WO2020062660A1 - 企业信用风险评估方法、装置、设备及存储介质 - Google Patents

企业信用风险评估方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2020062660A1
WO2020062660A1 PCT/CN2018/123278 CN2018123278W WO2020062660A1 WO 2020062660 A1 WO2020062660 A1 WO 2020062660A1 CN 2018123278 W CN2018123278 W CN 2018123278W WO 2020062660 A1 WO2020062660 A1 WO 2020062660A1
Authority
WO
WIPO (PCT)
Prior art keywords
credit risk
enterprise
risk assessment
model
credit
Prior art date
Application number
PCT/CN2018/123278
Other languages
English (en)
French (fr)
Inventor
朱玺道
李泓格
陈姗婷
Original Assignee
深圳壹账通智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2020062660A1 publication Critical patent/WO2020062660A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/02—Banking, e.g. interest calculation or account maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities

Definitions

  • the present application relates to the technical field of big data analysis and processing, and in particular, to an enterprise credit risk assessment method, device, device, and storage medium.
  • the main purpose of this application is to provide an enterprise credit risk assessment method, device, equipment, and storage medium, which are aimed at solving the technical problems that the existing technology cannot accurately and effectively perform credit risk assessment for different enterprises.
  • this application provides a method for assessing corporate credit risk, which includes the following steps:
  • an enterprise credit risk assessment device which includes: an industry determination module, a model search module, a risk assessment module, and a report generation module;
  • the industry determining module is configured to obtain enterprise data of a target enterprise within a preset period of time, and determine an industry category to which the target enterprise belongs according to the enterprise data;
  • the model search module is used to find a credit risk assessment model corresponding to the industry category in a pre-built mapping relationship
  • the risk evaluation module is configured to extract credit risk characteristic variables from the enterprise data according to a preset data dimension, input the extracted credit risk characteristic variables to the credit risk evaluation model, and obtain an enterprise credit risk evaluation result;
  • the report generating module is configured to generate an enterprise credit risk assessment report corresponding to the target enterprise according to the enterprise credit risk assessment result.
  • the present application also proposes an enterprise credit risk assessment device, the device includes: a memory, a processor, and an enterprise credit risk assessment program stored in the memory and operable on the processor
  • the enterprise credit risk assessment program is configured to implement the steps of the enterprise credit risk assessment method as described above.
  • the present application also proposes a storage medium storing an enterprise credit risk assessment program on the storage medium, and the enterprise credit risk assessment program realizes the enterprise credit risk as described above when executed by a processor. Steps of the evaluation method.
  • the enterprise category of the target enterprise is determined based on the enterprise data by obtaining the enterprise data of the target enterprise within a preset period of time; it is searched in the pre-built mapping relationship Credit risk assessment model corresponding to industry category; extracting credit risk feature variables from corporate data according to preset data dimensions, and inputting the extracted credit risk feature variables into the credit risk assessment model to obtain corporate credit risk assessment results; according to corporate credit risk The evaluation results generate a corporate credit risk assessment report corresponding to the target company. Because the industry category of the enterprise is determined based on the enterprise data, the corresponding credit risk assessment model is found based on the determined industry category, and then the credit risk characteristics are extracted from the enterprise data. Variables, and input the extracted credit risk characteristic variables into the credit risk assessment model for credit risk assessment, thereby making the corporate credit risk assessment more targeted and ensuring that the credit risk assessment results have higher accuracy and reliability .
  • FIG. 1 is a schematic structural diagram of an enterprise credit risk assessment device for a hardware operating environment involved in a solution according to an embodiment of the present application;
  • FIG. 2 is a schematic flowchart of a first embodiment of an enterprise credit risk assessment method for an application
  • FIG. 3 is a schematic flowchart of a second embodiment of an enterprise credit risk assessment method for an application
  • FIG. 4 is a schematic flowchart of a third embodiment of a credit risk assessment method for an application enterprise
  • FIG. 5 is a structural block diagram of a first embodiment of an enterprise credit risk assessment device of the present application.
  • FIG. 1 is a schematic structural diagram of an enterprise credit risk assessment device for a hardware operating environment involved in a solution according to an embodiment of the present application.
  • the enterprise credit risk assessment device may include a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display, an input unit such as a keyboard, and the optional user interface 1003 may further include a standard wired interface and a wireless interface.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WIreless-FIdelity (WI-FI) interface).
  • WI-FI WIreless-FIdelity
  • the memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or may be a stable non-volatile memory (Non-Volatile Memory, NVM), such as a magnetic disk memory.
  • RAM Random Access Memory
  • NVM Non-Volatile Memory
  • the memory 1005 may optionally be a storage device independent of the foregoing processor 1001.
  • FIG. 1 does not constitute a limitation on the enterprise credit risk assessment equipment, and may include more or fewer components than shown in the figure, or combine certain components, or different component arrangements .
  • the memory 1005 as a storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and an enterprise credit risk assessment program.
  • the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 in the enterprise credit risk assessment device of the present application
  • the memory 1005 may be provided in an enterprise credit risk assessment device.
  • the enterprise credit risk assessment device invokes an enterprise credit risk assessment program stored in the memory 1005 through the processor 1001 and executes the enterprise credit risk assessment method provided in the embodiment of the present application.
  • FIG. 2 is a schematic flowchart of a first embodiment of an enterprise credit risk assessment method of the present application.
  • the method for assessing corporate credit risk includes the following steps:
  • Step S10 Obtain enterprise data of the target enterprise within a preset period, and determine the industry category to which the target enterprise belongs according to the enterprise data;
  • the execution subject of the method in this embodiment may be a computing service device with network communication, data processing, and program running functions, such as a mobile phone, tablet computer, personal computer, server, etc. (the server is used as the description below for description) .
  • the target enterprise is an enterprise that needs to perform corporate credit risk assessment.
  • the preset time period may be a pre-entered custom time period, such as June 30, 2015 to June 30, 2018.
  • the enterprise data includes Judicial, public opinion, industrial and commercial, financial, credit information, business management and other data. Accordingly, the industry categories described in this embodiment include, but are not limited to, real estate, non-banking, banking, service, manufacturing, and other industries.
  • the judicial data can obtain the judicial information of the enterprise through the judicial system, and extract data such as the amount involved in the judgment documents, whether the enterprise is involved in major economic disputes, etc .
  • public opinion data can search the public opinion of the enterprise through the network. Screen out whether the public opinion on the company is positive, intermediate, or negative over a period of time; data on industry and commerce can be used to obtain data such as whether the registered capital of the company has decreased, and changes in shareholders' shareholdings; financial data can be passed The company's financial statements can be obtained, and the company's profitability, debt service level, growth, etc. can be analyzed; credit information can be obtained from the credit information base database of the credit center of the People's Bank of China; experience management data can be Obtained through corporate research reports.
  • the server may obtain the enterprise data of the target enterprise within a preset period of time, and determine the industry category to which the target enterprise belongs according to the enterprise type, enterprise name, and trademark information in the enterprise data.
  • Step S20 find a credit risk assessment model corresponding to the industry category in a pre-built mapping relationship
  • mapping relationship or correspondence between the industry category and the credit risk assessment model corresponding to the industry category can also be established, so that the server can implement the mapping relationship based on the mapping relationship after determining the industry category to which the target enterprise belongs.
  • the mapping end source is an industry category
  • the target end source is a credit risk assessment model.
  • the server may find a credit risk assessment model corresponding to the industry category in a pre-built mapping relationship.
  • Step S30 extracting credit risk characteristic variables from the enterprise data according to a preset data dimension, inputting the extracted credit risk characteristic variables to the credit risk evaluation model, and obtaining an enterprise credit risk evaluation result;
  • preset data dimensions include, but are not limited to, data, extraction dimensions such as law, public opinion, business information, financial information, and / or credit information.
  • the credit risk characteristic variable is information data (including quantified data and non-quantified data) extracted from enterprise data according to the preset dimension.
  • information data including quantified data and non-quantified data
