Diversified financing strategy intelligent recommendation system and method based on big data

Abstract

The invention discloses a diversified financing strategy intelligent recommendation system and method based on big data, relates to the technical field of big data finance, and solves the problems that various types of data of an enterprise are in a dispersed storage and isolated state, effective integration and correlation analysis are difficult to perform, the change condition of risks of the enterprise and a financing project cannot be captured in time, and the risk of the enterprise and the financing project cannot be timely captured. And the recommendation of financing strategies is influenced. Structured data, unstructured data and third-party data of an enterprise are collected, data sources are integrated, a static portrait, a dynamic portrait and a risk portrait of the enterprise are constructed, a heterogeneous graph is established, and information of the enterprise dispersed in all links can be integrated together. And in combination with the enterprise portrait information, comprehensive evaluation of the enterprise financing capability is formed. Information of multiple dimensions such as enterprise risk portraits, financing capability analysis results and financing project risk analysis results is comprehensively considered, and the financing project is comprehensively evaluated.

Description

Technical Field

[0001] This invention belongs to the field of big data finance, specifically a smart recommendation system and method for diversified financing strategies based on big data. Background Technology

[0002] Against the backdrop of fragmented financial resources and diversified corporate financing needs, intelligent recommendation systems integrate multi-dimensional data on corporate credit, debt repayment ability, and financing capacity, combined with machine learning and big data analytics, to provide companies with precisely matched financing project portfolios. A big data-based intelligent recommendation system for diversified financing strategies refers to a decision support platform that integrates multi-dimensional corporate data, uses artificial intelligence algorithms to dynamically analyze financing needs and risks, and intelligently matches and generates customized financing portfolios for different companies.

[0003] Existing financing strategy recommendations lack a system for building corporate risk profiles, analyzing financing capabilities, and analyzing financing project risks. Various corporate data, such as financial data, transaction data, market data, and information related to financing projects, are stored in a scattered and isolated state. Data from different departments or systems is difficult to effectively integrate and correlate. It is also difficult to establish a dynamic heterogeneous graph construction and a system for updating multiple profile data, making it impossible to capture changes in corporate and financing project risks in a timely manner. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art. To this end, this invention proposes a big data-based intelligent recommendation system and method for diversified financing strategies, which addresses the technical problem that various types of enterprise data are stored in a scattered and isolated state, making it difficult to effectively integrate and analyze them, and also unable to capture changes in enterprise and financing project risks in a timely manner, thus affecting the recommendation of financing strategies.

[0005] To address the aforementioned problems, the first aspect of this invention provides an intelligent recommendation system and method for diversified financing strategies based on big data, comprising:

[0006] Profile building module: Collects structured data, unstructured data and third-party data from enterprises, integrates the data sources, classifies enterprises based on structured data, unstructured data and annual report data, and builds static and dynamic profiles of enterprises; based on third-party data and industry risk indicators, it assesses the credit coefficient and solvency coefficient of enterprises and builds enterprise risk profiles.

[0007] Enterprise Analysis Module: By using suppliers, customers, and financial institutions within an enterprise as nodes and transaction records and contractual relationships as edges, a heterogeneous graph is constructed, and combined with enterprise profile information, the enterprise's financing capabilities are analyzed.

[0008] Project Analysis Module: Based on the financing needs information of financing projects, construct a knowledge graph of financing projects and conduct risk analysis of financing projects;

[0009] Financing Project Recommendation Module: Based on the results of financing project risk analysis, as well as the results of enterprise risk profile and enterprise financing capability analysis, candidate financing projects are screened, and a combination of financing projects is recommended from the candidate financing projects based on the results of enterprise risk profile and enterprise financing capability analysis.

[0010] Optionally, in one example of the above aspects, collecting structured data, unstructured data, and third-party data from the enterprise and integrating the data sources includes the following steps:

[0011] Collect structured data from enterprises: business registration information and tax data; unstructured data: enterprise website content data and supply chain contract data; third-party data: credit data, legal litigation records, enterprise annual reports and intellectual property data.

[0012] NLP technology is used to extract keywords and corresponding data information from structured data, unstructured data, and third-party data of enterprises, and time tags are added.

[0013] Optionally, in one example of the above aspects, based on the enterprise's structured data, as well as unstructured data and enterprise annual report data, the enterprise type is classified, and static and dynamic enterprise profiles are constructed, including the following steps:

[0014] Based on the basic information of the enterprise, the main business type of the enterprise is classified according to the registered business type. According to the enterprise's registered capital data and tax data in the enterprise's business information, the enterprise's scale is classified by setting thresholds, and the main business type and enterprise scale are used as the type data of the enterprise.

[0015] Based on enterprise type data, combined with enterprise registered capital data and enterprise tax data in the latest preset time period, a static profile of the enterprise is constructed.

