period, while also performing simple registration of information about the businesses that move in. Under this model, new property information or changes to business information require manual updates, which is inefficient and prone to errors. Furthermore, manually recorded data is limited to basic surface information, preventing in-depth analysis of a business's potential needs and industry chain connections. Data updates are severely delayed, failing to reflect the true status of properties in real time. Information is isolated, lacking integrated analysis of multi-source data, making it difficult to provide strong support for investment decisions. Management efficiency is low, significant manpower and material resources are consumed, and inaccurate information is prone to occur, severely restricting the park's development and investment promotion efforts.

[0004] Existing technologies have improved traditional park management methods to a certain extent. By introducing information management systems, digital storage and simple visual display of housing information have been realized. However, existing technologies collect less data and fail to deeply integrate them. The integration and analysis of corporate public opinion data with industrial and commercial data and housing data are insufficient, and it is impossible to comprehensively assess corporate risks and needs. There is also a lack of predictions on future development trends of enterprises, making it difficult to achieve accurate investment promotion and adapt to the diverse needs of parks of different sizes and types, and it is unable to effectively support the emerging business of industrial chain investment promotion.

[0005] In summary, the park management methods based on existing and traditional technologies are difficult to meet the needs of the current environment and park investment promotion. Summary of the Invention

[0006] Based on the above content, the present application discloses a system for generating intelligent housing status maps in a park based on multi-source data fusion, including a data fusion module, a demand prediction module, and a visualization module;

[0007] The data fusion module includes a data source acquisition module and a data processing module; the data source acquisition module obtains enterprise multi-source data and inputs it into the data processing module; the data processing module receives enterprise multi-source data for pre-processing, labels it, and generates a label data source;

[0008] The demand forecasting module includes a forecasting submodule and an industry chain analysis submodule. The forecasting submodule obtains a label data source, combines the company's historical expansion data and the industry's average growth rate into an area demand model, and outputs the company's required area. The industry chain analysis submodule obtains the upstream and downstream relationships between companies, associates companies, suppliers, and customers in the form of triples, and generates a list of recommended investment invitations.

[0009] The visualization module includes a room status map generation submodule and a decision execution module; the room status map generation submodule constructs a visual room status map covering the enterprise's labels by superimposing the enterprise's label data source with the room status information; the decision execution module generates a heat map by rendering and overlaying the visual room status map, obtains the risk level and vacancy days, and realizes investment promotion and efficient management.

[0010] Preferably, the enterprise multi-source data includes industrial and commercial data, public opinion data and housing data of enterprises in the park; industrial and commercial data includes registered capital, business status and industry classification; public opinion data includes legal risks, business anomalies, bidding information and intellectual property rights; housing data includes the enterprise's building, floor, room number, lease period and historical tenants.

[0011] Preferably, the data processing module removes duplicate values ​​from the enterprise's industrial and commercial data, public opinion data and housing database through a data cleaning sub-module, and performs address standardization processing; the data processing module also includes a credit portrait generation sub-module, which obtains the enterprise's comprehensive enterprise data and public opinion risk, and uses the credit portrait generation sub-module to label the enterprise, calculates the public opinion risk score based on legal disputes, business anomalies and insurance fluctuations, and combines other business data of the enterprise to generate high / low potential or high / low risk enterprise labels.

[0012] Preferably, the area demand model is an LSTM network model, which takes the acquired label data source, the enterprise's historical expansion data and the industry average growth rate as input, and outputs the enterprise's required area. The formula is: y = FC (LSTM ([x z ,x c ,x b ,x s ,r,k])), where the historical expansion data of the enterprise includes the number of branches x z , Growth rate of insured persons x c 、Increase in the number of winning bids x b 、Increase in the number of intellectual property rights x s, r is the industry average growth rate, k is the risk coefficient, multiple iterative calculations are performed, mapping conversion is performed through the fully connected layer, and the predicted value of the enterprise office area demand is output.

[0013] Preferably, the formation of the investment recommendation list is specifically as follows:

[0014] Through the industrial chain analysis submodule, the upstream and downstream relationships between enterprises in the park are obtained from the multi-source data of enterprises, and enterprises, suppliers and customers are associated in the form of triplets to build an enterprise knowledge graph, identify the enterprise types corresponding to the missing or weak links in the industrial chain, and screen out qualified potential enterprises. At the same time, the enterprise knowledge graph is used to evaluate the industry status, market influence and technological innovation indicators of enterprises to obtain the leading enterprises in the industry. The industrial chain analysis submodule follows the priority principle of giving priority to enterprises with industrial chain gaps over leading enterprises in the same industry. It internally ranks the selected enterprises with industrial chain gaps according to the importance of improving the industrial chain and the urgency of filling the gaps, and puts them at the top of the investment recommendation list. It also ranks the leading enterprises in the same industry according to their comprehensive strength and potential value to the park industry, and lists them after the enterprises with industrial chain gaps to generate an investment recommendation list.

