Credit report abstract generation method and device, computer equipment and storage medium

Abstract

The invention belongs to the field of financial information, and relates to a credit report abstract generation method and device, computer equipment and a storage medium. The credit report abstract generation comprises the following steps: obtaining enterprise credit index information; training the logistic regression model according to the financial analysis data to obtain an enterprise feature classification model; inputting the enterprise credit index information into an enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; inputting the enterprise advantage characteristics into a multi-dimensional index aggregation model to obtain industry advantage indexes and regional advantage indexes; analyzing the enterprise risk characteristics according to a multi-dimensional analysis network to obtain enterprise self-risk information and enterprise associated risk information; and fusing the industry advantage index, the regional advantage index, the enterprise self-risk information and the enterprise associated risk information to obtain an enterprise characteristic index set and generate an enterprise credit report abstract. According to the method, the accurate and effective enterprise credit report abstract can be effectively generated according to the enterprise credit report.

Description

Technical Field

[0001] The present application relates to the field of financial information, and in particular to a credit report summary generation method, device, computer equipment and storage medium. Background Art

[0002] In the modern business environment, corporate credit investigation is an important cornerstone for the sound operation of the financial market. Its accuracy and comprehensiveness are directly related to the credit decisions of financial institutions, the selection of supply chain partners and the overall health of the social economy. As the core output of this process, the corporate credit report summary is designed to quickly overview the credit status of the company and provide key information reference for decision makers. However, with the continuous development of the market economy and the increasing diversification of corporate forms, traditional credit assessment methods are facing unprecedented challenges.

[0003] Different companies show completely different credit characteristics and risk profiles due to factors such as their position in the industry, the size of their business, the economic environment of their region, and their operating strategies in specific business scenarios. The complex interaction between these factors makes it difficult for a single or standardized credit assessment model to fully capture and accurately reflect the true credit status of a company. For example, a start-up technology company may have a high future expected value due to its high growth potential, but its current financial performance may not be outstanding. If it is evaluated only based on traditional financial indicators, its credit level may be underestimated; on the contrary, although large traditional industry companies are financially sound, they may face potential risks due to the pressure of industry transformation. If this is ignored, their credit status may be overestimated. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a credit report summary generation method, apparatus, computer equipment and storage medium to solve the problem of being unable to generate an accurate and effective corporate credit report summary.

[0005] In order to solve the above technical problems, the present application embodiment provides a credit report summary generation method, which adopts the following technical solution:

[0006] Obtain enterprise credit index information;

[0007] Acquire financial analysis data, and train a preset logistic regression model based on the financial analysis data to obtain an enterprise feature classification model;

[0008] Inputting the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features;

[0009] Building a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and inputting the enterprise advantage characteristics into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators;

[0010] Analyze the risk characteristics of the enterprise according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and enterprise-related risk information;

[0011] The industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information are integrated to obtain a set of enterprise characteristic indicators;

[0012] A text is generated according to the enterprise characteristic indicator set to obtain a summary of the enterprise credit report.

[0013] Furthermore, the step of obtaining the enterprise credit index information specifically includes:

[0014] Obtain corporate credit reports;

[0015] Performing text recognition on the enterprise credit report to obtain enterprise credit text information;

[0016] The enterprise credit investigation text information is preprocessed to obtain the enterprise credit investigation index information.

[0017] Furthermore, the financial analysis data includes financial instance sample data, financial public data, industry report data, and regional report data. The steps of obtaining the financial analysis data, training a preset logistic regression model according to the financial analysis data, and obtaining the enterprise feature classification model specifically include:

[0018] Acquire the financial instance sample data, the financial public data, the industry report data, and the regional report data;

[0019] Preprocessing the financial instance sample data, the financial public data, the industry report data, and the regional report data to obtain standard financial instance sample data, standard financial public data, standard industry report data, and standard regional report data;

[0020] Extracting features from the standard financial instance sample data, the standard financial public data, the standard industry report data, and the standard regional report data to obtain full financial features, financial public features, industry report features, and regional report features;

[0021] The total amount of financial features, the financial public features, the industry report features, and the regional report features are input into a preset classification regression model for training to obtain the enterprise feature classification model.

[0022] Furthermore, the financial analysis data includes financial instance sample data, financial public data, industry report data, and regional report data. The step of constructing a model based on the financial analysis data to obtain a multi-dimensional indicator aggregation model specifically includes:

[0023] Extracting dimension indicators from the financial instance sample data, the financial public data, the industry report data, and the regional report data to obtain financial industry dimension features and financial regional dimension features;

[0024] Performing feature dimensionality upgrade on the financial industry dimension feature and the financial region dimension feature to obtain the financial industry dimensionality upgrade feature and the financial region dimensionality upgrade feature;

[0025] The financial industry dimensionality-raising features and the financial region dimensionality-raising features are input into a preset factor analysis model for training to obtain the multidimensional indicator aggregation model.

[0026] Furthermore, the step of analyzing the enterprise risk characteristics according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and the enterprise-related risk information specifically includes:

[0027] Parsing the financial analysis data to obtain financial instance sample data;

[0028] Performing transaction flow analysis on the sample data of the financial instance to obtain transaction association information, and constructing a heterogeneous graph based on the transaction association information and the enterprise risk characteristics to obtain a transaction information heterogeneous graph;

[0029] Acquire the multidimensional analysis network, wherein the multidimensional analysis network includes a heterogeneous graph attention network and a heterogeneous graph Transformer network;

[0030] Inputting the transaction information heterogeneous graph into the heterogeneous graph attention network for analysis to obtain enterprise risk coding, and using the enterprise risk coding as the enterprise's own risk information;

[0031] The transaction information heterogeneous graph is input into the heterogeneous graph Transformer network for analysis to obtain enterprise-related risk codes, and the enterprise-related risk codes are used as the enterprise-related risk information.

[0032] Furthermore, the step of fusing the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain a set of enterprise characteristic indicators specifically includes:

[0033] Performing correlation analysis on the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise's associated risk information to obtain indicator correlation information;

[0034] Obtain a preset weight distribution table, map the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to the weight distribution table, and obtain indicator weight information;

[0035] The industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information are integrated according to the indicator association information and the indicator weight information to obtain the enterprise characteristic indicator set.