[0016] By extracting keywords and corresponding data information from unstructured data, along with corresponding time tags, we draw a waveform chart of enterprise keyword data. At the same time, based on the enterprise's historical data and the latest annual report data, we draw waveform charts of the enterprise's operating revenue and total profit. All the obtained waveform charts are used as a dynamic profile of the enterprise.

[0017] Optionally, in one example of the above, based on third-party data and combined with industry risk indicators (such as the median industry debt ratio), the enterprise's credit rating and solvency ratio are assessed to construct a corporate risk profile, including the following steps:

[0018] Based on credit data and legal litigation records from third-party data, we obtain corporate debt ratio data, as well as corporate operation-related litigation records and the amounts involved. Based on corporate annual reports and intellectual property data, we extract corporate asset and liability amount data and intellectual property value data.

[0019] Combined with industry risk indicators, including: median industry debt ratio and median amount of debt involved in business-related litigation;

[0020] The credit rating of an enterprise is calculated as follows: α1 * (1 − industry debt ratio / median enterprise debt ratio) + β1 * (1 − litigation debt amount / median industry litigation debt); the solvency ratio of an enterprise is calculated as follows: α2 * current assets / current liabilities + β2 * earnings before interest and taxes (EBIT) / interest expense + γ * total revenue from intellectual property transfers and licensing in the past 3 years / intellectual property valuation, where α1, β1, α2, β2 and γ are the corresponding weights.

[0021] Corporate credit rating and debt repayment ability rating are used as corporate risk profiles.

[0022] Optionally, in one example of the above aspects, a heterogeneous graph is constructed by using suppliers, customers, and financial institutions within the enterprise as nodes and transaction records and contractual relationships as edges. This graph is then combined with enterprise profile information to analyze the enterprise's financing capabilities, including the following steps:

[0023] The enterprise's suppliers, customers, and financial institutions are treated as nodes, and enterprise type data is added as a node label for each node.

[0024] Using transaction records and contractual relationships between enterprises as edges, construct a heterogeneous graph of suppliers, customers, and financial institutions;

[0025] Calculate the sum of transaction records and contractual relationships corresponding to each edge, and filter the maximum amount corresponding to each edge in the heterogeneous graph;

[0026] The weight of each edge in the heterogeneous graph is calculated by dividing the sum of the transaction records and contract amounts corresponding to each edge by the maximum amount corresponding to the edge in the heterogeneous graph.

[0027] Based on the number of edges connecting each node in the heterogeneous graph and the weight of each edge, calculate the centrality PR of each node as PR = Ped * Np1 / Np2, where Ped is the sum of the weights of each edge connecting the node, Np1 is the number of edges connecting the node, and Np2 is the sum of the number of edges in the heterogeneous graph.

[0028] By analyzing corporate volatility, combining node centrality, corporate credit coefficient and solvency coefficient, and corporate return on equity, the financial capacity analysis value of each node in the heterogeneous graph is calculated.

[0029] Optionally, in one example of the above aspects, analyzing firm volatility, combining node centrality, firm credit coefficient and solvency coefficient, and firm return on equity, calculates the firm financing capacity analysis value for each node in the heterogeneous graph, including the following steps:

[0030] Based on the enterprise profile information, calculate the volatility coefficient Vf=σf / μf of each waveform in the enterprise dynamic profile, where σf is the standard deviation of the waveform and μf is the mean of the waveform. Calculate the mean of the volatility coefficients of each waveform as the enterprise volatility coefficient Sv.

[0031] Based on the company's registered capital data and the company's tax data in the latest preset time period, calculate the company's return on net assets (REO) = net income / (company's registered capital + debt * (1 - tax rate)).

[0032] Based on the enterprise volatility coefficient, enterprise credit coefficient, and solvency coefficient, as well as the enterprise's return on net assets, calculate the enterprise financing capacity analysis value for each node in the heterogeneous graph:

[0033]

[0034] Where Et is the corporate financing capacity analysis value, f1 is the corporate credit coefficient, f2 is the debt repayment capacity coefficient, and w1, w2, w3 and w4 are the corresponding weights.

[0035] Add the enterprise financing capability analysis value to the label of the corresponding node.

[0036] Optionally, in one example of the above aspects, a knowledge graph of the financing project is constructed based on the financing demand information of the financing project, and a risk analysis of the financing project is conducted, including the following steps:

[0037] The financing needs information for each financing project includes: project type, project information, funding amount, percentage of completion of the financing project, and expected completion time of the financing project.

[0038] Analyze the cosine similarity between keywords related to project type and project information and keywords of corresponding financing projects in historical financing projects, filter historical financing projects with similarity greater than a threshold, and calculate the investment return ratio and time difference between the initial investment and the peak return of historical financing projects.

[0039] The financing demand information of the financing projects is used to construct a knowledge graph of the financing projects, and the investment return ratio (ps) and the time difference between the initial investment and the peak return of similar historical financing projects are added as labels (ts) to the financing projects.