[0015] Preferably, the housing status map generation submodule establishes a mapping relationship M by integrating the enterprise's label data source and housing status information. ik , construct a visual housing status diagram, the formula is: where h ij is the jth attribute of property i, l mk is the kth label of enterprise m, w j and v n are the weight coefficients of property attributes and enterprise tags, respectively. a and b are the number of property attributes and enterprise tags, respectively. By integrating the property and enterprise tag information, M is calculated. ik , for each listing i, according to M ik The values ​​and set rules are visualized.

[0016] Preferably, the visual property status diagram supports a multi-level search function, searches property information through multiple dimensions, locates and displays the status of properties in the park, including whether they are vacant or rented, and the corresponding corporate tag information. Based on the property status information, vacant properties are screened out and presented on the visual property status diagram. At the same time, it supports combined search based on industry and risk tags, and screens out qualified properties in the visual diagram with superimposed corporate tags and property status.

[0017] Preferably, the decision execution module extracts and analyzes data from the visualized housing status diagram to obtain the correlation information between the enterprise and the housing, including the enterprise risk label and the contract lease period. Through the risk assessment system, the housing corresponding to the labeled enterprise is taken as the key assessment object, and the legal risk and operation abnormality information in the public opinion data are combined to perform quantitative scoring to determine the risk level of the housing; the current time of the housing and the end-of-term lease expiration time are obtained to obtain the number of vacant days. With the basic housing status layer as the bottom layer, the information representing the risk level and the number of vacant days are superimposed as different layers to output the park housing heat map.

[0018] Preferably, the decision execution module also includes an early warning module, which monitors and warns through contract expiration and enterprise risk indicators; for contract expiration warning, the early warning module continuously obtains housing rental information, compares and calculates the contract expiration date with the current date, and sets a contract date threshold. If the contract expiration date and the current date are less than the contract date threshold, the early warning is triggered, and the corresponding housing is prompted in the visual housing status diagram; for enterprise risk warning, the early warning module obtains integrated public opinion data and enterprise credit portraits, and dynamically monitors the enterprise risk status. When the risk score exceeds the set risk threshold, it is determined to be a high-risk enterprise, and the early warning module pushes risk control suggestions, and at the same time, special marks are made on the housing rented by high-risk enterprises.

[0019] A method for generating a smart housing state map in a park based on multi-source data fusion includes the following steps:

[0020] S1. Obtain the industrial and commercial data, public opinion data, and housing data of enterprises in the park for pre-processing, calculate the risk score, and generate a label data source;

[0021] S2. Obtain the label data source, combine the company's historical expansion data and the industry's average growth rate, input the area demand model, and output the company's office space demand forecast;

[0022] S3. Obtain the relationships between enterprises, suppliers, and customers, and construct an enterprise knowledge graph using triples to identify potential entrants and leading enterprises in the industry with missing or weak links in the industrial chain. Prioritize enterprises with gaps in the industrial chain over leading enterprises in the same industry, and generate a list of recommended investment invitations.

[0023] S4. Integrate the enterprise's tag data source with the park's housing status information, visualize each property, and build a visual housing status map that includes the enterprise's tags;

[0024] S5. Extract and analyze data from the visualized property status map to obtain enterprise risk tags and contract lease periods. Use the risk assessment system combined with public opinion data to determine the property risk level and calculate the number of vacancy days. Using the basic property status layer as the base layer, overlay the risk level and vacancy days information layers to generate a heat map of the park's properties.

[0025] S6. Monitor the expiration of corporate contracts and corporate risk indicators, push risk control suggestions and specially mark the properties rented by high-risk enterprises.

[0026] Compared with the prior art, the technical solution of this application has the following technical effects:

[0027] The present invention uses a prediction submodule, combined with historical enterprise expansion data, industry average growth rate, and label data sources, to accurately predict enterprise demand area using an LSTM network model. Simultaneously, the industrial chain analysis submodule constructs a knowledge graph based on the upstream and downstream relationships between enterprises, identifies enterprises with gaps in the industrial chain and leading enterprises in the industry, and generates a list of recommended investment promotions based on priority. This enables park investment promotion to accurately meet enterprise needs, prioritizing the introduction of enterprises that play an important role in improving the industrial chain. This greatly improves the accuracy of investment promotion, avoids blind investment promotion, and improves the efficiency of investment promotion resources.

[0028] In terms of property management, this invention uses a property status map generation submodule to integrate enterprise tag data sources with property status information to construct a visual property status map. This supports multi-level search, allowing for rapid location and screening of vacant properties, and also allows for combined searches based on industry and risk tags. The decision execution module analyzes the visual property status map to determine risk levels and vacancy days, generating a heat map to visually display the overall property status. These features help managers understand property dynamics in real time, quickly respond to property allocation needs, effectively reduce idle time, and improve the efficiency and refinement of property management.