[0036] Furthermore, the step of generating text according to the set of enterprise characteristic indicators to obtain a summary of the enterprise credit report specifically includes:

[0037] Get the text algorithm extraction identifier;

[0038] Extracting a summary text generation algorithm from a database according to the text algorithm extraction identifier;

[0039] The enterprise characteristic indicator set is converted into text according to a preset natural language processing algorithm and the summary text generation algorithm to obtain the enterprise credit report summary.

[0040] In order to solve the above technical problems, the embodiment of the present application also provides a credit report summary generation device, which adopts the following technical solution:

[0041] Information acquisition module, used to obtain corporate credit index information;

[0042] A model training module is used to obtain financial analysis data, train a preset logistic regression model according to the financial analysis data, and obtain an enterprise feature classification model;

[0043] A feature extraction module, used to input the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features;

[0044] An indicator acquisition module is used to construct a model based on the financial analysis data to obtain a multi-dimensional indicator aggregation model, and input the enterprise advantage characteristics into the multi-dimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators;

[0045] A risk analysis module is used to analyze the risk characteristics of the enterprise according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and enterprise-related risk information;

[0046] An information fusion module, used to fuse the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain a set of enterprise characteristic indicators;

[0047] The text generation module is used to generate text according to the enterprise characteristic indicator set to obtain the enterprise credit report summary.

[0048] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0049] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the credit report summary generation method as described in any one of the above items are implemented.

[0050] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0051] A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the credit report summary generation method as described in any one of the above items.

[0052] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0053] This embodiment obtains enterprise credit index information; obtains financial analysis data, trains a preset logistic regression model according to the financial analysis data, and obtains an enterprise feature classification model; inputs the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; constructs a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and inputs the enterprise advantage features into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; analyzes the enterprise risk features according to a preset multidimensional analysis network to obtain enterprise risk information and enterprise-related risk information; integrates the industry advantage indicators, the regional advantage indicators, the enterprise risk information, and the enterprise-related risk information to obtain an enterprise feature indicator set; generates text according to the enterprise feature indicator set to obtain an enterprise credit report summary. Thus, an accurate and effective enterprise credit report summary is effectively generated based on the enterprise credit report, so as to facilitate a quick understanding of the credit status of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 A flowchart of an embodiment of a method for generating a credit report summary according to the present application;

[0056] Figure 2 yes Figure 1 A flowchart of a specific implementation of step S10;

[0057] Figure 3 yes Figure 1 A flowchart of a specific implementation of step S20;

[0058] Figure 4 yes Figure 1 A flowchart of a specific implementation of step S30;

[0059] Figure 5 yes Figure 1 A flowchart of a specific implementation of step S50;

[0060] Figure 6 yes Figure 1 A flowchart of a specific implementation of step S60;

[0061] Figure 7 yes Figure 1 A flowchart of a specific implementation of step S70;

[0062] Figure 8 is a structural schematic diagram of an embodiment of a credit report summary generating device according to the present application;

[0063] Fig. 9 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0065] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0066] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0067] Continue to refer Figure 1 , showing a flow chart of an embodiment of a method for speaker production testing according to the present application. The credit report summary generation method comprises the following steps:

[0068] Step S10, obtaining enterprise credit index information;

[0069] In this embodiment, the enterprise credit index information refers to the enterprise credit report after text extraction and preprocessing. The enterprise credit report includes basic information of the enterprise: identity information: such as the name of the organization, unified social credit code, address, legal representative, etc.; registration information: including registered capital, enterprise type, business scope, establishment time, existence status, etc.; shareholder structure: display the shareholder information of the enterprise, investment ratio, etc.; management and employee information: some reports may include the profile of senior management and employees of the enterprise. Credit transaction information: credit information: recorded credit activities such as enterprise loans, factoring, bill discounting, trade financing, letters of credit, guarantees, bank acceptance bills, guarantees, etc.; non-credit credit information: such as public utility payment records (electricity, water, gas, etc.), housing provident fund payment information, court defaulter information, administrative penalty information, etc. Business information: main business and business transactions, including procurement, production, sales information and the evaluation of the credited enterprise by major suppliers and major customers. Financial information: including the balance sheet, profit and loss statement and main financial ratios of the enterprise, and notes will be made for abnormal financial data. Financial institution information: including the company's account number, bank name, etc. Public record information: mainly includes positive and negative information about the company recorded by administrative and judicial departments such as courts, industry and commerce, taxation, quality supervision, customs, and other organizations and social organizations with administrative management functions, as well as media and electronic trading platforms.

[0070] Step S20, obtaining financial analysis data, and training a preset logistic regression model according to the financial analysis data to obtain an enterprise feature classification model;

[0071] In this embodiment, financial analysis data refers to data used to analyze the financial status of an enterprise. Financial analysis data includes financial instance sample data, financial public data, industry report data, and regional report data. The preset logistic regression model uses a logistic regression model. The logistic regression model is used to process regression problems with dependent variables as two categories (or binary). The probability is estimated through a logistic function (the Sigmoid function can be used) and then classified. The Sigmoid function can map any real value to the interval (0,1), indicating the probability of an event occurring. Given a set of independent variables X (which may be one or more features), the logistic regression model calculates the probability of outputting Y=1 (i.e., the probability of an event occurring) as: P(Y=1|X)=1 / (1+e -(β0+β1X1+β2X2+…+βnXn) ), where e is the base of the natural logarithm, β0, β1, …, β n are the model parameters, X1, X2, …, X n is the independent variable (feature).