  • it can be quantified according to preset quantification standards and then used as a risk characteristic variable. , Positive, negative, neutral, etc.) positive and negative scoring of public opinion, and then count the number of major positive and negative public opinion corresponding to the enterprise over a period of time, so as to achieve the quantification of public opinion data.
  • the server may extract credit risk characteristic variables from the enterprise data according to data extraction dimensions such as law, public opinion, business information, financial information, and / or personal credit information, and then input the extracted credit risk characteristic variables to a pre-built Credit risk assessment model to obtain the corporate credit risk assessment results output by the credit risk assessment model.
  • data extraction dimensions such as law, public opinion, business information, financial information, and / or personal credit information
  • Step S40 Generate an enterprise credit risk assessment report corresponding to the target enterprise according to the enterprise credit risk assessment result.
  • the server can generate a corporate credit risk assessment report corresponding to the target enterprise based on the results. Specifically, the server can translate the corporate credit risk assessment results corresponding to The credit risk score is compared with a preset risk threshold, and whether the target company has credit risk is determined according to the comparison result, or the risk level of the credit risk score is determined according to a preset credit risk level, and then the target is determined. Whether the enterprise has credit risk; if there is credit risk, the risk data that causes the enterprise to have credit risk is screened from the enterprise data, and an enterprise credit risk assessment report corresponding to the target enterprise is generated based on these risk data.
  • This embodiment determines the industry category to which the target enterprise belongs by acquiring the enterprise data of the target enterprise within a preset period of time; finding the credit risk assessment model corresponding to the industry category in a pre-built mapping relationship; according to the preset data dimension
  • the credit risk characteristic variables are extracted from the corporate data, and the extracted credit risk characteristic variables are input to the credit risk assessment model to obtain the corporate credit risk assessment results; based on the corporate credit risk assessment results, a corporate credit risk assessment report corresponding to the target company is generated.
  • the risk assessment model conducts credit risk assessment, thereby making corporate credit risk assessment more targeted and ensuring that the credit risk assessment results have higher accuracy and reliability.
  • FIG. 3 is a schematic flowchart of a second embodiment of a credit risk assessment method for an application enterprise.
  • the method before the method for assessing corporate credit risk provided in this embodiment, before step S10, the method further includes:
  • Step S01 Obtain enterprise information of several enterprises, classify the enterprise information according to a preset industry category, and obtain industry information samples corresponding to each industry category;
  • the several enterprises may be enterprises or companies of different industry categories, and the preset industry categories may be six categories of real estate, non-banking, banking, service industry, manufacturing, and other industries.
  • the setting of the industry category is not limited in this embodiment.
  • the industry information sample contains enterprise information of several different enterprises in the industry.
  • the server may first perform user portraits for each enterprise according to the enterprise information to obtain corresponding enterprise credit risk portraits of each enterprise; and then according to the preset Industry categories are used to classify corporate credit risk portraits, so as to achieve classification of enterprises and enterprise information, and obtain industry information samples corresponding to each industry category.
  • the so-called portrait that is, the labeling of user (enterprise) information
  • a credit risk portrait of an enterprise is performed, that is, based on a large amount of enterprise data (such as laws, public opinion, business information, financial information, and / or credit information from the People's Bank, etc.), the basic situation and behavior model of the enterprise are performed. Comprehensive analysis to get the credit label of the enterprise.
  • the server obtains the enterprise information of several enterprises, classifies the enterprise information according to the preset industry categories such as real estate, non-banking, banking, service industry, manufacturing, and other industries, and obtains the corresponding information for each industry category.
  • Industry information samples are provided.
  • Step S02 extracting a credit risk characteristic variable from the industry information sample according to a preset data dimension, and establishing a credit risk assessment model corresponding to each industry category according to the extracted credit risk characteristic variable.
  • the server may extract the credit risk characteristic variables from the industry information samples according to a preset data dimension; and then discretize the extracted credit risk characteristic variables.
  • the variable factors are obtained by decomposition, and the variable factors are input to a preset neural network model for model training to obtain a credit risk assessment model corresponding to each industry category.
  • the initially trained models are often not very accurate. Therefore, in actual applications, it is necessary to optimize the initially trained models to improve the accuracy of the model output results.
  • the model optimization can be achieved by improving the accuracy when extracting credit risk characteristic variables (training data), that is, improving the accuracy or effectiveness of the training data. Therefore, in this embodiment, after obtaining the credit risk assessment model, the server can also select an effective combination of various variable factors according to the credit risk assessment model, thereby obtaining a higher-order correlation between the variable factors, thereby improving the characteristics of credit risk. Accuracy in variable extraction improves model performance.
  • the server in this embodiment may obtain all The variable information value corresponding to each variable factor in the initial credit risk assessment model; filtering the extracted credit risk characteristic variable according to the variable information value to obtain an effective characteristic variable and a data type corresponding to the effective characteristic variable;
  • a target credit risk characteristic variable is extracted from the industry information sample based on the data type; the target credit risk characteristic variable is discretized and decomposed and input to the recurrent neural network model for model training to obtain an effective credit risk assessment model.
  • the variable information value may be a variable coefficient corresponding to each variable factor.
  • the filtering of the extracted credit risk characteristic variable according to the variable information value may specifically be: variable information value corresponding to each variable factor. Compare with a preset threshold to obtain an effective variable factor whose variable information value is higher than the preset threshold; use the credit risk characteristic variable corresponding to the effective variable factor as the effective characteristic variable, and obtain data corresponding to the effective characteristic variable Types of.
  • variable coefficients corresponding to different variable factors in the model are not the same, such as the current asset liability ratio, the total number of significant negative public opinions in the past 30 days, and the corresponding variable coefficients are theoretically greater than the number of employees and companies.
  • the variable coefficient corresponding to a variable factor such as the number of shareholders, so in this embodiment, the server may compare the variable information value corresponding to each variable factor with a preset threshold, and then eliminate the variable factor that has a small impact on the output of the model according to the comparison result. To improve model calculation efficiency.
  • the preset neural network model in this embodiment is preferably a recurrent neural network model (also known as a recurrent neural network model) (Recurrent Neural Network) (RNN).
  • RNN Recurrent Neural Network
  • the credit risk assessment model can be split according to the model decomposition rules pre-configured by the staff to obtain each credit.
  • the credit risk assessment sub-model corresponding to the risk assessment model is to facilitate subsequent targeted optimization of each sub-model, so as to achieve the overall optimization of the credit risk assessment model and improve the evaluation accuracy.
  • the server may read a preset model decomposition rule from the database, and perform a model split on the credit risk assessment model corresponding to each industry category according to the model decomposition rule to obtain the credit risk assessment corresponding to each credit risk assessment model.
  • Sub-model weight allocation of the credit risk assessment sub-models corresponding to each credit risk assessment model to obtain preset weight values corresponding to each credit risk assessment sub-model.
  • the model decomposition rule can be configured to perform model splitting of credit risk assessment models in various industries according to dimensions such as finance, credit information and other categories, and decompose the credit risk assessment model into financial credit risk assessment models and credit
  • the enterprise information is obtained from several companies, and the enterprise information is classified according to a preset industry category to obtain a sample of industry information corresponding to each industry category; a credit risk characteristic variable is extracted from the sample of industry information according to a preset data dimension, And based on the extracted credit risk characteristic variables, credit risk assessment models corresponding to various industry categories are established to ensure that the established corporate credit risk assessment model has high accuracy.
  • FIG. 4 is a schematic flowchart of a third embodiment of an enterprise credit risk assessment method of the present application.
  • the credit risk assessment model includes multiple credit risk assessment sub-models.
  • step S30 in the method for assessing corporate credit risk provided in this embodiment may specifically include:
  • Step S301 extracting credit risk characteristic variables from the enterprise data according to a preset data dimension
  • the server may extract the credit risk characteristic variables from the enterprise data according to the data extraction dimensions such as law, public opinion, business information, financial information, and / or personal credit information.
  • Step S302 classify the credit risk characteristic variable according to a model category corresponding to the credit risk assessment sub-model, and obtain a classified credit risk characteristic variable;
  • each credit risk assessment submodel belongs to a different model category
  • Credit risk characteristic variables are classified to obtain classified credit risk characteristic variables. For example, credit risk characteristic variables such as industrial and commercial information and financial information are classified as financial credit risk assessment models, and credits such as legal and PBOC credit dimensions are classified.
  • the risk characteristic variables are classified as credit risk assessment models, and the credit risk characteristic variables of public opinion dimensions are classified as other types of credit risk assessment models.
  • Step S303 input the classified credit risk characteristic variables to the corresponding credit risk assessment sub-models, respectively;
  • the classified credit risk characteristic variables can be input to the corresponding credit risk assessment sub-models for credit risk assessment respectively.
  • Step S304 Obtain a credit risk score output by each credit risk assessment sub-model, and obtain an enterprise credit risk assessment result according to the credit risk score.
  • the server can obtain the corporate credit risk assessment results of the target enterprise by obtaining the credit risk scores output by each credit risk assessment sub-model.
  • the server may obtain the credit risk score output by each credit risk assessment sub-model, and query the database for a preset weight value corresponding to each credit risk assessment sub-model; and then, according to the preset weight value, use the following formula to compare Performing weighted summation on the credit risk score to obtain a summation result, and using the summation result as an enterprise credit risk assessment result;
  • S is the summation result
  • Y i is a credit risk score Credit Risk Assessment sub-model output
  • X i is a credit risk assessment preset weight value corresponding to the sub-model.
  • the preset weight value corresponding to the financial credit risk assessment model is 0.5
  • the preset weight value corresponding to the credit risk assessment model is 0.35
  • the preset weight value corresponding to other types of credit risk assessment model is 0.15
  • the credit risk score output by the financial credit risk assessment model is 80
  • the credit risk score output by the credit type credit risk assessment model is 75
  • the credit risk score output by other types of credit risk assessment models is 90
  • the server may find and compare with the enterprise data.
  • the target enterprise has an affiliated enterprise with an associated relationship, and queries in the database whether there is an affiliated enterprise credit risk assessment result corresponding to the affiliated enterprise, and if it exists, obtains the affiliated enterprise credit risk assessment result; and then according to the affiliated enterprise
  • the credit risk assessment result and the corporate credit risk assessment result corresponding to the target enterprise generate the corporate credit risk assessment report of the target enterprise.
  • the credit risk characteristic variables are extracted from the enterprise data according to a preset data dimension; the credit risk characteristic variables are classified according to the model category corresponding to the credit risk assessment sub-model, and the classified credit risk characteristic variables are obtained; The credit risk characteristic variables are input to the corresponding credit risk assessment sub-models; the credit risk scores output from each credit risk assessment sub-model are obtained, and the corporate credit risk assessment results are obtained based on the credit risk scores, which effectively improves the flexibility of corporate credit risk assessment And accuracy.
  • the above-mentioned storage media may be a read-only memory, a magnetic disk, or an optical disk.
  • FIG. 5 is a structural block diagram of a first embodiment of an enterprise credit risk assessment device of the present application.
  • the enterprise credit risk assessment device includes: an industry determination module 501, a model search module 502, a risk assessment module 503, and a report generation module 504;
  • the industry determination module 501 is configured to obtain enterprise data of a target enterprise within a preset period of time, and determine an industry category to which the target enterprise belongs according to the enterprise data;
  • the model searching module 502 is configured to find a credit risk assessment model corresponding to the industry category in a pre-built mapping relationship
  • the risk evaluation module 503 is configured to extract credit risk characteristic variables from the enterprise data according to a preset data dimension, input the extracted credit risk characteristic variables to the credit risk evaluation model, and obtain an enterprise credit risk evaluation result;
  • the report generating module 504 is configured to generate an enterprise credit risk assessment report corresponding to the target enterprise according to the enterprise credit risk assessment result.
  • This embodiment determines the industry category to which the target enterprise belongs by acquiring the enterprise data of the target enterprise within a preset period of time; finding the credit risk assessment model corresponding to the industry category in a pre-built mapping relationship; according to the preset data dimension
  • the credit risk characteristic variables are extracted from the corporate data, and the extracted credit risk characteristic variables are input to the credit risk assessment model to obtain the corporate credit risk assessment results; based on the corporate credit risk assessment results, a corporate credit risk assessment report corresponding to the target company is generated.
  • the risk assessment model conducts credit risk assessment, thereby making corporate credit risk assessment more targeted and ensuring that the credit risk assessment results have higher accuracy and reliability.
  • the risk assessment module 503 is further configured to extract a credit risk characteristic variable from the enterprise data according to a preset data dimension; and to the credit risk according to a model category corresponding to the credit risk assessment sub-model. Classify the characteristic variables to obtain the classified credit risk characteristic variables; input the classified credit risk characteristic variables to the corresponding credit risk assessment sub-models separately; obtain the credit risk scores output by each credit risk assessment sub-model, and The credit risk score is used to obtain the corporate credit risk assessment results.
  • the risk assessment module 503 is further configured to obtain a credit risk score output by each credit risk assessment sub-model, and query a database for a preset weight value corresponding to each credit risk assessment sub-model; according to the preset weight Value, the weighted summation of the credit risk score is obtained by the following formula, and the summation result is used as the corporate credit risk assessment result;
  • S is the summation result
  • Y i is a credit risk score Credit Risk Assessment sub-model output
  • X i is a credit risk assessment preset weight value corresponding to the sub-model.
  • the enterprise credit risk assessment device of this embodiment further includes a model establishment module, which is used to obtain enterprise information of several enterprises, classify the enterprise information according to a preset industry category, and obtain various industry categories Corresponding industry information samples; Credit risk characteristic variables are extracted from the industry information samples according to a preset data dimension, and a credit risk assessment model corresponding to each industry category is established based on the extracted credit risk characteristic variables.
  • a model establishment module which is used to obtain enterprise information of several enterprises, classify the enterprise information according to a preset industry category, and obtain various industry categories Corresponding industry information samples; Credit risk characteristic variables are extracted from the industry information samples according to a preset data dimension, and a credit risk assessment model corresponding to each industry category is established based on the extracted credit risk characteristic variables.
  • model building module is further configured to extract a credit risk characteristic variable from the industry information sample according to a preset data dimension; discretely decompose the extracted credit risk characteristic variable to obtain a variable factor, and
  • the above-mentioned variable factors are input to a preset neural network model for model training, and a credit risk assessment model corresponding to each industry category is obtained.
  • the model building module is further configured to read a preset model decomposition rule from a database, and perform a model split on a credit risk assessment model corresponding to each industry category according to the model decomposition rule to obtain each credit risk.
  • the risk assessment module 503 is further configured to find an affiliated company that has an associated relationship with the target enterprise according to the enterprise data, and query in the database whether there is an affiliated enterprise credit risk corresponding to the affiliated company.
  • the methods in the above embodiments can be implemented by means of software plus a necessary universal hardware platform, and of course, also by hardware, but in many cases the former is better.
  • Implementation Based on such an understanding, the technical solution of the present application, in essence, or a part that contributes to the existing technology, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access)
  • the memory, the magnetic disk, and the optical disc include several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in the embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Educational Administration (AREA)
  • Computational Linguistics (AREA)
  • Game Theory and Decision Science (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Technology Law (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