[0029] The early warning module within the decision-making execution module of this invention monitors contract expiration dates and enterprise risk indicators in real time. It continuously acquires property lease information and, when the difference between the contract expiration date and the current date is less than a set threshold, displays a prompt on a visual property status diagram, facilitating advance planning for subsequent investment opportunities or lease renewals. Furthermore, by acquiring public opinion data and corporate credit profiles, it dynamically monitors corporate risks, provides risk control recommendations to high-risk enterprises, and specifically labels their rental properties, effectively reducing park operational risks and ensuring their stable development.

[0030] The industrial chain analysis submodule of the present invention associates enterprises, suppliers and customers in the form of triples to construct a knowledge graph, identifies missing or weak links in the industrial chain, and introduces enterprises with gaps in the industrial chain in a targeted manner. This promotes upstream and downstream cooperation between enterprises in the park and forms a complete closed loop of the industrial chain. For example, after the entry of new energy vehicle companies, it will drive the gathering of related companies such as battery suppliers and charging pile service providers, realize the coordinated development of the industry, enhance the overall competitiveness of the park industry, attract more high-quality companies to settle in, and promote the virtuous cycle of the park's industrial ecology.

[0031] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application so that it can be implemented in accordance with the contents of the specification, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following is a detailed description of the preferred embodiment of the present application in conjunction with the accompanying drawings.

[0032] Based on the detailed description of the specific embodiments of the present application in conjunction with the accompanying drawings below, those skilled in the art will become more aware of the above and other objects, advantages and features of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without inventive work. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0034] Figure 1 This is a structural diagram of a park intelligent room status map generation system based on multi-source data fusion according to the present invention;

[0035] Figure 2 This is a flow chart of a method for generating a smart housing state map in a park based on multi-source data fusion according to the present invention;

[0036] Figure 3 This is a comparison chart of the occupancy matching degree of the smart room status map of the park of the present invention;

[0037] Figure 4 This is a comparison chart of the average idle time and the average time consumed in searching for housing resources in the intelligent housing status map of the park of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the following description, specific details such as specific configurations and components are provided only to help fully understand the embodiments of the present application. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, for clarity and brevity, the description of known functions and structures has been omitted in the embodiments.

[0039] It should be understood that references throughout this specification to "one embodiment" or "this embodiment" mean that a particular feature, structure, or characteristic associated with the embodiment is included in at least one embodiment of the present application. Therefore, the appearance of "one embodiment" or "this embodiment" throughout this specification does not necessarily refer to the same embodiment. Furthermore, these particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0040] In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplicity and clarity and does not in itself indicate the relationship between the various embodiments and / or settings discussed.

[0041] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" in this article describes another type of association object relationship, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the previous and subsequent associated objects are in an "or" relationship.

[0042] The term "at least one" in this article is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, at least one of A and B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0043] It should also be noted that, in this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprises," or any other variations thereof are intended to cover non-exclusive inclusion.

[0044] Example 1

[0045] This embodiment mainly describes a system for generating intelligent room status maps in a park based on multi-source data fusion. Figure 1 As shown, it includes data fusion module, demand forecasting module and visualization module;

[0046] The data fusion module includes a data source acquisition module and a data processing module; the data source acquisition module obtains enterprise multi-source data and inputs it into the data processing module; the data processing module receives enterprise multi-source data for pre-processing, labels it, and generates a label data source;

[0047] The demand forecasting module includes a forecasting submodule and an industry chain analysis submodule. The forecasting submodule obtains a label data source, combines the company's historical expansion data and the industry's average growth rate into an area demand model, and outputs the company's required area. The industry chain analysis submodule obtains the upstream and downstream relationships between companies, associates companies, suppliers, and customers in the form of triples, and generates a list of recommended investment invitations.

[0048] The visualization module includes a room status map generation submodule and a decision execution module; the room status map generation submodule constructs a visual room status map covering the enterprise's labels by superimposing the enterprise's label data source with the room status information; the decision execution module generates a heat map by rendering and overlaying the visual room status map, obtains the risk level and vacancy days, and realizes investment promotion and efficient management.

[0049] Furthermore, the multi-source enterprise data includes industrial and commercial data, public opinion data and housing data of enterprises in the park; industrial and commercial data includes registered capital, business status and industry classification; public opinion data includes legal risks, business anomalies, bidding information and intellectual property rights; housing data includes the enterprise's building, floor, room number, lease period and historical tenants.

[0050] Furthermore, the data processing module removes duplicate values ​​from the enterprise's industrial and commercial data, public opinion data and housing database through the data cleaning sub-module, and performs address standardization processing; the data processing module also includes a credit portrait generation sub-module, which obtains the enterprise's comprehensive enterprise data and public opinion risk, and uses the credit portrait generation sub-module to label the enterprise, calculate the public opinion risk score based on legal disputes, business anomalies and insurance fluctuations, and combine other business data of the enterprise to generate high / low potential or high / low risk enterprise labels.