[0072] Step S30, inputting the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features;

[0073] In this embodiment, the enterprise advantage characteristics refer to the characteristic information that characterizes the advantages of the enterprise in the industry or region, wherein the enterprise advantage characteristics include the enterprise industry advantage characteristics and the enterprise regional advantage characteristics, the enterprise industry advantage characteristics include technological innovation advantages, market share advantages, brand influence, cost advantages, channel advantages, strategic positioning advantages, etc., and the enterprise regional advantage characteristics include natural resource advantages, geographical location advantages, policy support advantages, talent resource advantages, industrial cluster advantages, market demand advantages, etc. Enterprise risk characteristics refer to the characteristic information that characterizes the risks of the enterprise itself and the risks of the enterprise-related objects, wherein the enterprise risk characteristics include the enterprise's own risk characteristics and the enterprise-related risk characteristics, the enterprise's own risk characteristics include subjective risk, objective risk, profit and loss, conditionality, time limit, diversity, dynamics, etc., and the enterprise-related risk characteristics include credit expansion, virtual guarantee, related transaction risk, debt evasion risk, information supervision difficulty, etc.

[0074] Step S40, constructing a model based on the financial analysis data to obtain a multi-dimensional indicator aggregation model, and inputting the enterprise advantage characteristics into the multi-dimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators;

[0075] In this embodiment, the multidimensional index aggregation model adopts a factor analysis model, and the steps of constructing the factor analysis model include KMO and Bartlett test: KMO (Kaiser-Meyer-Olkin) test and Bartlett sphericity test are performed to determine whether the data is suitable for factor analysis; extract common factors: use principal component analysis, maximum likelihood method and other methods to extract common factors, and the number of extracted factors can be determined according to the criteria such as characteristic root greater than 1 and cumulative variance contribution rate; factor rotation: in order to obtain factors that are easier to interpret, the factor loading matrix can be rotated, such as using the maximum variance orthogonal rotation method; factor naming and interpretation: according to the rotated factor loading matrix, the factors are named and interpreted to clarify the variable combination represented by each factor and its economic significance. Industry advantage indicators include technological innovation capability: reflecting the industry's ability in technology research and development and product innovation; market share: reflecting the industry's competitive position and influence in the market; profitability: measuring the industry's overall profitability and profitability; industrial chain integration capability: evaluating the industry's integration and optimization capabilities in the supply chain and value chain; sustainable development capability: examining the industry's performance in environmental protection, social responsibility and other aspects, as well as its long-term development potential. Regional advantage indicators include economic foundation: evaluating the region's total economic volume, industrial structure, economic growth rate, etc.; resource advantages: examining the region's advantages in natural resources, human resources, technological resources, etc.; policy environment: analyzing the support of regional government policies for economic development, business environment, etc.; transportation and logistics: evaluating the region's transportation network, logistics facilities and their efficiency; education and scientific research: measuring the region's strength in higher education, scientific research institutions, talent pool, etc.

[0076] Step S50, analyzing the enterprise risk characteristics according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and enterprise-related risk information;

[0077] In this embodiment, the preset multidimensional analysis network includes a heterogeneous graph attention network and a heterogeneous graph Transformer network, wherein the heterogeneous graph attention network and the heterogeneous graph Transformer network are networks for analyzing heterogeneous graphs, and heterogeneous graphs refer to graphs containing multiple types of nodes and edges, which represent special types of graph structures of different entities and relationships. The heterogeneous graph attention network (HAN) is a neural network model specifically used to process heterogeneous graphs. It dynamically allocates weights of different neighbor nodes by introducing an attention mechanism, and considers the importance of different meta-paths, thereby improving the model's ability to process heterogeneous graph structure information. The HAN model includes two levels of node-level attention and semantic-level attention, which can more accurately capture the complex relationships and semantic information between nodes in heterogeneous graphs. The heterogeneous graph Transformer network is a neural network architecture specifically used to process and analyze heterogeneous graphs, including self-attention mechanisms, embedding of heterogeneous nodes and edges, meta-path and semantic-level attention, and efficient parallel processing capabilities.

[0078] Step S60, integrating the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain an enterprise characteristic index set;

[0079] In this embodiment, the enterprise characteristic indicator set is a collection of multiple indicators and risk information that have been sorted, including multiple indicators and data points that describe various aspects of the enterprise. These indicators can cover multiple dimensions such as the enterprise's financial status, market performance, operational efficiency, innovation capabilities, industry status, regional advantages, and potential risks.

[0080] Step S70, generating text according to the enterprise characteristic indicator set to obtain a summary of the enterprise credit report.

[0081] In this embodiment, the corporate credit report summary is text data that integrates key information of various indicators. The corporate information report may include: corporate profile, financial overview, market performance and competitiveness, operational efficiency and innovation capabilities, industry status and regional advantages, potential risks, etc. After the corporate credit report summary is generated, it can be saved in the database for easy extraction and viewing.

[0082] This embodiment obtains enterprise credit index information; obtains financial analysis data, trains a preset logistic regression model according to the financial analysis data, and obtains an enterprise feature classification model; inputs the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; constructs a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and inputs the enterprise advantage features into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; analyzes the enterprise risk features according to a preset multidimensional analysis network to obtain enterprise risk information and enterprise-related risk information; integrates the industry advantage indicators, the regional advantage indicators, the enterprise risk information, and the enterprise-related risk information to obtain an enterprise feature indicator set; generates text according to the enterprise feature indicator set to obtain an enterprise credit report summary. Thus, an accurate and effective enterprise credit report summary is effectively generated based on the enterprise credit report, so as to facilitate a quick understanding of the credit status of the enterprise.

[0083] In this embodiment, the enterprise feature classification model can also adopt the Logistic regression three-classification model. When the enterprise feature classification model adopts the Logistic regression three-classification model, after the enterprise credit index information is input into the enterprise feature classification model, the enterprise advantage characteristics, enterprise standard characteristics, and enterprise risk characteristics are output, wherein the enterprise standard characteristics refer to the characteristics of the basic attributes of the enterprise, and the enterprise standard characteristics include enterprise income characteristics, enterprise scale characteristics, enterprise personnel distribution characteristics, enterprise business industry characteristics, etc. After obtaining industry advantage indicators, regional advantage indicators, enterprise's own risk information, and enterprise-related risk information, the enterprise standard characteristics are directly converted into enterprise standard information, and then the enterprise standard information, industry advantage indicators, regional advantage indicators, enterprise's own risk information, and enterprise-related risk information are integrated to obtain a set of enterprise feature indicators, and a summary of the enterprise credit report is generated based on the set of enterprise feature indicators.

