TW202147222A - Enterprise loan evaluation system for providing remote credit checking and fast loaning approval - Google Patents

Abstract

Disclosed is an enterprise loan evaluation system, which comprises: a bank computer system including a credit limit calculation and credit analysis module, i.e. a database. When the bank computer system receives an application documentation image related to an enterprise from a first mobile device, the application documentation image is first performed with image recognition, so as to generate an application documentation data containing the content of a plurality of paper application documentation. Based on the application documentation data, the credit limit calculation and credit analysis module obtains a joint credit data related to the enterprise, i.e. a client's data, from the database and a computer system of the Joint Credit Information Center, then generates an input data accordingly for being inputted to a machine learning module to calculate and generate a loanable credit. The bank computer system transmits the loanable credit to the first mobile device for notifying the enterprise in real time.

Description

Enterprise Loan Evaluation System

The present invention relates to a financial system, in particular to an enterprise loan application evaluation system for remote credit investigation and quick loan verification.

With the prosperity of entrepreneurship, many small and medium-sized or micro-enterprises gradually appear in the society, but the capital turnover problem that accompanies the operation of enterprises often makes many business owners exhausted. Funds are the key to maintaining the company's operations. Compared with personal credit loans, corporate loans require more documents due to existing regulatory requirements, making general banks accustomed to passively waiting for customers to consult back and forth and complete paper documents. , to start reviewing loan applications. As a result, not only does it take a long time for enterprises to apply for loans, but the final approved quota conditions often fall short of customer expectations, which causes inconvenience to the scheduling and development of operating funds for enterprises, which in turn affects the willingness of enterprises to apply for loans from banks. , therefore, becomes an open problem.

Therefore, the purpose of the present invention is to provide an enterprise loan application evaluation system for remote credit investigation and rapid loan verification.

Therefore, the present invention provides an enterprise loan application evaluation system, which is suitable for a first mobile device and a computer system of a joint credit reporting center, and includes a bank computer system. The bank computer system belongs to a bank, can establish a connection with the first mobile device, and includes a quota calculation and review module and a database, the database stores a machine learning model and a related data of an enterprise customer data.

Wherein, when the bank computer system receives an application document image from the first mobile device, it performs image recognition on the application document image to generate an image of the contents of a plurality of paper application documents included in the application document image. Application documents, the paper application documents are related to a loan application made by the enterprise.

The quota calculation and review module obtains a joint information about the enterprise from the computer system of the joint credit information center according to the application documents, and obtains the customer information of the enterprise from the database, and then according to the association collect data and the customer data, obtain an input data to input the machine learning model, and then calculate to generate a loanable amount, the bank computer system then transmits the loanable amount to the first mobile device for real-time notification the enterprise.

In some implementation aspects, the machine learning model is a gradient boosting decision tree (GBDT) model.

In some other implementation aspects, the quota calculation and review module obtains the variable data of an industry of the industry to which the enterprise belongs, and the first rating of the rating of the bank according to the joint collection data or the customer data variable data, a plurality of revenue variable data related to the revenue of the business, a plurality of asset variable data related to the assets of the business, a plurality of liability variable data related to the liabilities of the business, and a plurality of company seniority variables data as the input data.

In some implementations, the rating variable data includes one of a plurality of different values corresponding to a plurality of different ratings. The industry-specific variable data includes one of a plurality of different values corresponding to a plurality of different industries.

Such revenue variable data include the first variable of the company's revenue in the last 12 months, the first variable of the company's remittance passbook details in the last three months that can be identified as the monthly average of revenue, the company's recent The first and third variable of the monthly average amount of actual revenue in half a year, and the details of the passbook remitted by the company in the past three months can be identified as the first and third variable of the monthly average amount of revenue divided by the monthly average amount of actual revenue of the company in the past half year The 4th variable, the record of the company and the person in charge and the spouse of the company in the last three months divided by the company's revenue in the last 12 months. The fifth variable, and the company's revenue in the last 12 months minus last year's revenue at least one of a sixth variable of .

The asset variable information includes a seventh variable of whether the enterprise and the person in charge of the enterprise and his spouse and children have real estate, and an eighth variable of the record of the enterprise and the person in charge of the enterprise and his spouse in the past three months. at least one of them.