[0051] Furthermore, the area demand model is an LSTM network model, which takes the acquired label data source, enterprise historical expansion data and industry average growth rate as input and outputs the enterprise demand area. The formula is: y = FC (LSTM ([x z ,x c ,x b ,x s,r,k])), where the historical expansion data of the enterprise includes the number of branches x z , Growth rate of insured persons x c 、Increase in the number of winning bids x b 、Increase in the number of intellectual property rights x s , r is the industry average growth rate, k is the risk coefficient, multiple iterative calculations are performed, mapping conversion is performed through the fully connected layer, and the predicted value of the enterprise office area demand is output.

[0052] Furthermore, the formation of the investment recommendation list is specifically as follows:

[0053] Through the industrial chain analysis submodule, the upstream and downstream relationships between enterprises in the park are obtained from the multi-source data of enterprises, and enterprises, suppliers and customers are associated in the form of triplets to build an enterprise knowledge graph, identify the enterprise types corresponding to the missing or weak links in the industrial chain, and screen out qualified potential enterprises. At the same time, the enterprise knowledge graph is used to evaluate the industry status, market influence and technological innovation indicators of enterprises to obtain the leading enterprises in the industry. The industrial chain analysis submodule follows the priority principle of giving priority to enterprises with industrial chain gaps over leading enterprises in the same industry. It internally ranks the selected enterprises with industrial chain gaps according to the importance of improving the industrial chain and the urgency of filling the gaps, and puts them at the top of the investment recommendation list. It also ranks the leading enterprises in the same industry according to their comprehensive strength and potential value to the park industry, and lists them after the enterprises with industrial chain gaps to generate an investment recommendation list.

[0054] Furthermore, the housing status map generation submodule establishes a mapping relationship M by integrating the enterprise's label data source and housing status information. ik , construct a visual housing status diagram, the formula is: where h ij is the jth attribute of property i, l mk is the kth label of enterprise m, w j and v n are the weight coefficients of property attributes and enterprise tags, respectively. a and b are the number of property attributes and enterprise tags, respectively. By integrating the property and enterprise tag information, M is calculated. ik , for each listing i, according to M ik The values ​​and set rules are visualized.

[0055] Furthermore, the visual property status diagram supports multi-level search functions, searches property information through multiple dimensions, locates and displays the status of properties in the park, including whether they are vacant or rented, and the corresponding corporate tag information. Based on the property status information, vacant properties are screened out and presented on the visual property status diagram. At the same time, it supports combined search based on industry and risk tags, and screens out qualified properties in the visual diagram with superimposed corporate tags and property status.

[0056] Furthermore, the decision execution module extracts and analyzes data from the visualized housing status diagram to obtain the correlation information between the enterprise and the housing, including the enterprise risk label and the contract lease period. Through the risk assessment system, the housing corresponding to the labeled enterprise is taken as the key assessment object, and the legal risk and operation abnormality information in the public opinion data are combined to perform quantitative scoring to determine the risk level of the housing; the current time of the housing and the end-of-term lease expiration time are obtained to obtain the number of vacant days. With the basic housing status layer as the bottom layer, the information representing the risk level and the number of vacant days are superimposed as different layers to output the park housing heat map.

[0057] Furthermore, the decision execution module also includes an early warning module, which monitors and warns through contract expiration and enterprise risk indicators; for contract expiration warning, the early warning module continuously obtains housing rental information, compares and calculates the contract expiration date with the current date, and sets a contract date threshold. If the contract expiration date and the current date are less than the contract date threshold, the early warning is triggered, and the corresponding housing is prompted in the visual housing status diagram; for enterprise risk warning, the early warning module obtains integrated public opinion data and enterprise credit portraits, and dynamically monitors the enterprise risk status. When the risk score exceeds the set risk threshold and is determined to be a high-risk enterprise, the early warning module pushes risk control suggestions and specially marks the housing rented by high-risk enterprises.

[0058] This embodiment describes in detail how to deeply integrate corporate business data, public opinion data, and housing data through a multi-source data fusion module. This not only accurately obtains basic corporate information, but also assesses corporate risks through public opinion data and accurately labels companies, achieving deep data mining and fusion. The LSTM network model is used to predict corporate area requirements, and combined with industrial chain analysis, a recommended list of investment promotion opportunities is generated, greatly improving the accuracy and efficiency of investment promotion. In terms of visualization and management, the constructed visual housing status diagram supports multi-level retrieval and heat map display, making it convenient for managers to quickly obtain the required information. The early warning module monitors contract expiration and corporate risks in real time, and promptly pushes risk control suggestions, effectively reducing park operation risks.