[0084] Continue to refer Figure 2 In some optional implementations of this embodiment, step S10 includes the following steps:

[0085] Step S101, obtaining a corporate credit report;

[0086] In this embodiment, the enterprise credit report can be extracted from the database through the credit information identifier, which is a unique identifier corresponding to the enterprise credit report and is used to perform a matching query in the database to obtain the enterprise credit report.

[0087] Step S102, performing text recognition on the enterprise credit report to obtain enterprise credit text information;

[0088] In this embodiment, the corporate credit report refers to an image file of the corporate credit report, which may be a photo, screenshot, or scanned file, etc. The corporate credit report image file is preprocessed by denoising, tilt correction, binarization, etc., so as to obtain a clear and effective corporate credit report image. The preprocessed corporate credit report image is then segmented into single characters or character groups, and feature information of the characters, such as strokes, structure, contour, etc., is extracted. A convolutional neural network (CNN) model is used to match and identify the extracted character features, thereby completing the text recognition step of the corporate credit report to effectively obtain the corporate credit text information.

[0089] Step S103, pre-processing the enterprise credit investigation text information to obtain the enterprise credit investigation index information.

[0090] In this embodiment, preprocessing includes removing irrelevant information: deleting advertisements, watermarks, headers and footers, copyright statements and other information in the text that is irrelevant to the corporate credit; removing noise: removing garbled characters, special symbols, non-text characters and other noise data in the text; blank processing: removing extra spaces, blank lines, tabs, etc. to ensure that the text is neat; unifying the format: unifying special format data such as dates, amounts, telephone numbers, etc. in the text into a standardized format for easy subsequent processing; text encoding: ensuring that the text encoding is consistent to avoid garbled characters; case conversion: converting the text to all lowercase or all uppercase as needed to eliminate the impact of case differences on subsequent processing.

[0091] This embodiment obtains a corporate credit report; performs text recognition on the corporate credit report to obtain corporate credit text information; and pre-processes the corporate credit text information to obtain standard and effective corporate credit indicator information to provide a reliable data basis for subsequent processing.

[0092] Continue to refer Figure 3 In some optional implementations of this embodiment, step S20 includes the following steps:

[0093] Step S201, obtaining the financial instance sample data, the financial public data, the industry report data, and the regional report data;

[0094] In this embodiment, the sample data of financial examples include stock data: including stock price, trading volume, price-earnings ratio, price-to-book ratio, yield, etc.; bond data: involving bond interest rate, yield, expiration date, face value, market value, etc.; futures data: including futures contract price, trading volume, opening price, closing price, leverage ratio, etc.; option data: such as option contract price, trading volume, strike price, withdrawal price, volatility, etc.; foreign exchange data: involving exchange rate changes and trading volume between different currencies. Financial public data includes policy data: including financial policies, regulations, announcements issued by the government, etc.; macroeconomic data: such as GDP growth rate, inflation rate, unemployment rate, etc. Industry report data is report information on the market size, growth rate, competitive landscape, etc. of the enterprise industry. The regional report data includes economic data: regional gross domestic product (GDP); industrial structure: including the proportion and development of various industries such as agriculture, industry, and services; fixed asset investment: showing the region's investment in infrastructure and industrial development; consumption and retail: the consumption capacity and market activity of regional residents; imports and exports and foreign trade: the region's foreign economic relations and international trade; fiscal revenue and expenditure: government fiscal revenue and expenditure; demographics: including population size, structure, distribution, mobility, etc.; employment and unemployment: reflecting the supply and demand situation and employment quality of the regional labor market; residents' income and consumption: reflecting the living standards and consumption habits of regional residents; regional development plans: including short-term, medium-term and long-term development goals and plans; policy documents: relevant policies, regulations, notices and other documents issued by the government; major projects: important investment projects and infrastructure construction in the region.

[0095] In this embodiment, financial instance sample data, financial public data, industry report data, and regional report data are stored in a database, and can be extracted by performing a matching query in the database through a financial information extraction identifier.

[0096] Step S202, pre-processing the financial instance sample data, the financial public data, the industry report data, and the regional report data to obtain standard financial instance sample data, standard financial public data, standard industry report data, and standard regional report data;

[0097] In this embodiment, the preprocessing includes removing irrelevant information: deleting advertisements, watermarks, headers and footers, copyright statements and other information irrelevant to the enterprise credit in the text; removing noise: removing noise data such as garbled characters, special symbols, non-text characters in the text; blank processing: removing extra spaces, blank lines, tabs, etc. to ensure the neatness of the text; unifying the format: unifying the special format data such as date, amount, telephone number, etc. in the text into a standardized format for subsequent processing; text encoding: ensuring that the text encoding is consistent to avoid garbled characters; case conversion: converting the text to all lowercase or all uppercase as needed to eliminate the impact of case differences on subsequent processing. By preprocessing the financial instance sample data, financial public data, and financial report data, accurate and effective standard financial instance sample data, standard financial public data, and standard financial report data are obtained.

[0098] Step S203, extracting features from the standard financial instance sample data, the standard financial public data, the standard industry report data, and the standard regional report data to obtain full financial features, financial public features, industry report features, and regional report features;

[0099] In this embodiment, feature extraction of standard financial instance sample data, standard financial public data, standard industry report data, and standard regional report data can be achieved through feature definition and feature calculation. Feature definition includes selecting features that are helpful for prediction or analysis from standard full financial data, standard financial public data, standard industry report data, and standard regional report data. The feature can be selected based on business experience, industry knowledge, or statistical methods (such as chi-square test, information gain, etc.). Each selected feature is clearly defined, including its calculation method, data source, physical meaning, etc. For example, the rate of return can be defined as "(current price-previous period price) / previous period price", and volatility can be defined as "standard deviation of price changes over a period of time", etc. Feature calculation includes extracting corresponding data fields from the original data according to the feature definition. For features that need to be calculated, such as technical indicators, financial indicators, etc., the basic data required for calculation must be extracted first, and the extracted data is calculated according to the calculation method in the feature definition to obtain the feature value. This calculation step can be implemented using programming languages ​​(such as Python, R, etc.) and corresponding data processing libraries (such as pandas, numpy, etc.). The full range of financial characteristics include price-related characteristics, volatility characteristics, technical indicator characteristics, and liquidity characteristics; the financial public characteristics include macroeconomic characteristics, policy-oriented characteristics, and industry characteristics; the industry report characteristics include financial status characteristics and operating results characteristics; the regional report characteristics include regional financial situation characteristics, regional economic indicator characteristics, regional resource information characteristics, and regional policy status characteristics.