[0040] Calculate the risk analysis value of the financing project:

[0041]

[0042] Where Rs is the risk analysis value of the financing project, Zs is the non-self-raised funds of the financing project, Zs0 is the total funds required for the financing project, pl is the completed percentage of the financing project, ts0 is the expected remaining time period for the completion of the financing project, and α1, α2 and α3 are the corresponding weights for calculating the risk analysis value of the financing project.

[0043] Optionally, in one example of the above aspects, candidate financing projects are screened based on the results of the financing project risk analysis, as well as the results of the enterprise risk profile and enterprise financing capacity analysis, including the following steps:

[0044] Based on the results of the risk analysis of the financing project, as well as the results of the enterprise risk profile and enterprise financing capability analysis,

[0045] Extract keywords from the enterprise profile information of nodes in the heterogeneous graph, and extract keywords from the financing project knowledge graph. Select financing projects with a cosine similarity greater than the threshold as candidate financing projects.

[0046] Optionally, in one example of the above aspects, a portfolio of financing projects is recommended from alternative financing projects based on the enterprise risk profile and financing capacity analysis results, including the following steps:

[0047] Based on the enterprise risk profile and the analysis results of enterprise financing capacity, thresholds are set for enterprise credit coefficient and debt repayment capacity coefficient, as well as thresholds for enterprise financing capacity analysis value;

[0048] For companies whose credit rating, debt repayment ability rating and financing ability analysis value are all greater than the threshold, the top 80% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

[0049] For companies with two items exceeding the threshold, the top 60% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

[0050] For companies with only one risk analysis value exceeding the threshold, the top 40% of financing projects in the candidate financing projects are selected as the recommended financing project portfolio.

[0051] For companies where none of the three criteria exceed the threshold, the top 10% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

[0052] According to another aspect of this disclosure, a method for intelligent recommendation of diversified financing strategies based on big data is provided. This method uses the aforementioned intelligent recommendation system for diversified financing strategies based on big data to achieve intelligent recommendation of diversified financing strategies.

[0053] This invention constructs a dynamic profile by tracking changes in a company's operating data, market dynamics, and business expansion, reflecting the company's development trends and operational status in real time. This helps to promptly identify potential opportunities and risks, providing support for dynamic adjustments to financing strategies. By combining third-party data and industry risk indicators to assess a company's credit rating and solvency ratio, a risk profile is constructed, enabling a more objective and accurate measurement of the company's risk level. This helps financial institutions rationally determine financing conditions, such as interest rates and terms, reducing financing risks. Based on a comprehensive and accurate company profile, the intelligent recommendation solution can deeply understand a company's unique needs and risk characteristics, tailoring the most suitable financing strategy for each company.

[0054] This invention constructs a heterogeneous graph by using suppliers, customers, and financial institutions as nodes and transaction records and contractual relationships as edges, integrating information scattered across various stages of a company. Combined with company profile information, it forms a comprehensive assessment of a company's financing capabilities. The heterogeneous graph can also reveal complex relationships between different entities; furthermore, as transaction records and contractual relationships are continuously updated, the graph can reflect the company's operational dynamics in real time. By combining this with information such as financial growth trends in the company profile, the assessment of the company's financing capabilities can be dynamically adjusted, making the assessment results more timely. This provides an analysis of a company's financing capabilities from multiple dimensions. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a schematic diagram of the system framework of the present invention;

[0057] Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0058] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Please see Figures 1-2The first aspect of this invention provides an intelligent recommendation system and method for diversified financing strategies based on big data, including:

[0060] The profiling module collects structured, unstructured, and third-party data from enterprises, integrates these data sources, and classifies enterprises based on their structured, unstructured, and annual report data to construct static and dynamic profiles. Furthermore, it assesses enterprise credit ratings and solvency ratios using third-party data and industry risk indicators to build enterprise risk profiles.

[0061] Specifically, in this embodiment, structured and unstructured data of enterprises are collected, such as contracts and news reports in text form, as well as third-party data, such as industry reports. This breaks the limitations of data sources and covers all dimensions of enterprise operations and external environment information, providing a rich and comprehensive data foundation for the formulation of financing strategies.

[0062] Different types of data corroborate and complement each other, helping to reduce the errors and biases that may exist from a single data source. Financial indicators in structured data can be cross-referenced with business contract information in unstructured data, and third-party data can verify the authenticity of the data provided by the company itself, thereby improving the overall accuracy and reliability of the data. Classifying companies based on structured data, unstructured data, and annual report data allows for a more detailed and accurate understanding of the characteristics and attributes of companies. Different types of companies differ in their financing needs and risk characteristics; accurate classification lays the foundation for subsequently developing personalized financing strategies.