本申请涉及大数据分析处理领域，公开一种企业信用风险评估方法、装置、设备及存储介质，该方法包括：获取目标企业在预设时段内的企业数据，根据企业数据确定出目标企业所属的行业类别；查找行业类别对应的信用风险评估模型；按预设数据维度从企业数据中提取信用风险特征变量并将其输入至信用风险评估模型，获得企业信用风险评估结果；然后根据企业信用风险评估结果生成企业信用风险评估报告。

Description

企业信用风险评估方法、装置、设备及存储介质
本申请要求于2018年09月27日提交中国专利局、申请号为201811135377.2、发明名称为“企业信用风险评估方法、装置、设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及大数据分析处理技术领域，尤其涉及一种企业信用风险评估方法、装置、设备及存储介质。
背景技术
目前，部分金融机构在对企业进行信用风险评估时，对所有的企业都采用相同或相似的评估模型来进行，但实际上企业所属行业的不同，其经营模式、业务范围、资产配置等企业数据都不相同，笼统的将不同企业的企业数据输入到同一个评估模型进行信用风险评估就会导致最终获得的评估结果准确性较低，严重时甚至会出现评估错误的情况。因此，如何准确有效的对不同的企业进行信用风险评估，是一个亟待解决的问题。
发明内容
本申请的主要目的在于提供了一种企业信用风险评估方法、装置、设备及存储介质，旨在解决现有技术无法准确有效的对不同的企业进行信用风险评估的技术问题。
为实现上述目的，本申请提供了一种企业信用风险评估方法，所述方法包括以下步骤:
获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
此外，为实现上述目的，本申请还提出一种企业信用风险评估装置，所述装置包括：行业确定模块、模型查找模块、风险评估模块和报告生成模块；
其中，所述行业确定模块，用于获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
所述模型查找模块，用于在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
所述风险评估模块，用于按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
所述报告生成模块，用于根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
此外，为实现上述目的，本申请还提出一种企业信用风险评估设备，所述设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的企业信用风险评估程序，所述企业信用风险评估程序配置为实现如上文所述的企业信用风险评估方法的步骤。
此外，为实现上述目的，本申请还提出一种存储介质，所述存储介质上存储有企业信用风险评估程序，所述企业信用风险评估程序被处理器执行时实现如上文所述的企业信用风险评估方法的步骤。
本实施例的企业信用风险评估方法、装置、设备及存储介质，通过获取目标企业在预设时段内的企业数据，根据企业数据确定出目标企业所属的行业类别；在预先构建的映射关系中查找行业类别对应的信用风险评估模型；按预设数据维度从企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至信用风险评估模型，获得企业信用风险评估结果；根据企业信用风险评估结果生成目标企业对应的企业信用风险评估报告，由于是先根据企业数据确定企业所属的行业类别，然后根据确定出的行业类别查找对应的信用风险评估模型，再从企业数据中提取信用风险特征变量，并将提取到的信用风险特征 变量输入至信用风险评估模型进行信用风险评估，从而使得企业信用风险评估更具有针对性，确保了信用风险评估结果具有较高的准确性及可靠性。
附图说明
图1是本申请实施例方案涉及的硬件运行环境的企业信用风险评估设备的结构示意图；
图2为本申请企业信用风险评估方法第一实施例的流程示意图；
图3为本申请企业信用风险评估方法第二实施例的流程示意图；
图4为本申请企业信用风险评估方法第三实施例的流程示意图；
图5为本申请企业信用风险评估装置第一实施例的结构框图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。
参照图1，图1为本申请实施例方案涉及的硬件运行环境的企业信用风险评估设备结构示意图。
如图1所示，该企业信用风险评估设备可以包括：处理器1001，例如中央处理器(Central Processing Unit，CPU)，通信总线1002、用户接口1003，网络接口1004，存储器1005。其中，通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏(Display)、输入单元比如键盘(Keyboard)，可选用户接口1003还可以包括标准的有线接口、无线接口。网络接口1004可选的可以包括标准的有线接口、无线接口(如无线保真(WIreless-FIdelity，WI-FI)接口)。存储器1005可以是高速的随机存取存储器(Random Access Memory，RAM)存储器，也可以是稳定的非易失性存储器(Non-Volatile Memory，NVM)，例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解，图1中示出的结构并不构成对企业信用风险评估设备的限定，可以包括比图示更多或更少的部件，或者组 合某些部件，或者不同的部件布置。
如图1所示，作为一种存储介质的存储器1005中可以包括操作系统、数据存储模块、网络通信模块、用户接口模块以及企业信用风险评估程序。
在图1所示的企业信用风险评估设备中，网络接口1004主要用于与网络服务器进行数据通信；用户接口1003主要用于与用户进行数据交互；本申请企业信用风险评估设备中的处理器1001、存储器1005可以设置在企业信用风险评估设备中，所述企业信用风险评估设备通过处理器1001调用存储器1005中存储的企业信用风险评估程序，并执行本申请实施例提供的企业信用风险评估方法。
本申请实施例提供了一种企业信用风险评估方法，参照图2，图2为本申请企业信用风险评估方法第一实施例的流程示意图。
本实施例中，所述企业信用风险评估方法包括以下步骤：
步骤S10：获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
需要说明的是，本实施例方法的执行主体可以是具有网络通信、数据处理以及程序运行功能的计算服务设备，如手机、平板电脑、个人电脑、服务器等(以下以服务器为执行主体进行说明)。所述目标企业为需要进行企业信用风险评估的企业，所述预设时间段可以是预先输入的自定义时间段，例如2015年6月30日-2018年6月30日，所述企业数据包括企业的司法、舆情、工商、财务、征信、经营管理等方面的数据。相应地，本实施例中所述行业类别包括但不限于房地产、非银、银行业、服务业、制造业以及其它行业。
其中，司法数据可通过司法系统获取企业的司法信息，并从中提取裁判文书中的涉诉金额、企业是否涉及重大经济类纠纷等数据信息；舆情数据可通过网络途径搜索企业在社会上的舆情，筛选出一段时期内公众对于该企业发生的舆情是积极、中级还是消极情绪；工商数据可通过工商系统获取企业的注册资本是否有减少、股东持股比例的变化情况等数据信息；财务数据可通过企业财务报表来获得，并以此来分析企业的盈利水平、偿债水平、成长性等；征信数据可通过中国人 民银行征信中心的信用信息基础数据库获取企业征信情况；经验管理数据可通过企业调研报告获得。
在具体实现中，服务器可获取目标企业在预设时段内的企业数据，根据企业数据中的企业类型、企业名称、商标等信息确定出所述目标企业所属的行业类别。
步骤S20：在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
需要说明的是，执行本步骤之前，需要根据不同行业企业的企业数据先构建用于对不同行业的企业进行信用风险评估的(行业)信用风险评估模型，且服务器在构建好各个行业对应的信用风险评估模型后，还可建立一个行业类别和行业类别对应的信用风险评估模型之间的映射关系或对应关系，以便于服务器能够在确定出目标企业所属的行业类别后，根据所述映射关系实现对目标信用风险评估模型的快速确定并获取。在所述映射关系中，映射端源为行业类别，目标端源为信用风险评估模型。
在具体实现中，服务器在确定出到目标企业所属的行业类别后，可在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型。
步骤S30：按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
需要说明的是，所述预设数据维度包括但不限于法律、舆情、工商信息、财务信息和/或人行征信等数据提取维度。
应理解的是，所述信用风险特征变量即按所述预设维度从企业数据中提取出的信息数据(包括量化数据及非量化数据)。对于非量化数据，可根据预设量化标准对其进行量化后再作为风险特征变量，例如对于企业的舆情数据可通过网络途径获取企业在社会上的舆情，然后按照预先设定的打分标准(例如，利好、利空、中性等)对舆情进行正负面打分，再统计一段时间内企业对应的重大正负面舆情的数量，从而实现对舆情数据的量化。
在具体实现中，服务器可按照法律、舆情、工商信息、财务信息和/或人行征信等数据提取维度从企业数据中提取信用风险特征变量，然后将提取到的信用风险特征变量输入至预先构建的信用风险评估模型，获得信用风险评估模型输出的企业信用风险评估结果。