[0100] Step S204: input the total financial features, the financial public features, the industry report features, and the regional report features into a preset classification regression model for training to obtain the enterprise feature classification model.

[0101] In this embodiment, the preset classification regression model adopts the Logistic regression model. The training steps of the Logistic regression model include initializing the Logistic regression model: setting the initial parameters of the model, such as learning rate, number of iterations, regularization terms, etc.; training the model: using the full amount of financial features, financial public features, and financial report features to train the Logistic regression model. During the training process, the model will learn the mapping relationship between features and labels, and optimize the model parameters to minimize the loss function (cross entropy loss); verifying the model: during the training process or after the training is completed, the model is verified using the verification set data to evaluate the performance of the model and adjust the model parameters to obtain an effective enterprise feature classification model.

[0102] This embodiment obtains the financial instance sample data, the financial public data, the industry report data, and the regional report data; pre-processes the financial instance sample data, the financial public data, the industry report data, and the regional report data to obtain standard financial instance sample data, standard financial public data, standard industry report data, and standard regional report data; extracts features from the standard financial instance sample data, the standard financial public data, the standard industry report data, and the standard regional report data to obtain full financial features, financial public features, industry report features, and regional report features; inputs the full financial features, the financial public features, the industry report features, and the regional report features into a preset classification regression model for training, thereby effectively obtaining an enterprise feature classification model trained based on the financial instance sample data, the financial public data, the industry report data, and the regional report data, so as to facilitate the subsequent classification of enterprise credit indicator information according to the enterprise feature classification model.

[0103] Continue to refer Figure 4 In some optional implementations of this embodiment, step S40 includes the following steps:

[0104] Step S401, extracting dimension indicators from the financial instance sample data, the financial public data, and the financial report data to obtain financial industry dimension features and financial region dimension features;

[0105] In this embodiment, dimension indicators can be designed to extract dimension indicators from financial instance sample data, financial public data, and financial report data. The dimension characteristics of the financial industry include business types: such as banking, securities, insurance, etc.; financial indicators: profit margin, debt-to-asset ratio, ROE (return on equity), etc.; market performance: stock price, trading volume, price-earnings ratio, price-to-book ratio, etc. The dimension characteristics of financial regions include geographical location: country, province, city, etc.; economic development level: total GDP, per capita GDP, GDP growth rate, etc.; financial resource distribution: number of financial institutions, number of financial practitioners, financial market size, etc.; policy environment: local government financial policies, tax incentives, regulatory environment, etc.

[0106] Step S402, performing feature dimensionality upgrade on the financial industry dimension feature and the financial region dimension feature to obtain a financial industry dimensionality upgrade feature and a financial region dimensionality upgrade feature;

[0107] In this embodiment, the financial industry dimension features and financial region dimension features are upgraded by aggregation, derivation, crossover, etc. to increase the feature dimensions. Aggregation refers to merging multiple related or similar features into a new feature. The aggregation of financial industry dimension features includes: aggregating financial indicators (such as profit margin, debt-to-asset ratio) of different financial institutions, calculating the average value, median, etc. of the entire industry, and aggregating transaction data of different time periods, such as monthly transaction volume, quarterly transaction volume, etc.; the aggregation of financial region dimension features includes aggregating certain indicators of all financial institutions in the same region, such as total assets and total liabilities of financial institutions in the region, and calculating the overall performance of the financial market in the region, such as regional GDP growth rate, financial market transaction volume, etc. Derivation refers to calculating new features based on original features. The derivation of financial industry dimension features includes: calculating the ratio between financial indicators, such as the ratio of profit margin to total assets (ROE); differentiating time series data to obtain growth rate or change rate, such as monthly profit growth rate; using mathematical functions (such as logarithm, exponent, square root) to transform the original features. The derivation processing of the financial region dimension features includes: calculating the growth rate or change rate of the number of financial institutions in the region, and deriving the activity index of the regional financial market, such as the ratio of transaction volume to GDP. Crossover refers to combining two or more features to generate new features. The crossover processing of the financial industry dimension features includes: crossing the characteristics of the financial industry with the characteristics of the financial region. The crossover processing of the financial region dimension features includes: crossing the characteristics within the financial industry or within the financial region, such as "bank asset size*profit margin" or "city GDP*financial market transaction volume".

[0108] Step S403: input the financial industry dimension-raising features and the financial region dimension-raising features into a preset factor analysis model for training to obtain the multi-dimensional indicator aggregation model.

[0109] In this embodiment, before performing factor analysis, a suitability test is required to ensure that the data is suitable for factor analysis. The test methods include KMO (Kaiser-Meyer-Olkin) test and Bartlett's sphericity test. The closer the value of the KMO test is to 1, the more common factors there are between the variables, and the more suitable it is for factor analysis. The P value of the Bartlett's sphericity test should be less than the significance level (such as 0.05), indicating that there is a correlation between the variables and it is suitable for factor analysis.

[0110] This embodiment extracts dimensional indicators from the financial instance sample data, the financial public data, and the financial report data to obtain financial industry dimensional characteristics and financial region dimensional characteristics; performs feature dimensionality upgrade on the financial industry dimensional characteristics and the financial region dimensional characteristics to obtain upgraded dimensional characteristics of the financial industry and the financial region; inputs the upgraded dimensional characteristics of the financial industry and the financial region into a preset factor analysis model for training, thereby effectively obtaining a multidimensional indicator clustering model constructed from the financial instance sample data, the financial public data, and the financial report data, so as to facilitate subsequent clustering of enterprise advantage characteristics according to the multidimensional indicator clustering model.