[0059] Based on Example 1, this implementation describes in detail the enterprise knowledge graph identifying the enterprise types corresponding to the missing or weak links in the industrial chain, and screening out potential enterprises that meet the conditions, specifically:

[0060] The industrial chain analysis submodule starts with enterprise multi-source data, linking enterprises, suppliers, and customers in the form of triplets, and then constructing an enterprise knowledge graph. Multi-source data covers enterprise business data, public opinion data, and real estate data. With each data as support, each enterprise is regarded as a core node, and its suppliers and customers are respectively connected to it as upstream and downstream nodes, forming an "enterprise-supplier-customer" triple. Through the triples, the upstream and downstream related information of many enterprises is integrated, gradually building a complex network structure to form an enterprise knowledge graph;

[0061] Using the enterprise knowledge graph, we can identify the types of enterprises corresponding to missing or weak links in the industrial chain and screen out potential enterprises that meet the requirements. The knowledge graph presents the structure of the entire industrial chain and the distribution of enterprises. By analyzing the nodes and edges in the graph, we can find that there are fewer enterprises in certain links, the connections are not tight enough, or there is an insufficient supply of key technologies and products, which are missing or weak links in the industrial chain.

[0062] Through the enterprise knowledge graph, we evaluate the industry status, market influence and technological innovation indicators of enterprises, and then obtain the leading enterprises in the industry. The knowledge graph can evaluate the market share and business activity of enterprises in the industry from the scale and frequency of cooperation with upstream and downstream enterprises. Enterprises with large market share and frequent business often occupy an important position in the industry; the breadth and depth of connections in the enterprise knowledge graph reflect market influence; the breadth of connections means that there are cooperative relationships with many different types of enterprises, indicating that the business radiation range of the enterprise in the industry is wide; the depth of connections reflects the closeness and long-term stability of cooperation with upstream and downstream enterprises. Enterprises with deep connections are more influential in the industry; in terms of technological innovation indicators, through public opinion data and the enterprise's own intellectual property data, we can obtain the enterprise's technology R&D investment, number of patents, and application of innovative achievements, and comprehensively analyze and rank all enterprises in the knowledge graph based on various evaluation indicators. The leading enterprises are the leading enterprises in the industry.

[0063] This embodiment describes in detail how to construct an enterprise knowledge graph through the industrial chain analysis submodule and screen potential resident enterprises, acquire leading enterprises in the industry, accurately locate missing or weak links in the industrial chain, introduce adaptive enterprises to improve the industrial chain, enhance industrial synergy and competitiveness, evaluate enterprises with the help of the graph, and the acquired leading enterprises in the industry can play a leading role, attract upstream and downstream enterprises to gather, form an industrial cluster effect, optimize the overall industrial layout of the park, promote efficient allocation of resources, promote the sustainable development of the park industry, and enhance the park's attractiveness and comprehensive strength in the market.

[0064] Example 2

[0065] This embodiment describes in detail a method for generating a smart room state map based on multi-source data fusion. Figure 2 As shown, the following steps are included:

[0066] S1. Obtain the industrial and commercial data, public opinion data, and housing data of enterprises in the park for pre-processing, calculate the risk score, and generate a label data source;

[0067] S2. Obtain the label data source, combine the company's historical expansion data and the industry's average growth rate, input the area demand model, and output the company's office space demand forecast;

[0068] S3. Obtain the relationships between enterprises, suppliers, and customers, and construct an enterprise knowledge graph using triples to identify potential entrants and leading enterprises in the industry with missing or weak links in the industrial chain. Prioritize enterprises with gaps in the industrial chain over leading enterprises in the same industry, and generate a list of recommended investment invitations.

[0069] S4. Integrate the enterprise's tag data source with the park's housing status information, visualize each property, and build a visual housing status map that includes the enterprise's tags;

[0070] S5. Extract and analyze data from the visualized property status map to obtain enterprise risk tags and contract lease periods. Use the risk assessment system combined with public opinion data to determine the property risk level and calculate the number of vacancy days. Using the basic property status layer as the base layer, overlay the risk level and vacancy days information layers to generate a heat map of the park's properties.

[0071] S6. Monitor the expiration of corporate contracts and corporate risk indicators, push risk control suggestions and specially mark the properties rented by high-risk enterprises.

[0072] In S1, the data source acquisition module is used to collect the industrial and commercial data, public opinion data and housing data of enterprises in the park. The industrial and commercial data includes registered capital, operating status and industry classification; the public opinion data covers legal risks and operating anomalies; the housing data includes building and floor information; the collected multi-source enterprise data is input into the data processing module, and the data cleaning sub-module is used to remove duplicate data values ​​and standardize the addresses; the credit profile generation sub-module in the data processing module integrates enterprise data and public opinion risks, calculates the public opinion risk score based on legal disputes, operating anomalies and insurance fluctuations, and combines other business data of the enterprise to label the enterprise as a high / low potential or high / low risk enterprise to generate a labeled data source.