[0063] By constructing a static profile of a company based on relatively stable aspects such as its basic information, equity structure, and asset status, financial institutions or investors can quickly understand the company's fundamentals and development foundation, providing a basis for preliminary assessment of the company's value and financing potential.

[0064] By tracking changes in a company's operational data, market dynamics, and business expansion, a dynamic profile can be built, reflecting the company's development trends and operational status in real time. This helps to identify potential opportunities and risks in a timely manner, providing support for the dynamic adjustment of financing strategies.

[0065] By combining third-party data and industry risk indicators to assess a company's credit rating and solvency ratio, a corporate risk profile can be constructed, enabling a more objective and accurate measurement of a company's risk level. This helps financial institutions to rationally determine financing terms, such as interest rates and terms, thereby reducing financing risks.

[0066] Based on comprehensive and accurate enterprise profiles, intelligent recommendation solutions can gain a deep understanding of an enterprise's unique needs and risk characteristics, tailoring the most suitable financing strategy for each company. This personalized recommendation better meets the actual needs of enterprises, improving the success rate and efficiency of financing.

[0067] Dynamic and risk profiles can predict a company's future development trends and potential risks, giving financing strategy recommendations a degree of foresight. Financial institutions or investors can take proactive measures to address potential risks while seizing opportunities for company growth, achieving a balance between risk and return.

[0068] By setting up enterprise analysis and project analysis modules, a heterogeneous graph is constructed by using suppliers, customers, and financial institutions within the enterprise as nodes and transaction records and contractual relationships as edges. Combined with enterprise profile information, the financing capability of the enterprise is analyzed. Based on the financing demand information of financing projects, a financing project knowledge graph is constructed, and financing project risk analysis is conducted.

[0069] Specifically, in this embodiment, suppliers, customers, and financial institutions are used as nodes, and transaction records and contractual relationships are used as edges to construct a heterogeneous graph, which integrates information scattered across various stages of the enterprise. Combined with enterprise profile information, a comprehensive assessment of the enterprise's financing capabilities is formed. The heterogeneous graph can reveal the complex relationships between different entities.

[0070] As transaction records and contractual relationships are continuously updated, heterogeneous graphs can reflect a company's operational dynamics in real time. If a company recently adds several important clients or signs large contracts, the heterogeneous graph can promptly capture these changes and, combined with information such as financial growth trends in the company profile, dynamically adjust the assessment of the company's financing capabilities, making the assessment results more timely. It analyzes a company's financing capabilities from multiple dimensions, considering not only the company's own financial condition and creditworthiness but also relationships with upstream and downstream partners in the supply chain.

[0071] In-depth analysis of corporate financing capabilities based on heterogeneous graphs and corporate profiles can provide a more accurate understanding of a company's financing needs and affordability.

[0072] A financing project knowledge graph can integrate various information related to a financing project, including the project's basic information, market environment, policies and regulations, and participating parties. By analyzing the relationships between this information, it is possible to comprehensively identify various factors that may affect the risks of a financing project. For example, when analyzing a real estate financing project, the knowledge graph can link information from multiple aspects such as local real estate policies, market demand, and the developer's creditworthiness, thereby more accurately identifying project risks.

[0073] Using data from knowledge graphs facilitates the quantitative assessment of risk factors. As financing projects progress and the external environment changes, knowledge graphs can be updated in real time to dynamically assess changes in risk.

[0074] The financing project recommendation module filters candidate financing projects based on the results of financing project risk analysis, as well as the results of enterprise risk profile and enterprise financing capability analysis. Based on the results of enterprise risk profile and enterprise financing capability analysis, it recommends a combination of financing projects from the candidate financing projects.

[0075] Specifically, in this embodiment, the enterprise risk profile comprehensively reflects the enterprise's risk characteristics and tolerance capabilities in terms of credit, market, and operations. By combining this profile with the results of financing project risk analysis, alternative financing projects whose risk levels match the enterprise's tolerance can be selected. The enterprise financing capacity analysis considers factors such as the enterprise's financial condition, asset size, profitability, and cash flow. Based on these analysis results, financing projects that the enterprise is capable of undertaking and operating can be selected from the alternative financing projects.

[0076] Corporate risk profiling and financing project risk analysis are dynamic. As corporate operating conditions, market environment, and policies and regulations change, the company's risk tolerance and the risk level of financing projects will also change accordingly. Based on this dynamic analysis, the financing project portfolio can be adjusted in a timely manner to optimize risk allocation. For example, when a company faces intensified market competition and declining operating performance, it can appropriately reduce the proportion of high-risk financing projects and increase low-risk, stable cash flow financing projects to ensure the company's financial stability.

[0077] The recommended portfolio of financing projects can rationally allocate the direction and duration of funds based on the company's funding needs and investment plans, thereby improving the efficiency of capital utilization. Furthermore, by selecting financing projects with potential for value appreciation, additional investment returns can be generated, enhancing the company's overall value.