[0111] Continue to refer Figure 5 In some optional implementations of this embodiment, step S50 includes the following steps:

[0112] Step S501, parsing the financial analysis data to obtain financial instance sample data;

[0113] In this embodiment, the financial analysis data may be queried according to the query information corresponding to the financial instance sample data, and the queried financial analysis data may be extracted to obtain the financial instance sample data.

[0114] Step S502, performing transaction flow analysis on the financial instance sample data to obtain transaction association information, and constructing a heterogeneous graph based on the transaction association information and the enterprise risk characteristics to obtain a transaction information heterogeneous graph;

[0115] In this embodiment, the steps of transaction flow analysis include transaction statistics and transaction correlation analysis, the processing of enterprise risk characteristics includes: risk indicator construction and risk assessment, and the steps of heterogeneous graph construction include defining nodes and edges, constructing heterogeneous graphs, and graph data analysis.

[0116] Step S503, obtaining the multidimensional analysis network, wherein the multidimensional analysis network includes a heterogeneous graph attention network and a heterogeneous graph Transformer network;

[0117] In this embodiment, the heterogeneous graph attention network and the heterogeneous graph Transformer network can be extracted from the database by analyzing the network extraction identifier, wherein the analysis network extraction identifier includes a first analysis network extraction identifier and a second analysis network extraction identifier, and the first analysis network extraction identifier and the second analysis network extraction identifier are unique identification information corresponding to the heterogeneous graph attention network and the heterogeneous graph Transformer network, respectively.

[0118] Step S504, inputting the transaction information heterogeneous graph into the heterogeneous graph attention network for analysis to obtain enterprise risk codes, and using the enterprise risk codes as the enterprise's own risk information;

[0119] In this embodiment, the enterprise risk code is the output result after the heterogeneous graph attention network analyzes the heterogeneous graph of transaction information, and the output result is a code representing the enterprise risk level and characteristics. The enterprise risk code includes risk level: enterprises are divided into different risk levels, such as low risk, medium risk, high risk, etc.; risk type: contains specific information about the risk type; key risk factors: key factors or paths that affect the enterprise risk level, including specific transaction patterns, associations with high-risk entities, abnormal financial activities, etc.

[0120] Step S505: input the transaction information heterogeneous graph into the heterogeneous graph Transformer network for analysis to obtain the enterprise-related risk code, and use the enterprise-related risk code as the enterprise-related risk information.

[0121] In this embodiment, the enterprise-related risk coding is the output result after the heterogeneous graph Transformer network analyzes the transaction information heterogeneous graph, and the output result represents the risk situation of other enterprises that have direct or indirect transaction relations with the enterprise. The enterprise-related risk coding includes risk correlation: an indicator indicating the degree of risk correlation between the enterprise and other enterprises in the network; risk propagation path: path information of risk propagation between enterprises through a specific transaction path; risk type and level: risk type and level information of associated enterprises.

[0122] This embodiment obtains financial instance sample data by parsing the financial analysis data; performs transaction flow analysis on the financial instance sample data to obtain transaction association information, and constructs a heterogeneous graph based on the transaction association information to obtain a transaction information heterogeneous graph; obtains the multidimensional analysis network, wherein the multidimensional analysis network includes a heterogeneous graph attention network and a heterogeneous graph Transformer network; inputs the transaction information heterogeneous graph into the heterogeneous graph attention network for analysis to obtain enterprise risk coding, and uses the enterprise risk coding as the enterprise's own risk information; inputs the transaction information heterogeneous graph into the heterogeneous graph Transformer network for analysis to obtain enterprise-related risk coding, and uses the enterprise-related risk coding as the enterprise-related risk information. In this way, the risk of the enterprise is effectively analyzed based on the financial analysis data, and the enterprise's own risk information and enterprise-related risk information are obtained, so as to facilitate the subsequent fusion based on the enterprise's own risk information and enterprise-related risk information.

[0123] Continue to refer Figure 6 In some optional implementations of this embodiment, step S60 includes the following steps:

[0124] Step S601, performing correlation analysis on the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain index correlation information;

[0125] In this embodiment, the correlation between each indicator and risk information can be quantified by calculating the Pearson coefficient between the industry advantage indicator, the regional advantage indicator, the enterprise's own risk information, and the enterprise-related risk information, and then the correlation between the indicator and the risk information is determined based on the quantified correlation to obtain the indicator correlation information.

[0126] Step S602, obtaining a preset weight allocation table, mapping the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to the weight allocation table to obtain index weight information;

[0127] In this embodiment, the preset weight allocation table includes industry advantage indicators, regional advantage indicators, enterprise risk information, enterprise-related risk information, and their corresponding weights, which can be preset values ​​based on historical data. The industry advantage indicators, regional advantage indicators, enterprise risk information, and enterprise-related risk information are mapped to the indicator weight allocation table to match and extract the corresponding weight values, thereby obtaining indicator weight information.

[0128] Step S603, according to the indicator association information and the indicator weight information, the industry advantage indicator, the regional advantage indicator, the enterprise's own risk information, and the enterprise-related risk information are integrated to obtain the enterprise characteristic indicator set.

[0129] In this embodiment, indicator association information is used to determine whether there is an association relationship between indicators and indicators, indicators and risk information, and risk information and risk information, and then the associated indicators and risk information are fused based on the association relationship. In this embodiment, the weighted average method is used to fuse each indicator and risk information, thereby effectively obtaining a set of enterprise characteristic indicators.

[0130] This embodiment obtains indicator correlation information by performing correlation analysis on the industry advantage indicators, the regional advantage indicators, the enterprise's own risk information, and the enterprise-related risk information; obtains a preset weight allocation table, maps the industry advantage indicators, the regional advantage indicators, the enterprise's own risk information, and the enterprise-related risk information to the weight allocation table, and obtains indicator weight information; and fuses the industry advantage indicators, the regional advantage indicators, the enterprise's own risk information, and the enterprise-related risk information according to the indicator correlation information, thereby effectively obtaining a set of enterprise characteristic indicators that integrates multiple indicators and risk information, so as to facilitate subsequent text generation processing.