[0073] The industrial chain analysis submodule in S3 obtains the upstream and downstream relationships between enterprises in the park from multi-source data of enterprises, and constructs an enterprise knowledge graph in the form of enterprise, supplier, and customer triples. Through the knowledge graph, the enterprise types corresponding to the missing or weak links in the industrial chain are identified, and potential resident enterprises are screened. At the same time, the industry status, market influence, and technological innovation indicators of the enterprises are evaluated to obtain the leading enterprises in the industry. According to the principle that enterprises with gaps in the industrial chain are given priority over leading enterprises in the same industry, enterprises with gaps in the industrial chain are ranked according to the importance of improving the industrial chain and the urgency of filling the gaps, and leading enterprises in the same industry are ranked according to their comprehensive strength and potential value to the park industry, and a list of investment recommendation is generated.

[0074] This embodiment describes in detail a method for generating an intelligent housing status map for a park, improves park management efficiency, integrates multi-source data, accurately generates housing status maps, and supports multi-level retrieval, making it easy to quickly grasp the status of housing resources. The prediction module accurately estimates corporate demand, and industrial chain analysis helps with precise investment promotion, improves investment promotion efficiency and reduces vacancy rates. The early warning module monitors risks in real time, ensures stable operation of the park, optimizes resource allocation, promotes coordinated industrial development, significantly enhances the comprehensive competitiveness of the park, and realizes intelligent, efficient, and scientific park management.

[0075] Example 3

[0076] This embodiment describes the implementation process of this application in detail, specifically:

[0077] We selected a comprehensive industrial park with 34 companies, nine buildings, and a total of 436 properties as a sample. We collected multi-source data from the companies within the park, including industrial and commercial data, public opinion data, and property data. The industrial and commercial data covered registered capital, operating status, and industry classification; public opinion data included legal risks, operational anomalies, bidding information, and intellectual property rights; and property data included building, floor, room number, lease term, and historical tenant information. Table 1 shows some data from eight randomly selected companies.

[0078] Table 1 Enterprise multi-source data

[0079]

[0080]

[0081] The housing management system currently commonly used in industrial parks and based on simple data statistics and experience judgment was selected as the object of comparison with existing technologies. It mainly records basic housing information and corporate leasing status, and makes investment promotion and housing management decisions through simple manual analysis.

[0082] The data cleaning submodule is used to remove duplicate data values ​​and standardize addresses. The credit profile generation submodule calculates the public opinion risk score based on legal disputes, business anomalies, and insurance fluctuations, and generates enterprise labels in combination with other business data. Taking enterprise 001 as an example, its legal dispute weight is 0.4, its business anomaly weight is 0.3, and its insurance fluctuation weight is 0.3. The public opinion risk score is calculated as follows: (3×0.4+1×0.3+0×0.3) / (0.4+0.3+0.3)=1.5 (full score is 5 points). Combined with other business data, the final label is "medium risk, medium potential". According to this method, 10 randomly selected enterprises were processed to generate label data sources. Some of the results are shown in Table 2 below:

[0083] Table 2 Enterprise label data source

[0084] Enterprise Number Public Opinion Risk Score Corporate Label 001 1.59 Medium risk, medium potential 002 0.385 Low risk, low potential 003 0.86 Medium risk, medium potential 004 0.09 Low risk, low potential 005 1.965 High risk, high potential 006 0.92 Medium risk, medium potential 007 0.49 Low risk, medium potential 008 1.62 Medium risk, high potential

[0085] The prediction submodule obtains the label data source, combines the company's historical expansion data (number of branches, growth rate of insured persons, increase in number of successful bids, increase in number of intellectual property rights) and the industry average growth rate, inputs it into the LSTM network model, and predicts the office space demand of 8 randomly selected companies one by one. After multiple iterative calculations and mapping transformation through the fully connected layer, it predicts their office space demand in the next 6 months, as shown in Table 3 below:

[0086] Table 3 Forecast demand area

[0087]

[0088] The industrial chain analysis submodule obtains upstream and downstream relationships of enterprises from multi-source data, constructs enterprise knowledge graphs based on triples, and uses the graphs to identify missing or weak links in the industrial chain. For example, in the electronic information industry chain, there are relatively few companies in the high-end chip manufacturing sector. Potential entrants in this sector are identified as those with high-end chip manufacturing technology and production capabilities. At the same time, the industry status, market influence, and technological innovation indicators of enterprises are evaluated to identify leading enterprises in the industry. Enterprises with gaps in the industrial chain are prioritized and a list of recommended investment opportunities is generated.

[0089] The housing status map generation sub-module integrates the enterprise label data source and housing status information, establishes a mapping relationship and then visualizes it. The decision execution module analyzes the visualized housing status map, obtains the risk level and vacancy days, and generates a heat map. After processing all housing resources, it forms an overall housing heat map of the park, which intuitively displays the housing status.

[0090] The early warning module monitors contract expiration and enterprise risk indicators, and sets the contract date threshold at 3 months. Taking Enterprise 001 as an example, when the remaining lease term of the contract is 2.5 months, a prompt will be displayed on the visual property status diagram. At the same time, enterprise risks are monitored in real time. When the risk score of Enterprise 001 exceeds the set threshold of 4, risk control recommendations will be pushed and its rental properties will be specially marked.