[0078] The entire screening and recommendation process is based on a large amount of enterprise and financing project data. Through scientific methods such as data analysis, model calculation, and risk assessment, it provides an objective and accurate basis for decision-making. This avoids subjective assumptions and blind decisions that may exist in traditional decision-making processes, and improves the scientific nature and rationality of the decisions.

[0079] A comprehensive evaluation of financing projects was conducted, taking into account information from multiple dimensions, including corporate risk profiles, financing capacity analysis results, and financing project risk analysis results. This evaluation not only focused on project risks and returns but also considered the alignment of projects with the company's overall strategy and financial situation, ensuring that the recommended portfolio of financing projects can provide strong support for the company's long-term development.

[0080] In one embodiment of the present invention, collecting structured data, unstructured data, and third-party data from enterprises and integrating the data sources includes the following steps:

[0081] Collect structured data from enterprises: business registration information and tax data; unstructured data: enterprise website content data and supply chain contract data; third-party data: credit data, legal litigation records, enterprise annual reports and intellectual property data.

[0082] Structured data specifically includes: business registration information: registered capital, equity structure; tax data: tax records, invoice data; bank statements: income / expenditure frequency, abnormal transactions;

[0083] Unstructured data specifically includes: corporate website content data: product introductions, customer reviews; supply chain data: purchase / sales contracts;

[0084] Third-party data specifically includes: credit data: overdue records, debt ratio; legal litigation records: litigation records related to the company's operations and the amounts involved; company annual reports: asset status; intellectual property data: patent / trademark value;

[0085] NLP technology is used to extract keywords and corresponding data information from structured data, unstructured data, and third-party data of enterprises, and time tags are added.

[0086] In one embodiment of the present invention, based on the structured data, unstructured data, and annual report data of enterprises, enterprise types are classified, and static and dynamic enterprise profiles are constructed, including the following steps:

[0087] Based on the basic information of the enterprise, the main business type of the enterprise is classified according to the registered business type. According to the enterprise's registered capital data and tax data in the enterprise's business information, the enterprise's scale is classified by setting thresholds, and the main business type and enterprise scale are used as the type data of the enterprise.

[0088] Based on enterprise type data, combined with enterprise registered capital data and enterprise tax data in the latest preset time period, a static profile of the enterprise is constructed.

[0089] By extracting keywords and corresponding data information from unstructured data, along with corresponding time tags, we draw a waveform chart of enterprise keyword data. At the same time, based on the enterprise's historical data and the latest annual report data, we draw waveform charts of the enterprise's operating revenue and total profit. All the obtained waveform charts are used as a dynamic profile of the enterprise.

[0090] In one embodiment of the present invention, based on third-party data and combined with industry risk indicators (such as the median industry debt ratio), the credit rating and solvency ratio of enterprises are assessed to construct an enterprise risk profile, including the following steps:

[0091] Based on credit data and legal litigation records from third-party data, we obtain corporate debt ratio data, as well as corporate operation-related litigation records and the amounts involved. Based on corporate annual reports and intellectual property data, we extract corporate asset and liability amount data and intellectual property value data.

[0092] Combined with industry risk indicators, including: median industry debt ratio and median amount of debt involved in business-related litigation;

[0093] The credit rating of an enterprise is calculated as follows: α1 * (1 − industry debt ratio / median enterprise debt ratio) + β1 * (1 − litigation debt amount / median industry litigation debt); the solvency ratio of an enterprise is calculated as follows: α2 * current assets / current liabilities + β2 * earnings before interest and taxes (EBIT) / interest expense + γ * total revenue from intellectual property transfers and licensing in the past 3 years / intellectual property valuation, where α1, β1, α2, β2 and γ are the corresponding weights.

[0094] Corporate credit rating and debt repayment ability rating are used as corporate risk profiles.

[0095] In this embodiment, α1=0.7, the debt ratio weight, reflects financial soundness; β1=0.3, the litigation risk weight, reflects legal compliance; α2=0.4, reflects short-term debt repayment weight; β2=0.4, reflects interest payment weight; and γ=0.2, reflects asset liquidity weight.

[0096] In one embodiment of the present invention, suppliers, customers, and financial institutions within an enterprise are used as nodes, and transaction records and contractual relationships are used as edges to construct a heterogeneous graph. This graph is then combined with enterprise profile information to analyze the enterprise's financing capabilities, including the following steps:

[0097] The enterprise's suppliers, customers, and financial institutions are treated as nodes, and enterprise type data is added as a node label for each node.

[0098] Using transaction records and contractual relationships between enterprises as edges, construct a heterogeneous graph of suppliers, customers, and financial institutions;

[0099] Calculate the sum of transaction records and contractual relationships corresponding to each edge, and filter the maximum amount corresponding to each edge in the heterogeneous graph;

[0100] The weight of each edge in the heterogeneous graph is calculated by dividing the sum of the transaction records and contract amounts corresponding to each edge by the maximum amount corresponding to the edge in the heterogeneous graph.