[0131] Continue to refer Figure 7 In some optional implementations of this embodiment, step S70 includes the following steps:

[0132] Step S701, obtaining a text algorithm extraction identifier;

[0133] In this embodiment, the text algorithm extraction identifier is a unique information identifier corresponding to the summary text generation algorithm. The text algorithm extraction identifier can be generated according to a trigger event preset by the system. The trigger event can be triggered when the enterprise characteristic indicator set is generated.

[0134] Step S702, extracting a summary text generation algorithm from a database according to the text algorithm extraction identifier;

[0135] In this embodiment, a text algorithm extraction identifier is used to perform a matching query in an algorithm database to obtain a text generation algorithm.

[0136] Step S703, performing text conversion on the enterprise characteristic indicator set according to a preset natural language processing algorithm and the summary text generation algorithm to obtain the enterprise credit report summary.

[0137] In this embodiment, the preset natural language processing algorithm is an NLP algorithm, and the summary text generation algorithm can adopt a summary algorithm based on a Transformer model. The Transformer model can select a BERT model, and the NLP algorithm is used to perform word segmentation on the text of the enterprise feature indicator set to obtain standard enterprise feature indicator words, and then the enterprise feature indicator words are input into the trained BERT model for summary generation to obtain a summary of the enterprise credit report.

[0138] This embodiment obtains a text algorithm extraction identifier; extracts a summary text generation algorithm from a database according to the text algorithm extraction identifier; and performs text conversion on the enterprise characteristic indicator set according to a preset natural language processing algorithm and the summary text generation algorithm, thereby effectively obtaining a summary of the enterprise credit report to facilitate rapid understanding of the credit status of the enterprise.

[0139] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0140] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0141] Further references Figure 8 , as a response to the above Figure 1 In order to realize the method shown in the figure, the present application provides an embodiment of a credit report summary generation device, and the device embodiment is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0142] like Figure 8As shown, the credit report summary generating device 800 described in this embodiment includes: an information acquisition module 801, a model training module 802, a feature extraction module 803, an indicator acquisition module 804, a risk analysis module 805, an information fusion module 806, and a text generation module 807. Among them:

[0143] Information acquisition module 801, used to obtain enterprise credit index information;

[0144] Model training module 802, used to obtain financial analysis data, train a preset logistic regression model according to the financial analysis data, and obtain an enterprise feature classification model;

[0145] The feature extraction module 803 is used to input the enterprise credit index information into the enterprise feature classification model to obtain the enterprise advantage features and enterprise risk features;

[0146] An indicator acquisition module 804 is used to construct a model based on the financial analysis data to obtain a multi-dimensional indicator aggregation model, and input the enterprise advantage characteristics into the multi-dimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators;

[0147] The risk analysis module 805 is used to analyze the enterprise risk characteristics according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and enterprise-related risk information;

[0148] The information fusion module 806 is used to fuse the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain a set of enterprise characteristic indicators;

[0149] The text generation module 807 is used to generate text according to the enterprise characteristic indicator set to obtain the enterprise credit report summary.

[0150] This embodiment obtains enterprise credit index information; obtains financial analysis data, trains a preset logistic regression model according to the financial analysis data, and obtains an enterprise feature classification model; inputs the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; constructs a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and inputs the enterprise advantage features into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; analyzes the enterprise risk features according to a preset multidimensional analysis network to obtain enterprise risk information and enterprise-related risk information; integrates the industry advantage indicators, the regional advantage indicators, the enterprise risk information, and the enterprise-related risk information to obtain an enterprise feature indicator set; generates text according to the enterprise feature indicator set to obtain an enterprise credit report summary. Thus, an accurate and effective enterprise credit report summary is effectively generated based on the enterprise credit report, so as to facilitate a quick understanding of the credit status of the enterprise.

[0151] To solve the above technical problems, the present application also provides a computer device. Fig. 9 , Fig. 9 This is a basic structural block diagram of the computer device in this embodiment.

[0152] The computer device 9 includes a memory 91, a processor 92, and a network interface 93 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 9 with components 91-93, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0153] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0154] The memory 91 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 91 can be an internal storage unit of the computer device 9, such as a hard disk or memory of the computer device 9. In other embodiments, the memory 91 can also be an external storage device of the computer device 9, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 9. Of course, the memory 91 can also include both the internal storage unit of the computer device 9 and its external storage device. In this embodiment, the memory 91 is generally used to store the operating system and various application software installed on the computer device 9, such as computer-readable instructions of the credit report summary generation method, etc. In addition, the memory 91 can also be used to temporarily store various types of data that have been output or are to be output.

[0155] The processor 92 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 92 is generally used to control the overall operation of the computer device 9. In this embodiment, the processor 92 is used to run the computer-readable instructions stored in the memory 91 or process data, such as computer-readable instructions for running the credit report summary generation method.

[0156] The network interface 93 may include a wireless network interface or a wired network interface. The network interface 93 is generally used to establish a communication connection between the computer device 9 and other electronic devices.

[0157] By adopting the above-mentioned computer device, this embodiment can obtain enterprise credit index information; obtain financial analysis data, train a preset logistic regression model according to the financial analysis data, and obtain an enterprise feature classification model; input the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; construct a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and input the enterprise advantage features into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; analyze the enterprise risk features according to the preset multidimensional analysis network to obtain enterprise risk information and enterprise related risk information; fuse the industry advantage indicators, the regional advantage indicators, the enterprise risk information, and the enterprise related risk information to obtain an enterprise feature indicator set; generate text according to the enterprise feature indicator set to obtain an enterprise credit report summary. Thus, an accurate and effective enterprise credit report summary is effectively generated based on the enterprise credit report, so as to facilitate a quick understanding of the credit status of the enterprise.

[0158] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the credit report summary generation method as described above.