步骤S40：根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
在具体实现中，服务器在获取到信用风险评估模型输出的企业信用风险评估结果后，可根据该结果生成目标企业对应的企业信用风险评估报告，具体的，服务器可将企业信用风险评估结果对应的信用风险评分与预设风险阈值进行比较，根据比较结果来判断目标企业是否存在信用风险，又或是根据预先设定的信用风险等级来确定所述信用风险评分处于何种风险等级，进而判断目标企业是否存在信用风险；若存在信用风险则从企业数据中筛选出导致企业存在信用风险的风险数据，并根据这些风险数据生成目标企业对应的企业信用风险评估报告。
本实施例通过获取目标企业在预设时段内的企业数据，根据企业数据确定出目标企业所属的行业类别；在预先构建的映射关系中查找行业类别对应的信用风险评估模型；按预设数据维度从企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至信用风险评估模型，获得企业信用风险评估结果；根据企业信用风险评估结果生成目标企业对应的企业信用风险评估报告，由于是先根据企业数据确定企业所属的行业类别，然后根据确定出的行业类别查找对应的信用风险评估模型，再从企业数据中提取信用风险特征变量，并将提取到的信用风险特征变量输入至信用风险评估模型进行信用风险评估，从而使得企业信用风险评估更具有针对性，确保了信用风险评估结果具有较高的准确性及可靠性。
参考图3，图3为本申请企业信用风险评估方法第二实施例的流程示意图。
基于上述第一实施例，在本实施例提供的企业信用风险评估方法在所述步骤S10之前，还包括：
步骤S01：获取若干个企业的企业信息，按预设行业类别对所述企业信息进行分类，获得各行业类别对应的行业信息样本；
需要说明的是，所述若干个企业可以是不同行业类别的企业或公司，所述预设行业类别可以是房地产、非银、银行业、服务业、制造业以及其它行业六大类，当然具体行业类别的设定本实施例不作限制。所述行业信息样本中包含有若干个本行业不同企业的企业信息。
此外，在本实施例中，服务器在获取到若干个企业的企业信息后，可先根据企业信息为各企业进行用户画像，以获得各企业对应的企业信用风险画像；然后再根据所述预设行业类别对企业信用风险画像进行分类，从而实现对企业以及企业信息的分类，获得各行业类别对应的行业信息样本。
应理解的是，所谓画像，即用户(企业)信息标签化，就是通过收集与分析用户社会属性、生活习惯、消费行为等主要信息的数据之后，完美地抽象出一个用户的商业全貌。在本实施例中对企业进行信用风险画像，即在大量企业数据(如法律、舆情、工商信息、财务信息和/或人行征信等)信息的基础上对企业的基本情况、行为模式等进行综合分析，得到企业的信用标签。
在具体实现中，服务器获取若干个企业的企业信息，按预先设定的房地产、非银、银行业、服务业、制造业以及其它行业等行业类别对企业信息进行分类，获得各行业类别对应的行业信息样本。
步骤S02：按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型。
在具体实现中，服务器在获得各行业类别对应的行业信息样本后，可按预设数据维度从所述行业信息样本中提取出信用风险特征变量；然后对提取出的信用风险特征变量进行离散化分解获得变量因子，并将所述变量因子输入至预设神经网络模型进行模型训练，获得各行业类别对应的信用风险评估模型。
可以理解的是，最初训练出的模型往往准确性不高，因此在实际应用中有必要对最初训练出的模型进行优化以提高模型输出结果的 准确性。具体的，可通过提高对信用风险特征变量(训练数据)提取时的精准度来实现模型优化，即提高训练数据的准确度或有效性。因此，本实施例中服务器在获得信用风险评估模型后，还可以根据信用风险评估模型来挑选各变量因子的有效组合，由此获得变量因子之间高阶的关联关系，从而提高对信用风险特征变量提取时的精准度，提升模型表现。
进一步地，为保证提取出的信用风险特征变量具有较高的精准度，避免提取过多对信用风险评估影响程度较小的特征变量，导致服务器的计算量增加，本实施例中服务器可获取所述初始信用风险评估模型中各变量因子对应的变量信息值；根据所述变量信息值对所述提取出信用风险特征变量的进行筛选，获取有效特征变量以及所述有效特征变量对应的数据类型；基于所述数据类型从所述行业信息样本中提取目标信用风险特征变量；将所述目标信用风险特征变量离散化分解后输入至所述递归神经网络模型进行模型训练，获得有效信用风险评估模型。
其中，所述变量信息值可以是各变量因子对应的变量系数；所述根据所述变量信息值对所述提取出信用风险特征变量的进行筛选具体可以是：将各变量因子对应的变量信息值与预设阈值进行比较，获取变量信息值高于所述预设阈值的有效变量因子；将所述有效变量因子对应的信用风险特征变量作为有效特征变量，并获取所述有效特征变量对应的数据类型。
应理解的是，在模型中不同的变量因子对应的变量系数并不相同，例如流动资产负债率、过去30天重大负面舆情总数这类变量因子对应的变量系数理论上要大于企业员工数量、企业股东人数这类变量因子对应的变量系数，因此本实施例中，服务器可将各变量因子对应的变量信息值与预设阈值进行比较，然后根据比较结果剔除对模型输出结果影响较小的变量因子，以提高模型计算效率。
考虑到本实施例提出的企业信用风险评估方法需要将筛选出的信用风险特征变量与各信用风险特征变量对应的好坏标签一并输入到模型中进行训练，且各信用风险特征变量具有一定的关联性，彼此 并非相互独立，本实施例中所述预设神经网络模型优选为递归神经网络模型，(又称循环神经网络模型)(Recurrent Neural Network，RNN)。
应理解的是，在对企业进行信用风险评估时，由于涉及的企业数据种类繁多、类目繁杂，若将所有的企业数据汇聚到某一单个模型中进行训练，并不利于后续的模型优化操作，因此，本实施例提供的企业信用风险评估方法在构建出各行业对应的信用风险评估模型后，还可以按照工作人员预先配置的模型分解规则来对信用风险评估模型进行拆分，获得各信用风险评估模型对应的信用风险评估子模型，以便于后续有针对性的对各子模型进行优化，从而实现对信用风险评估模型的整体优化，提高评估准确率。
具体的，服务器可从数据库中读取预先设定的模型分解规则，根据所述模型分解规则对各行业类别对应的信用风险评估模型进行模型拆分，获得各信用风险评估模型对应的信用风险评估子模型；分别对各信用风险评估模型对应的信用风险评估子模型进行权重配置，得到各信用风险评估子模型对应的预设权重值。
其中，所述模型分解规则可配置为按照财务类、征信类和其它类等维度对各行业信用风险评估模型进行模型拆分，将信用风险评估模型分解为财务类信用风险评估模型、征信类信用风险评估模型和其它类信用风险评估模型等三个子模型，并为各子模型配置相应的预设权重值。
本实施例通过获取若干个企业的企业信息，按预设行业类别对企业信息进行分类，获得各行业类别对应的行业信息样本；按预设数据维度从行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型，保证了建立的企业信用风险评估模型具有较高的准确度。
参考图4，图4为本申请企业信用风险评估方法第三实施例的流程示意图。
基于上述各实施例，在本实施例中，所述信用风险评估模型包括多个信用风险评估子模型。
相应地，本实施例提供的企业信用风险评估方法中所述步骤S30 可具体包括：
步骤S301：按预设数据维度从所述企业数据中提取信用风险特征变量；
在具体实现中，服务器可按照法律、舆情、工商信息、财务信息和/或人行征信等数据提取维度从企业数据中提取信用风险特征变量。
步骤S302：根据所述信用风险评估子模型对应的模型类别对所述信用风险特征变量进行分类，获得分类后的信用风险特征变量；
应理解的是，由于各信用风险评估子模型所属的模型类别不同，因此服务器在从企业数据中提取出信用风险特征变量时，需要根据预先拆分好的信用风险评估子模型对应的模型类别对信用风险特征变量进行分类，获得分类后的信用风险特征变量，例如将工商信息、财务信息等维度的信用风险特征变量归类为财务类信用风险评估模型，将法律、人行征信等维度的信用风险特征变量归类为征信类信用风险评估模型，将舆情维度的信用风险特征变量归类为其它类信用风险评估模型等。
步骤S303：将所述分类后的信用风险特征变量分别输入至对应的信用风险评估子模型；
在具体实现中，服务器在完成对信用风险特征变量的分类后，即可将分类后的信用风险特征变量分别输入至对应的信用风险评估子模型进行信用风险评估。
步骤S304：获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果。
在具体实现中，服务器可通过获取各信用风险评估子模型输出的信用风险评分来获得目标企业的企业信用风险评估结果。
具体的，服务器可通过获取各信用风险评估子模型输出的信用风险评分，并在数据库中查询各信用风险评估子模型对应的预设权重值；然后根据所述预设权重值，通过下式对所述信用风险评分进行加权求和获得求和结果，并将所述求和结果作为企业信用风险评估结果；
Figure PCTCN2018123278-appb-000001
式中，S为求和结果，Y i为信用风险评估子模型输出的信用风险评分，X i为信用风险评估子模型对应的预设权重值。
例如，若财务类信用风险评估模型对应的预设权重值为0.5，征信类信用风险评估模型对应的预设权重值为0.35，其它类信用风险评估模型对应的预设权重值为0.15，那么当财务类信用风险评估模型输出的信用风险评分为80、征信类信用风险评估模型输出的信用风险评分为75、其它类信用风险评估模型输出的信用风险评分为90时，对信用风险评分进行加权求和的求和结果则为80*0.5+75*0.35+90*0.15＝79.75。