[0101] Based on the number of edges connecting each node in the heterogeneous graph and the weight of each edge, calculate the centrality PR of each node as PR = Ped * Np1 / Np2, where Ped is the sum of the weights of each edge connecting the node, Np1 is the number of edges connecting the node, and Np2 is the sum of the number of edges in the heterogeneous graph.

[0102] By analyzing corporate volatility, combining node centrality, corporate credit coefficient and solvency coefficient, and corporate return on equity, the financial capacity analysis value of each node in the heterogeneous graph is calculated.

[0103] In one embodiment of the present invention, the analysis of enterprise volatility, combined with the centrality of nodes, enterprise credit coefficient and solvency coefficient, and enterprise return on net assets, calculates the enterprise financing capacity analysis value of each node in the heterogeneous graph, including the following steps:

[0104] Based on the enterprise profile information, calculate the volatility coefficient Vf=σf / μf of each waveform in the enterprise dynamic profile, where σf is the standard deviation of the waveform and μf is the mean of the waveform. Calculate the mean of the volatility coefficients of each waveform as the enterprise volatility coefficient Sv.

[0105] Based on the company's registered capital data and the company's tax data in the latest preset time period, calculate the company's return on net assets (REO) = net income / (company's registered capital + debt * (1 - tax rate)).

[0106] Based on the enterprise volatility coefficient, enterprise credit coefficient, and solvency coefficient, as well as the enterprise's return on net assets, calculate the enterprise financing capacity analysis value for each node in the heterogeneous graph:

[0107]

[0108] Where Et is the corporate financing capacity analysis value, f1 is the corporate credit coefficient, f2 is the debt repayment capacity coefficient, and w1, w2, w3 and w4 are the corresponding weights.

[0109] Add the enterprise financing capability analysis value to the label of the corresponding node.

[0110] Specifically, in this embodiment, w1, w2, w3, and w4 are set to 0.3, 0.4, 0.15, and 0.15, respectively;

[0111] In one embodiment of the present invention, a knowledge graph of financing projects is constructed based on the financing demand information of the financing projects, and a risk analysis of the financing projects is performed, including the following steps:

[0112] The financing needs information for each financing project includes: project type, project information, funding amount, percentage of completion of the financing project, and expected completion time of the financing project.

[0113] Analyze the cosine similarity between keywords related to project type and project information and keywords of corresponding financing projects in historical financing projects, filter historical financing projects with similarity greater than a threshold, and calculate the investment return ratio and time difference between the initial investment and the peak return of historical financing projects.

[0114] The financing demand information of the financing projects is used to construct a knowledge graph of the financing projects, and the investment return ratio (ps) and the time difference between the initial investment and the peak return of similar historical financing projects are added as labels (ts) to the financing projects.

[0115] Calculate the risk analysis value of the financing project:

[0116]

[0117] Where Rs is the risk analysis value of the financing project, Zs is the non-self-raised funds of the financing project, Zs0 is the total funds required for the financing project, pl is the completed percentage of the financing project, ts0 is the expected remaining time period for the completion of the financing project, and α1, α2 and α3 are the corresponding weights for calculating the risk analysis value of the financing project.

[0118] In one embodiment of the present invention, candidate financing projects are screened based on the results of the financing project risk analysis, as well as the results of the enterprise risk profile and enterprise financing capacity analysis, including the following steps:

[0119] Based on the results of the risk analysis of the financing project, as well as the results of the enterprise risk profile and enterprise financing capability analysis,

[0120] Extract keywords from the enterprise profile information of nodes in the heterogeneous graph, and extract keywords from the financing project knowledge graph. Select financing projects with a cosine similarity greater than the threshold as candidate financing projects.

[0121] In one embodiment of the present invention, based on the enterprise risk profile and the analysis results of the enterprise financing capacity, a portfolio of financing projects is recommended from the candidate financing projects, including the following steps:

[0122] Based on the enterprise risk profile and the analysis results of enterprise financing capacity, thresholds are set for enterprise credit coefficient and debt repayment capacity coefficient, as well as thresholds for enterprise financing capacity analysis value;

[0123] For companies whose credit rating, debt repayment ability rating and financing ability analysis value are all greater than the threshold, the top 80% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

[0124] For companies with two items exceeding the threshold, the top 60% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

[0125] For companies with only one risk analysis value exceeding the threshold, the top 40% of financing projects in the candidate financing projects are selected as the recommended financing project portfolio.