[0159] By adopting the above-mentioned computer-readable storage medium, this embodiment can obtain enterprise credit index information; obtain financial analysis data, train a preset logistic regression model according to the financial analysis data, and obtain an enterprise feature classification model; input the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; construct a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and input the enterprise advantage features into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; analyze the enterprise risk features according to a preset multidimensional analysis network to obtain enterprise risk information and enterprise-related risk information; fuse the industry advantage indicators, the regional advantage indicators, the enterprise risk information, and the enterprise-related risk information to obtain an enterprise feature indicator set; generate text according to the enterprise feature indicator set to obtain an enterprise credit report summary. In this way, an accurate and effective enterprise credit report summary is effectively generated based on the enterprise credit report, so as to facilitate a quick understanding of the credit status of the enterprise.

[0160] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0161] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for generating a credit report summary, characterized in that: include: Obtain enterprise credit index information; Acquire financial analysis data, and train a preset logistic regression model based on the financial analysis data to obtain an enterprise feature classification model; Inputting the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; Building a model based on the financial analysis data to obtain a multidimensional indicator aggregation model, and inputting the enterprise advantage characteristics into the multidimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; Analyze the risk characteristics of the enterprise according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and enterprise-related risk information; The industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information are integrated to obtain a set of enterprise characteristic indicators; A text is generated according to the enterprise characteristic indicator set to obtain a summary of the enterprise credit report.

2. The credit report summary generation method according to claim 1, characterized in that: The step of obtaining the enterprise credit index information specifically includes: Obtain corporate credit reports; Performing text recognition on the enterprise credit report to obtain enterprise credit text information; The enterprise credit investigation text information is preprocessed to obtain the enterprise credit investigation index information.

3. The credit report summary generation method according to claim 1, characterized in that: The financial analysis data includes financial instance sample data, financial public data, industry report data, and regional report data. The steps of obtaining the financial analysis data, training a preset logistic regression model according to the financial analysis data, and obtaining an enterprise feature classification model specifically include: Acquire the financial instance sample data, the financial public data, the industry report data, and the regional report data; Preprocessing the financial instance sample data, the financial public data, the industry report data, and the regional report data to obtain standard financial instance sample data, standard financial public data, standard industry report data, and standard regional report data; Extracting features from the standard financial instance sample data, the standard financial public data, the standard industry report data, and the standard regional report data to obtain full financial features, financial public features, industry report features, and regional report features; The total amount of financial features, the financial public features, the industry report features, and the regional report features are input into a preset classification regression model for training to obtain the enterprise feature classification model.

4. The credit report summary generation method according to claim 1, characterized in that: The financial analysis data includes financial instance sample data, financial public data, industry report data, and regional report data. The step of constructing a model based on the financial analysis data to obtain a multi-dimensional indicator aggregation model specifically includes: Extracting dimension indicators from the financial instance sample data, the financial public data, the industry report data, and the regional report data to obtain financial industry dimension features and financial regional dimension features; Performing feature dimensionality upgrade on the financial industry dimension feature and the financial region dimension feature to obtain the financial industry dimensionality upgrade feature and the financial region dimensionality upgrade feature; The financial industry dimensionality-raising features and the financial region dimensionality-raising features are input into a preset factor analysis model for training to obtain the multidimensional indicator aggregation model.

5. The credit report summary generation method according to claim 1, characterized in that: The step of analyzing the enterprise risk characteristics according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and the enterprise-related risk information specifically includes: Parsing the financial analysis data to obtain financial instance sample data; Performing transaction flow analysis on the sample data of the financial instance to obtain transaction association information, and constructing a heterogeneous graph based on the transaction association information and the enterprise risk characteristics to obtain a transaction information heterogeneous graph; Acquire the multidimensional analysis network, wherein the multidimensional analysis network includes a heterogeneous graph attention network and a heterogeneous graph Transformer network; Inputting the transaction information heterogeneous graph into the heterogeneous graph attention network for analysis to obtain enterprise risk coding, and using the enterprise risk coding as the enterprise's own risk information; The transaction information heterogeneous graph is input into the heterogeneous graph Transformer network for analysis to obtain enterprise-related risk codes, and the enterprise-related risk codes are used as the enterprise-related risk information.

6. The credit report summary generation method according to claim 1, characterized in that: The step of fusing the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain a set of enterprise characteristic indicators specifically includes: Performing correlation analysis on the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise's associated risk information to obtain indicator correlation information; Obtain a preset weight distribution table, map the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to the weight distribution table, and obtain indicator weight information; The industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information are integrated according to the indicator association information and the indicator weight information to obtain the enterprise characteristic indicator set.

7. The credit report summary generation method according to claim 1, characterized in that: The step of generating text according to the set of enterprise characteristic indicators to obtain a summary of the enterprise credit report specifically includes: Get the text algorithm extraction identifier; Extracting a summary text generation algorithm from a database according to the text algorithm extraction identifier; The enterprise characteristic indicator set is converted into text according to a preset natural language processing algorithm and the summary text generation algorithm to obtain the enterprise credit report summary.

8. A credit report summary generating device, characterized in that: include: Information acquisition module, used to obtain corporate credit index information; A model training module is used to obtain financial analysis data, train a preset logistic regression model according to the financial analysis data, and obtain an enterprise feature classification model; A feature extraction module, used to input the enterprise credit index information into the enterprise feature classification model to obtain enterprise advantage features and enterprise risk features; An indicator acquisition module is used to construct a model based on the financial analysis data to obtain a multi-dimensional indicator aggregation model, and input the enterprise advantage characteristics into the multi-dimensional indicator aggregation model to obtain industry advantage indicators and regional advantage indicators; A risk analysis module is used to analyze the risk characteristics of the enterprise according to a preset multi-dimensional analysis network to obtain the enterprise's own risk information and enterprise-related risk information; An information fusion module, used to fuse the industry advantage index, the regional advantage index, the enterprise's own risk information, and the enterprise-related risk information to obtain a set of enterprise characteristic indicators; The text generation module is used to generate text according to the enterprise characteristic indicator set to obtain the enterprise credit report summary.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the credit report summary generation method as described in any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the credit report summary generation method according to any one of claims 1 to 7.

Images