[0091] like Figure 3 As shown, based on enterprise demand forecasts and industry chain analysis, the match between investment recommendations and actual occupancy in the comprehensive industrial park under this application's technical solution reached 85.4%, while the existing technology, which primarily relies on experience, had a match rate of only 40.2%. In terms of the number of successful investment companies, this application successfully attracted seven companies during the experimental period, six of which were in line with the park's industrial plan; the existing technology attracted two companies, only one of which met the plan.

[0092] like Figure 4 As shown, the technology of this application supports multi-level search, and it takes an average of 10.2 seconds to find a specific property, and the average idle time of vacant properties is shortened to 30.6 days. The existing technology takes an average of 72.8 seconds to find a property, and the idle time of vacant properties is as long as 112.5 days. In terms of the timeliness of property allocation, this application can complete the allocation within 8 hours after the enterprise's needs change, while the existing technology takes 42 hours.

[0093] This application conducts real-time monitoring of contract expiration and enterprise risks of some enterprises in comprehensive industrial parks, with an early warning accuracy rate of 86.5%; the early warning accuracy rate of existing technologies is only 42.8%, and the average early warning delay time is 3.2 days.

[0094] Therefore, the present invention fully verifies the excellent performance of the park intelligent room status map generation technology based on multi-source data fusion. In terms of enterprise demand forecasting, investment recommendation, room management and risk warning, the technical solution of this application shows advantages far exceeding the existing technology.

[0095] Through specific implementation comparison, this embodiment verifies the technical solution of the present invention, which can effectively improve the intelligence and precision of park management, optimize resource allocation, promote the coordinated development of park industries, and provide strong support for the efficient operation and competitiveness improvement of the park.

[0096] The above are only preferred embodiments of the present invention, which do not limit the scope of protection of the present invention. For those skilled in the art, the present invention can be modified and varied in various ways. Any changes, modifications, replacements, integrations and parameter changes to these embodiments through conventional substitutions or that can achieve the same functions without departing from the principles and spirit of the present invention fall within the scope of protection of the present invention.

Claims

1. A system for generating intelligent housing state diagrams in a park based on multi-source data fusion, characterized in that: Including data fusion module, demand forecasting module, and visualization module; The data fusion module includes a data source acquisition module and a data processing module; The data source acquisition module obtains enterprise multi-source data and inputs it into the data processing module; The data processing module receives enterprise multi-source data for pre-processing, labels, and generates label data sources; The demand forecasting module includes a forecasting submodule and an industry chain analysis submodule. The forecasting submodule obtains the label data source, combines the enterprise's historical expansion data and the industry's average growth rate into the area demand model, and outputs the enterprise's required area. The industrial chain analysis submodule obtains the upstream and downstream relationships between enterprises, associates enterprises, suppliers, and customers in the form of triplets, and generates a list of recommended investment invitations; The visualization module includes a room status map generation submodule and a decision execution module; the room status map generation submodule constructs a visual room status map covering the enterprise's labels by superimposing the enterprise's label data source with the room status information; the decision execution module generates a heat map by rendering and overlaying the visual room status map, obtains the risk level and vacancy days, and realizes investment promotion and efficient management.

2. A system for generating intelligent housing state diagrams based on multi-source data fusion according to claim 1, characterized in that: The multi-source enterprise data includes the industrial and commercial data, public opinion data and housing data of enterprises in the park; the industrial and commercial data includes registered capital, business status and industry classification; the public opinion data includes legal risks, business anomalies, bidding information and intellectual property rights; the housing data includes the enterprise's building, floor, room number, lease period and historical tenants.

3. The system for generating intelligent housing state diagram based on multi-source data fusion according to claim 1 is characterized in that: The data processing module removes duplicate values ​​from the enterprise's business data, public opinion data, and housing database through the data cleaning submodule, and performs address standardization processing; The data processing module also includes a credit profile generation sub-module. By obtaining the company's comprehensive corporate data and public opinion risks, the credit profile generation sub-module is used to label the company. Based on legal disputes, business anomalies and insurance fluctuations, the public opinion risk score is calculated, and combined with other business data of the company, high / low potential or high / low risk enterprise labels are generated.

4. The system for generating intelligent housing state diagrams based on multi-source data fusion according to claim 1 is characterized in that: The area demand model is an LSTM network model, which takes the acquired label data source, enterprise historical expansion data and industry average growth rate as input and outputs the enterprise demand area. The formula is: y = FC (LSTM ([x z ,x c ,x b ,x s ,r,k])), where the historical expansion data of the enterprise includes the number of branches x z , Growth rate of insured persons x c 、Increase in the number of winning bids x b 、Increase in the number of intellectual property rights x s , r is the industry average growth rate, k is the risk coefficient, multiple iterative calculations are performed, mapping conversion is performed through the fully connected layer, and the predicted value of the enterprise office area demand is output.