[0126] For companies where none of the three criteria exceed the threshold, the top 10% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

[0127] Specifically, based on the enterprise risk profile and the analysis results of enterprise financing capacity, thresholds are set for the enterprise credit coefficient and debt repayment capacity coefficient, as well as the threshold for the enterprise financing capacity analysis value. The thresholds can be set by filtering historical data from regional enterprise historical financing data, where both enterprise financing and financing projects generate revenue, calculating the enterprise credit coefficient, debt repayment capacity coefficient, and enterprise financing capacity analysis value, and taking the average as the corresponding threshold. A buffer zone of 5% above and below the threshold is set, which can be fine-tuned within the buffer zone according to the enterprise. In this embodiment, the enterprise credit coefficient threshold is set to 0.7; the debt repayment capacity coefficient threshold is set to 0.6; and the enterprise financing capacity analysis value threshold is set to 0.75.

[0128] In another embodiment of the present invention, a method for intelligent recommendation of diversified financing strategies based on big data is provided. This method uses the intelligent recommendation system for diversified financing strategies based on big data as described above to achieve intelligent recommendation of diversified financing strategies.

[0129] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A big data-based diversified financing strategy intelligent recommendation system and method, characterized in that, include: Profile building module: Collects structured data, unstructured data and third-party data from enterprises, integrates data sources, classifies enterprise types based on structured data, unstructured data and enterprise annual report data, and builds static and dynamic profiles of enterprises. Based on third-party data and industry risk indicators, we assess corporate credit rating and solvency ratio to construct a corporate risk profile. Enterprise Analysis Module: By using suppliers, customers, and financial institutions within an enterprise as nodes and transaction records and contractual relationships as edges, a heterogeneous graph is constructed, and combined with enterprise profile information, the enterprise's financing capabilities are analyzed. Project Analysis Module: Based on the financing needs information of financing projects, construct a knowledge graph of financing projects and conduct risk analysis of financing projects; Financing Project Recommendation Module: Based on the results of financing project risk analysis, as well as the results of enterprise risk profile and enterprise financing capability analysis, candidate financing projects are screened, and a combination of financing projects is recommended from the candidate financing projects based on the results of enterprise risk profile and enterprise financing capability analysis.

2. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, Collecting structured data, unstructured data, and third-party data from enterprises, and integrating the data sources, includes the following steps: Collect structured data from enterprises: business registration information and tax data; unstructured data: enterprise website content data and supply chain contract data; third-party data: credit data, legal litigation records, enterprise annual reports and intellectual property data. NLP technology is used to extract keywords and corresponding data information from structured data, unstructured data, and third-party data of enterprises, and time tags are added.

3. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, Based on structured data, unstructured data, and annual report data of enterprises, enterprise types are classified, and static and dynamic enterprise profiles are constructed, including the following steps: Based on the basic information of the enterprise, the main business type of the enterprise is classified according to the registered business type. According to the enterprise's registered capital data and tax data in the enterprise's business information, the enterprise's scale is classified by setting thresholds, and the main business type and enterprise scale are used as the type data of the enterprise. Based on enterprise type data, combined with enterprise registered capital data and enterprise tax data in the latest preset time period, a static profile of the enterprise is constructed. By extracting keywords and corresponding data information from unstructured data, along with corresponding time tags, we draw a waveform chart of enterprise keyword data. At the same time, based on the enterprise's historical data and the latest annual report data, we draw waveform charts of the enterprise's operating revenue and total profit. All the obtained waveform charts are used as a dynamic profile of the enterprise.

4. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, Based on third-party data and combined with industry risk indicators (such as the median industry debt ratio), the credit rating and solvency ratio of enterprises are assessed to construct a risk profile of the enterprise, including the following steps: Based on credit data and legal litigation records from third-party data, we obtain corporate debt ratio data, as well as corporate operation-related litigation records and the amounts involved. Based on corporate annual reports and intellectual property data, we extract corporate asset and liability amount data and intellectual property value data. Combined with industry risk indicators, including: median industry debt ratio and median amount of debt involved in business-related litigation; The credit rating of an enterprise is calculated as follows: α1 * (1 − industry debt ratio / median enterprise debt ratio) + β1 * (1 − litigation debt amount / median industry litigation debt); the solvency ratio of an enterprise is calculated as follows: α2 * current assets / current liabilities + β2 * earnings before interest and taxes (EBIT) / interest expense + γ * total revenue from intellectual property transfers and licensing in the past 3 years / intellectual property valuation, where α1, β1, α2, β2 and γ are the corresponding weights. Corporate credit rating and debt repayment ability rating are used as corporate risk profiles.

5. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, By treating suppliers, customers, and financial institutions within a company as nodes and transaction records and contractual relationships as edges, a heterogeneous graph is constructed. Combined with company profile information, an analysis of the company's financing capabilities is performed, including the following steps: The enterprise's suppliers, customers, and financial institutions are treated as nodes, and enterprise type data is added as a node label for each node. Using transaction records and contractual relationships between enterprises as edges, construct a heterogeneous graph of suppliers, customers, and financial institutions; Calculate the sum of transaction records and contractual relationships corresponding to each edge, and filter the maximum amount corresponding to each edge in the heterogeneous graph; The weight of each edge in the heterogeneous graph is calculated by dividing the sum of the transaction records and contract amounts corresponding to each edge by the maximum amount corresponding to the edge in the heterogeneous graph. Based on the number of edges connecting each node in the heterogeneous graph and the weight of each edge, calculate the centrality PR of each node as PR = Ped * Np1 / Np2, where Ped is the sum of the weights of each edge connecting the node, Np1 is the number of edges connecting the node, and Np2 is the sum of the number of edges in the heterogeneous graph. By analyzing corporate volatility, combining node centrality, corporate credit coefficient and solvency coefficient, and corporate return on equity, the financial capacity analysis value of each node in the heterogeneous graph is calculated.

6. The intelligent recommendation system for diversified financing strategies based on big data according to claim 5, characterized in that, To analyze corporate volatility, and combining node centrality, corporate credit coefficients, solvency coefficients, and return on equity, the financial capacity analysis value of each node in the heterogeneous graph is calculated, including the following steps: Based on the enterprise profile information, calculate the volatility coefficient Vf=σf / μf of each waveform in the enterprise dynamic profile, where σf is the standard deviation of the waveform and μf is the mean of the waveform. Calculate the mean of the volatility coefficients of each waveform as the enterprise volatility coefficient Sv. Based on the company's registered capital data and the company's tax data in the latest preset time period, calculate the company's return on net assets (REO) = net income / (company's registered capital + debt * (1 - tax rate)). Based on the enterprise volatility coefficient, enterprise credit coefficient, and solvency coefficient, as well as the enterprise's return on net assets, calculate the enterprise financing capacity analysis value for each node in the heterogeneous graph: Where Et is the corporate financing capacity analysis value, f1 is the corporate credit coefficient, f2 is the debt repayment capacity coefficient, and w1, w2, w3 and w4 are the corresponding weights. Add the enterprise financing capability analysis value to the label of the corresponding node.

7. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, Based on the financing needs information of the financing projects, a knowledge graph of the financing projects is constructed, and a risk analysis of the financing projects is conducted, including the following steps: The financing needs information for each financing project includes: project type, project information, funding amount, percentage of completion of the financing project, and expected completion time of the financing project. Analyze the cosine similarity between keywords related to project type and project information and keywords of corresponding financing projects in historical financing projects, filter historical financing projects with similarity greater than a threshold, and calculate the investment return ratio and time difference between the initial investment and the peak return of historical financing projects. The financing demand information of the financing projects is used to construct a knowledge graph of the financing projects, and the investment return ratio (ps) and the time difference between the initial investment and the peak return of similar historical financing projects are added as labels (ts) to the financing projects. Calculate the risk analysis value of the financing project: Where Rs is the risk analysis value of the financing project, Zs is the non-self-raised funds of the financing project, Zs0 is the total funds required for the financing project, pl is the completed percentage of the financing project, ts0 is the expected remaining time period for the completion of the financing project, and α1, α2 and α3 are the corresponding weights for calculating the risk analysis value of the financing project.

8. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, Based on the results of the risk analysis of financing projects, as well as the results of the enterprise risk profile and financing capacity analysis, candidate financing projects are screened, including the following steps: Based on the results of the risk analysis of the financing project, as well as the results of the enterprise risk profile and enterprise financing capability analysis, Extract keywords from the enterprise profile information of nodes in the heterogeneous graph, and extract keywords from the financing project knowledge graph. Select financing projects with a cosine similarity greater than the threshold as candidate financing projects.

9. The intelligent recommendation system for diversified financing strategies based on big data according to claim 1, characterized in that, Based on the analysis of corporate risk profiles and financing capabilities, a portfolio of financing projects is recommended from the available financing options, including the following steps: Based on the enterprise risk profile and the analysis results of enterprise financing capacity, thresholds are set for enterprise credit coefficient and debt repayment capacity coefficient, as well as thresholds for enterprise financing capacity analysis value; For companies whose credit rating, debt repayment ability rating and financing ability analysis value are all greater than the threshold, the top 80% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio. For companies with two items exceeding the threshold, the top 60% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio. For companies with only one risk analysis value exceeding the threshold, the top 40% of financing projects in the candidate financing projects are selected as the recommended financing project portfolio. For companies where none of the three criteria exceed the threshold, the top 10% of financing projects in terms of risk analysis value among the candidate financing projects are selected as the recommended financing project portfolio.

10. A method for intelligently recommending diversified financing strategies based on big data, characterized in that: This method employs a big data-based intelligent recommendation system for diversified financing strategies, as described in any one of claims 1-9, to achieve intelligent recommendation of diversified financing strategies.

Images