进一步地，考虑到企业与企业之间大多会存在一定的业务往来或直接投资、债务、股东等关联性。因此一个企业出现风险，其关联企业也可能会受到影响，因此考虑到企业之间的关联性，为进一步提高信用风险评估的准确性，本实施例中服务器可根据所述企业数据查找与所述目标企业存在关联关系的关联企业，并在所述数据库中查询是否存在所述关联企业对应的关联企业信用风险评估结果，若存在则获取所述关联企业信用风险评估结果；然后根据所述关联企业信用风险评估结果以及所述目标企业对应的企业信用风险评估结果生成所述目标企业的企业信用风险评估报告。
本实施例通过按预设数据维度从企业数据中提取信用风险特征变量；根据信用风险评估子模型对应的模型类别对信用风险特征变量进行分类，获得分类后的信用风险特征变量；将分类后的信用风险特征变量分别输入至对应的信用风险评估子模型；获取各信用风险评估子模型输出的信用风险评分，根据信用风险评分获得企业信用风险评估结果，有效地提高了企业信用风险评估的灵活性和准确性。
需要说明的是，本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
参照图5，图5为本申请企业信用风险评估装置第一实施例的结 构框图。
如图5所示，本申请实施例提出的企业信用风险评估装置包括：行业确定模块501、模型查找模块502、风险评估模块503和报告生成模块504；
其中，所述行业确定模块501，用于获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
所述模型查找模块502，用于在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
所述风险评估模块503，用于按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
所述报告生成模块504，用于根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
本实施例通过获取目标企业在预设时段内的企业数据，根据企业数据确定出目标企业所属的行业类别；在预先构建的映射关系中查找行业类别对应的信用风险评估模型；按预设数据维度从企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至信用风险评估模型，获得企业信用风险评估结果；根据企业信用风险评估结果生成目标企业对应的企业信用风险评估报告，由于是先根据企业数据确定企业所属的行业类别，然后根据确定出的行业类别查找对应的信用风险评估模型，再从企业数据中提取信用风险特征变量，并将提取到的信用风险特征变量输入至信用风险评估模型进行信用风险评估，从而使得企业信用风险评估更具有针对性，确保了信用风险评估结果具有较高的准确性及可靠性。
基于本申请上述企业信用风险评估装置第一实施例，提出本申请企业信用风险评估装置的第二实施例。
在本实施例中，所述风险评估模块503，还用于按预设数据维度从所述企业数据中提取信用风险特征变量；根据所述信用风险评估子模型对应的模型类别对所述信用风险特征变量进行分类，获得分类后 的信用风险特征变量；将所述分类后的信用风险特征变量分别输入至对应的信用风险评估子模型；获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果。
进一步地，所述风险评估模块503，还用于获取各信用风险评估子模型输出的信用风险评分，并在数据库中查询各信用风险评估子模型对应的预设权重值；根据所述预设权重值，通过下式对所述信用风险评分进行加权求和获得求和结果，并将所述求和结果作为企业信用风险评估结果；
Figure PCTCN2018123278-appb-000002
式中，S为求和结果，Y i为信用风险评估子模型输出的信用风险评分，X i为信用风险评估子模型对应的预设权重值。
进一步地，本实施例企业信用风险评估装置还包括模型建立模块，所述模型建立模块，用于获取若干个企业的企业信息，按预设行业类别对所述企业信息进行分类，获得各行业类别对应的行业信息样本；按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型。
进一步地，所述模型建立模块，还用于按预设数据维度从所述行业信息样本中提取出信用风险特征变量；对提取出的信用风险特征变量进行离散化分解获得变量因子，并将所述变量因子输入至预设神经网络模型进行模型训练，获得各行业类别对应的信用风险评估模型。
进一步地，所述模型建立模块，还用于从数据库中读取预先设定的模型分解规则，根据所述模型分解规则对各行业类别对应的信用风险评估模型进行模型拆分，获得各信用风险评估模型对应的风险评估子模型；分别对各信用风险评估模型对应的风险评估子模型进行权重配置，得到各信用风险评估子模型对应的预设权重值。
进一步地，所述风险评估模块503，还用于根据所述企业数据查找与所述目标企业存在关联关系的关联企业，并在所述数据库中查询是否存在所述关联企业对应的关联企业信用风险评估结果；若存在， 则获取所述关联企业信用风险评估结果；相应地，所述报告生成模块504，还用于根据所述关联企业信用风险评估结果以及所述目标企业对应的企业信用风险评估结果生成所述目标企业的企业信用风险评估报告。
本申请企业信用风险评估装置的其他实施例或具体实现方式可参照上述各方法实施例，此处不再赘述。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者系统不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者系统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如只读存储器/随机存取存储器、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种企业信用风险评估方法，其特征在于，所述方法包括：
    获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
    在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
    按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
    根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
  2. 如权利要求1所述的方法，其特征在于，所述信用风险评估模型包括多个信用风险评估子模型；
    所述按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果的步骤包括：
    按预设数据维度从所述企业数据中提取信用风险特征变量；
    根据所述信用风险评估子模型对应的模型类别对所述信用风险特征变量进行分类，获得分类后的信用风险特征变量；
    将所述分类后的信用风险特征变量分别输入至对应的信用风险评估子模型；
    获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果。
  3. 如权利要求2所述的方法，其特征在于，所述获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果的步骤，包括：
    获取各信用风险评估子模型输出的信用风险评分，并在数据库中查询各信用风险评估子模型对应的预设权重值；
    根据所述预设权重值，通过下式对所述信用风险评分进行加权求 和获得求和结果，并将所述求和结果作为企业信用风险评估结果；
    Figure PCTCN2018123278-appb-100001
    式中，S为求和结果，Y i为信用风险评估子模型输出的信用风险评分，X i为信用风险评估子模型对应的预设权重值。
  4. 如权利要求1所述的方法，其特征在于，所述获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别的步骤之前，所述方法还包括：
    获取若干个企业的企业信息，按预设行业类别对所述企业信息进行分类，获得各行业类别对应的行业信息样本；
    按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型。
  5. 如权利要求4所述的方法，其特征在于，所述按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型的步骤，包括：
    按预设数据维度从所述行业信息样本中提取出信用风险特征变量；
    对提取出的信用风险特征变量进行离散化分解获得变量因子，并将所述变量因子输入至预设神经网络模型进行模型训练，获得各行业类别对应的信用风险评估模型。
  6. 如权利要求5所述的方法，其特征在于，所述获得各行业类别对应的信用风险评估模型的步骤之后，所述方法还包括：
    从数据库中读取预先设定的模型分解规则，根据所述模型分解规则对各行业类别对应的信用风险评估模型进行模型拆分，获得各信用风险评估模型对应的信用风险评估子模型；
    分别对各信用风险评估模型对应的信用风险评估子模型进行权重配置，得到各信用风险评估子模型对应的预设权重值。