5. The system for generating intelligent housing state diagram based on multi-source data fusion according to claim 1 is characterized in that: The formation of the investment recommendation list is specifically as follows: The industrial chain analysis submodule uses multi-source enterprise data to obtain upstream and downstream relationships between enterprises in the park. Enterprises, suppliers, and customers are associated in the form of triplets to construct an enterprise knowledge graph. This identifies the types of enterprises corresponding to missing or weak links in the industrial chain and screens out potential enterprises that meet the requirements. At the same time, the enterprise knowledge graph is used to evaluate the industry status, market influence, and technological innovation indicators of enterprises to identify leading enterprises in the industry. The industrial chain analysis sub-module follows the priority principle of giving priority to enterprises with gaps in the industrial chain over leading enterprises in the same industry. It internally ranks the selected enterprises with gaps in the industrial chain according to their importance to improving the industrial chain and the urgency of filling the gaps, and places them at the top of the investment recommendation list. It also ranks the leading enterprises in the same industry according to their comprehensive strength and potential value to the park's industries, and lists them after the enterprises with gaps in the industrial chain to generate an investment recommendation list.

6. The system for generating a smart housing state map based on multi-source data fusion according to claim 1 is characterized in that: The room status map generation submodule establishes a mapping relationship M by integrating the enterprise's label data source and room status information. ik , construct a visual housing status diagram, the formula is: where h ij is the jth attribute of property i, l mk is the kth label of enterprise m, w j and v n are the weight coefficients of property attributes and enterprise tags, respectively. a and b are the number of property attributes and enterprise tags, respectively. By integrating the property and enterprise tag information, M is calculated. ik , for each listing i, according to M ik The values ​​and set rules are visualized.

7. A system for generating intelligent housing state diagrams based on multi-source data fusion according to claim 6, characterized in that: The visual property status diagram supports a multi-level search function, searches property information through multiple dimensions, locates and displays the status of properties in the park, including whether they are vacant or rented, and the corresponding enterprise tag information. Based on the property status information, vacant properties are screened out and presented on the visual property status diagram. At the same time, it supports combined search based on industry and risk tags, and screens out qualified properties in the visual diagram with superimposed enterprise tags and property status.

8. The system for generating intelligent housing state diagrams based on multi-source data fusion according to claim 1 is characterized in that: The decision execution module extracts and analyzes data from the visual property status diagram to obtain information related to companies and properties, including company risk tags and contract lease periods. Using the risk assessment system, the properties corresponding to the tagged companies are selected as key assessment targets. Combined with the legal risks and operational anomalies in the public opinion data, a quantitative score is generated to determine the risk level of the property. Obtain the current time of the property and the end-of-term lease expiration date to obtain the number of vacant days. Using the basic property status layer as the bottom layer, overlay the information representing the risk level and vacancy days as different layers to output the park property heat map.

9. The system for generating intelligent housing state diagrams based on multi-source data fusion according to claim 1, characterized in that: The decision execution module also includes an early warning module, which monitors and issues early warnings based on contract expiration and enterprise risk indicators. For early warnings of contract expiration, the early warning module continuously obtains information on the rental period of the property, compares the contract expiration date with the current date, and sets a contract date threshold. If the difference between the contract expiration date and the current date is less than the contract date threshold, an early warning is triggered, and the corresponding property is prompted in the visual property status diagram. For enterprise risk warning, the early warning module obtains integrated public opinion data and enterprise credit portraits, and dynamically monitors the enterprise's risk status. When the risk score exceeds the set risk threshold and is judged to be a high-risk enterprise, the early warning module pushes risk control suggestions and specially marks the properties rented by high-risk enterprises.

10. A method for generating a smart room state map based on multi-source data fusion, based on the smart room state map generation system described in 1-9 above, characterized in that: The following steps are involved: S1. Obtain the industrial and commercial data, public opinion data, and housing data of enterprises in the park for pre-processing, calculate the risk score, and generate a label data source; S2. Obtain the label data source, combine the company's historical expansion data and the industry's average growth rate, input the area demand model, and output the company's office space demand forecast; S3. Obtain the relationships between enterprises, suppliers, and customers, and construct an enterprise knowledge graph using triples to identify potential entrants and leading enterprises in the industry with missing or weak links in the industrial chain. Prioritize enterprises with gaps in the industrial chain over leading enterprises in the same industry, and generate a list of recommended investment invitations. S4. Integrate the enterprise's tag data source with the park's housing status information, visualize each property, and build a visual housing status map that includes the enterprise's tags; S5. Extract and analyze data from the visualized property status map to obtain enterprise risk tags and contract lease periods. Use the risk assessment system combined with public opinion data to determine the property risk level and calculate the number of vacancy days. Using the basic property status layer as the base layer, overlay the risk level and vacancy days information layers to generate a heat map of the park's properties. S6. Monitor the expiration of corporate contracts and corporate risk indicators, push risk control suggestions and specially mark the properties rented by high-risk enterprises.

Images