  7. 如权利要求1所述的方法，其特征在于，所述根据所述企业 信用风险评估结果生成所述目标企业对应的企业信用风险评估报告的步骤之前，所述方法还包括：
    根据所述企业数据查找与所述目标企业存在关联关系的关联企业，并在所述数据库中查询是否存在所述关联企业对应的关联企业信用风险评估结果；
    若存在，则获取所述关联企业信用风险评估结果；
    所述根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告的步骤，包括：
    根据所述关联企业信用风险评估结果以及所述目标企业对应的企业信用风险评估结果生成所述目标企业的企业信用风险评估报告。
  8. 一种企业信用风险评估装置，其特征在于，所述装置包括：行业确定模块、模型查找模块、风险评估模块和报告生成模块；
    其中，所述行业确定模块，用于获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
    所述模型查找模块，用于在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
    所述风险评估模块，用于按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
    所述报告生成模块，用于根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
  9. 一种企业信用风险评估设备，其特征在于，所述企业信用风险评估设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的企业信用风险评估程序，所述企业信用风险评估程序配置为实现如以下步骤：
    获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
    在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
    按预设数据维度从所述企业数据中提取信用风险特征变量，将提 取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
    根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
  10. 如权利要求9所述的企业信用风险评估设备，其特征在于，所述信用风险评估模型包括多个信用风险评估子模型；
    所述按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果的步骤包括：
    按预设数据维度从所述企业数据中提取信用风险特征变量；
    根据所述信用风险评估子模型对应的模型类别对所述信用风险特征变量进行分类，获得分类后的信用风险特征变量；
    将所述分类后的信用风险特征变量分别输入至对应的信用风险评估子模型；
    获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果。
  11. 如权利要求10所述的企业信用风险评估设备，其特征在于，所述获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果的步骤，包括：
    获取各信用风险评估子模型输出的信用风险评分，并在数据库中查询各信用风险评估子模型对应的预设权重值；
    根据所述预设权重值，通过下式对所述信用风险评分进行加权求和获得求和结果，并将所述求和结果作为企业信用风险评估结果；
    Figure PCTCN2018123278-appb-100002
    式中，S为求和结果，Y i为信用风险评估子模型输出的信用风险评分，X i为信用风险评估子模型对应的预设权重值。
  12. 如权利要求9所述的企业信用风险评估设备，其特征在于，所述获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别的步骤之前，所述方法还包括：
    获取若干个企业的企业信息，按预设行业类别对所述企业信息进行分类，获得各行业类别对应的行业信息样本；
    按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型。
  13. 如权利要求12所述的企业信用风险评估设备，其特征在于，所述按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型的步骤，包括：
    按预设数据维度从所述行业信息样本中提取出信用风险特征变量；
    对提取出的信用风险特征变量进行离散化分解获得变量因子，并将所述变量因子输入至预设神经网络模型进行模型训练，获得各行业类别对应的信用风险评估模型。
  14. 如权利要求13所述的企业信用风险评估设备，其特征在于，所述获得各行业类别对应的信用风险评估模型的步骤之后，所述方法还包括：
    从数据库中读取预先设定的模型分解规则，根据所述模型分解规则对各行业类别对应的信用风险评估模型进行模型拆分，获得各信用风险评估模型对应的信用风险评估子模型；
    分别对各信用风险评估模型对应的信用风险评估子模型进行权重配置，得到各信用风险评估子模型对应的预设权重值。
  15. 如权利要求9所述的企业信用风险评估设备，其特征在于，所述根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告的步骤之前，所述方法还包括：
    根据所述企业数据查找与所述目标企业存在关联关系的关联企业，并在所述数据库中查询是否存在所述关联企业对应的关联企业信用风险评估结果；
    若存在，则获取所述关联企业信用风险评估结果；
    所述根据所述企业信用风险评估结果生成所述目标企业对应的 企业信用风险评估报告的步骤，包括：
    根据所述关联企业信用风险评估结果以及所述目标企业对应的企业信用风险评估结果生成所述目标企业的企业信用风险评估报告。
  16. 一种存储介质，其特征在于，所述存储介质上存储有企业信用风险评估程序，所述企业信用风险评估程序被处理器执行时实现以下步骤：
    获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别；
    在预先构建的映射关系中查找所述行业类别对应的信用风险评估模型；
    按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果；
    根据所述企业信用风险评估结果生成所述目标企业对应的企业信用风险评估报告。
  17. 如权利要求16所述的可读存储介质，其特征在于，所述信用风险评估模型包括多个信用风险评估子模型；
    所述按预设数据维度从所述企业数据中提取信用风险特征变量，将提取到的信用风险特征变量输入至所述信用风险评估模型，获得企业信用风险评估结果的步骤包括：
    按预设数据维度从所述企业数据中提取信用风险特征变量；
    根据所述信用风险评估子模型对应的模型类别对所述信用风险特征变量进行分类，获得分类后的信用风险特征变量；
    将所述分类后的信用风险特征变量分别输入至对应的信用风险评估子模型；
    获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果。
  18. 如权利要求17所述的可读存储介质，其特征在于，所述获取各信用风险评估子模型输出的信用风险评分，根据所述信用风险评分获得企业信用风险评估结果的步骤，包括：
    获取各信用风险评估子模型输出的信用风险评分，并在数据库中查询各信用风险评估子模型对应的预设权重值；
    根据所述预设权重值，通过下式对所述信用风险评分进行加权求和获得求和结果，并将所述求和结果作为企业信用风险评估结果；
    Figure PCTCN2018123278-appb-100003
    式中，S为求和结果，Y i为信用风险评估子模型输出的信用风险评分，X i为信用风险评估子模型对应的预设权重值。
  19. 如权利要求16所述的可读存储介质，其特征在于，所述获取目标企业在预设时段内的企业数据，根据所述企业数据确定出所述目标企业所属的行业类别的步骤之前，所述方法还包括：
    获取若干个企业的企业信息，按预设行业类别对所述企业信息进行分类，获得各行业类别对应的行业信息样本；
    按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型。
  20. 如权利要求19所述的可读存储介质，其特征在于，所述按预设数据维度从所述行业信息样本中提取出信用风险特征变量，并根据提取出的信用风险特征变量建立各行业类别对应的信用风险评估模型的步骤，包括：
    按预设数据维度从所述行业信息样本中提取出信用风险特征变量；
    对提取出的信用风险特征变量进行离散化分解获得变量因子，并将所述变量因子输入至预设神经网络模型进行模型训练，获得各行业类别对应的信用风险评估模型。
PCT/CN2018/123278 2018-09-27 2018-12-25 企业信用风险评估方法、装置、设备及存储介质 WO2020062660A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811135377.2A CN109409677A (zh) 2018-09-27 2018-09-27 企业信用风险评估方法、装置、设备及存储介质
CN201811135377.2 2018-09-27

Publications (1)

Publication Number Publication Date
WO2020062660A1 true WO2020062660A1 (zh) 2020-04-02

Family

ID=65465400

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/123278 WO2020062660A1 (zh) 2018-09-27 2018-12-25 企业信用风险评估方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN109409677A (zh)
WO (1) WO2020062660A1 (zh)

Families Citing this family (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110417721B (zh) * 2019-03-07 2021-10-26 腾讯科技（深圳）有限公司 安全风险评估方法、装置、设备及计算机可读存储介质
CN109961368A (zh) * 2019-03-18 2019-07-02 京东数字科技控股有限公司 基于机器学习的数据处理方法及装置
CN110443458A (zh) * 2019-07-05 2019-11-12 深圳壹账通智能科技有限公司 风险评估方法、装置、计算机设备和存储介质
CN110443489A (zh) * 2019-07-31 2019-11-12 北京明略软件系统有限公司 目标行业的获取方法及装置、存储介质、电子装置
CN110400215B (zh) * 2019-07-31 2023-11-03 浪潮软件集团有限公司 面向企业家族的小微企业信用评估模型构建方法及系统
CN110472680B (zh) * 2019-08-08 2021-05-25 京东城市(北京)数字科技有限公司 目标分类方法、装置和计算机可读存储介质
CN110717821A (zh) * 2019-09-09 2020-01-21 上海凯京信达科技集团有限公司 一种车辆贷款评估方法、装置及计算机存储介质、电子设备
CN110889589A (zh) * 2019-10-23 2020-03-17 今稠科技(上海)有限公司 一种企业在线风控服务系统
CN112734598A (zh) * 2019-10-29 2021-04-30 北京国双科技有限公司 建立外包平台模型的方法、外包方查找方法及装置
CN111191091A (zh) * 2019-12-30 2020-05-22 成都数联铭品科技有限公司 一种数据分类方法及系统
CN111260484A (zh) * 2020-01-14 2020-06-09 支付宝(杭州)信息技术有限公司 一种人伤识别的数据处理方法、装置、服务器及系统
CN110930250A (zh) * 2020-02-12 2020-03-27 成都数联铭品科技有限公司 企业信用风险的预测方法及系统、存储介质及电子设备
CN111340365A (zh) * 2020-02-26 2020-06-26 深圳壹账通智能科技有限公司 企业数据处理方法、装置、计算机设备和存储介质
CN111369260A (zh) * 2020-03-10 2020-07-03 支付宝(杭州)信息技术有限公司 保护隐私的风险预测方法及装置
CN111507822A (zh) * 2020-04-13 2020-08-07 深圳微众信用科技股份有限公司 一种基于特征工程的企业风险评估方法
CN111915155A (zh) * 2020-07-13 2020-11-10 上海冰鉴信息科技有限公司 小微企业风险等级识别方法、装置及计算机设备
US11960993B2 (en) * 2020-11-10 2024-04-16 Equifax Inc. Machine-learning techniques involving monotonic recurrent neural networks
CN112561682A (zh) * 2020-12-10 2021-03-26 中信银行股份有限公司 一种针对小微企业的银行授信风险评估方法及系统
CN112613755A (zh) * 2020-12-25 2021-04-06 北京知因智慧科技有限公司 利用置信度评估企业风险的方法、装置及电子设备
CN112633709A (zh) * 2020-12-26 2021-04-09 中国农业银行股份有限公司 一种企业征信评估方法和装置
CN112734559B (zh) * 2020-12-30 2024-04-30 北京知因智慧科技有限公司 企业信用风险评价方法、装置及电子设备
CN112734099A (zh) * 2020-12-31 2021-04-30 北京知因智慧科技有限公司 企业风险的预测方法、装置及服务器
CN112907351A (zh) * 2021-02-05 2021-06-04 中国工商银行股份有限公司 金融报文异常识别方法及装置
CN112990707A (zh) * 2021-03-12 2021-06-18 深圳工盟科技有限公司 施工风险评估方法、装置、设备及存储介质
CN113177701A (zh) * 2021-04-15 2021-07-27 国任财产保险股份有限公司 一种用户信用评估方法和装置
CN114266022A (zh) * 2021-12-14 2022-04-01 上海淇玥信息技术有限公司 一种基于数据挖掘的反欺诈方法、装置及电子设备
CN114004212B (zh) * 2021-12-30 2022-03-25 深圳希施玛数据科技有限公司 数据处理方法、装置及存储介质
CN114757783A (zh) * 2022-04-26 2022-07-15 深圳平安综合金融服务有限公司 信息识别方法、装置、设备及存储介质
CN114781937A (zh) * 2022-06-20 2022-07-22 华网领业(杭州)软件有限公司 一种预付卡企业风险预警方法和装置以及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050119961A1 (en) * 2003-12-02 2005-06-02 Dun & Bradstreet, Inc. Enterprise risk assessment manager system
CN104966227A (zh) * 2015-06-11 2015-10-07 安徽融信金模信息技术有限公司 一种基于多种经营数据的企业风险评估系统
CN104992234A (zh) * 2015-06-11 2015-10-21 安徽融信金模信息技术有限公司 一种基于多种运营数据的企业风险评估方法
CN106779755A (zh) * 2016-12-31 2017-05-31 湖南文沥征信数据服务有限公司 一种网络电商借贷风险评估方法及模型

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050119961A1 (en) * 2003-12-02 2005-06-02 Dun & Bradstreet, Inc. Enterprise risk assessment manager system
CN104966227A (zh) * 2015-06-11 2015-10-07 安徽融信金模信息技术有限公司 一种基于多种经营数据的企业风险评估系统
CN104992234A (zh) * 2015-06-11 2015-10-21 安徽融信金模信息技术有限公司 一种基于多种运营数据的企业风险评估方法
CN106779755A (zh) * 2016-12-31 2017-05-31 湖南文沥征信数据服务有限公司 一种网络电商借贷风险评估方法及模型

Also Published As

Publication number Publication date
CN109409677A (zh) 2019-03-01

Similar Documents

Publication Publication Date Title
WO2020062660A1 (zh) 企业信用风险评估方法、装置、设备及存储介质
CN107945024B (zh) 识别互联网金融借贷企业经营异常的方法、终端设备及存储介质
US9990356B2 (en) Device and method for analyzing reputation for objects by data mining
CN109635117A (zh) 一种基于知识图谱识别用户意图方法及装置
WO2021027595A1 (zh) 用户画像生成方法、装置、计算机设备和计算机可读存储介质
Sehgal et al. Sops: stock prediction using web sentiment
CN110688454A (zh) 咨询对话处理的方法、装置、设备及存储介质
JP6713238B2 (ja) 電子装置、小売店舗評価モデルを構築する方法、システム及び記憶媒体
WO2018184518A1 (zh) 微博数据处理方法、装置、计算机设备及存储介质
JP2021111367A (ja) Ａｐｉ評価のための自動パラメータ値解決
CN113342976B (zh) 一种自动采集处理数据的方法、装置、存储介质及设备
WO2021098651A1 (zh) 一种风险实体获取方法及装置
WO2023029065A1 (zh) 数据集质量评估方法、装置、计算机设备及存储介质
CN113392920B (zh) 生成作弊预测模型的方法、装置、设备、介质及程序产品
WO2021129368A1 (zh) 一种客户类型的确定方法及装置
CN117633500A (zh) 风险反诈识别方法、装置、设备及存储介质
WO2020119533A1 (zh) 基于循环神经网络算法的舆情预警方法、装置、终端及介质
CN117196630A (zh) 交易风险预测方法、装置、终端设备以及存储介质
CN111222032A (zh) 舆情分析方法及相关设备
CN114417974B (zh) 模型训练方法、信息处理方法、装置、电子设备和介质
CN114398562B (zh) 一种店铺数据管理方法、装置、设备及存储介质
CN115936748A (zh) 一种商业大数据分析方法及系统
KR20230103025A (ko) 기업 신용 분석 및 평가 정보 제공 방법, 장치 및 시스템
CN111209397B (zh) 一种确定企业产业类别的方法
TWM625689U (zh) 雲端智慧動態產業分析系統

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18935618

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 08/07/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 18935618

Country of ref document: EP

Kind code of ref document: A1