The debt variable data includes the ninth variable of the latest total credit balance of the enterprise, the tenth variable of the total credit balance of the person in charge of the enterprise, and the monthly principal and interest expenses of the enterprise and the person in charge and the spouse of the enterprise. The 11th variable, the latest total credit balance of the company divided by the revenue in the last 12 months. The 12th variable, the sum of the company's latest total credit balance and the total credit balance of the person in charge of the company divided by the company's last 12 months The 13th variable of revenue, the record of the company and the person in charge of the company in the past three months divided by the monthly principal and interest expenses of the company and the person in charge of the company and the spouse, and the company and the company. The 14th variable The latest total credit balance of the person in charge of the enterprise and the spouse minus at least one of the 15th variables of the total credit balance of the enterprise and the person in charge of the enterprise and the spouse of the same period last year.

The variable data of company qualifications include a 16th variable of whether the company has contacts with a leasing company, a 17th variable of how many banks the company has been jointly queried in the past three months, and how many banks the company has contacted with the bank. At least one of the 18th variable of the enterprise and the number of correspondent banks that do not include the bank, the 20th variable of how many consecutive months the enterprise has been operating, and an economic indicator. The prosperity index includes one of a plurality of different values corresponding to a plurality of different prosperity levels.

In some other implementations, the machine learning model is firstly trained with a training input data and a training target data, the training input data includes belonging to the bank and within a training time interval and between relevant The industry-specific variable data, the rating variable data, the revenue variable data, and the asset variables for multiple corporate clients between a first ratio and a second ratio of a revenue-to-credit ratio data, the variable data of liabilities, and the variable data of company seniority, and the training target data is a plurality of actual loan amounts corresponding to the corporate customers who should train the input data respectively.

In some other implementations, one of the variable data of revenue is the revenue of the company in a revenue range within a predetermined time interval, and the quota calculation and review module is based on the variable data of the industry, the At least one of the rating variable data and the income of the revenue range, to determine at least one correspondence among an industry note, a rating note, and a revenue range note, so as to determine the industry note, the rating note , and at least the counterpart of the revenue range note as part of the input data.

In some implementations, each of the industry note, the rating note, and the revenue range note includes two values. When the quota calculation and review module judges that the industry to which the enterprise belongs is equal to and not equal to a set industry according to the variable data of the industry, it determines that the industry mark is respectively equal to a first value and a second value, and according to the rating The variable data determines that when the rating of the enterprise is higher than or equal to and lower than a predetermined rating, respectively, the rating note is determined to be equal to a third value and a fourth value, and according to the enterprise's performance in the predetermined time interval When the revenue of the revenue range is greater than or equal to and less than a set amount, respectively, it is determined that the revenue of the revenue range is equal to a fifth value and a sixth value respectively.

In some other implementations, the machine learning model is firstly trained with a training input data and a training target data, the training input data includes belonging to the bank and within a training time interval and between relevant The industry-specific variable data, the rating variable data, the revenue variable data, and the asset variables for multiple corporate clients between a first ratio and a second ratio of a revenue-to-credit ratio data, the variable data of liabilities, the variable data of company seniority, the industry notes, the rating notes, and the revenue range notes, the training target data are for the corporate clients who should input the data for training respectively multiple actual loan amounts.

In some other implementation aspects, the ratio of revenue to credit extension = (financial liabilities of the enterprise - long-term guarantees and long-term loans of the enterprise + financial liabilities of the person in charge of the enterprise - long-term liabilities of the person in charge of the enterprise Guaranteed and long-term loan loan + revolving card fee of the credit card of the person in charge of the enterprise + loan balance of the cash card of the person in charge of the enterprise + financial liabilities of the spouse of the person in charge of the enterprise - long-term of the spouse of the person in charge of the enterprise Guaranteed and long-term loan loans + revolving card fee of the credit card of the spouse of the person in charge of the enterprise + loan balance of the cash card of the spouse of the person in charge of the enterprise) / revenue of the enterprise's 401 form in the past 12 months.

In some other implementation aspects, the enterprise loan application evaluation system is further applicable to a credit investigator and a second mobile device, after the first mobile device notifies the enterprise of the loanable amount, and sends a confirmation application After the loan instruction is transmitted to the bank computer system, the bank computer system generates a credit question group corresponding to the industry to which the enterprise belongs, and transmits the credit question group to the second mobile device, the second mobile device Obtain a credit report corresponding to the credit investigation group by the credit investigator, complete the writing of the credit report and transmit it to the bank's computer system, and the bank's computer system decides whether to approve the corresponding credit report according to the credit report. The loan application for the loanable amount.

The effect of the present invention is that: when the bank computer system receives the image of the application document of the first mobile device, it first obtains the data of the application document according to the image of the application document, and then the computer system of the joint credit reporting center and the The database obtains the joint information and the customer information. The limit calculation and review module obtains the input data according to the joint data and the customer data, and inputs the machine learning model to calculate the loanable limit, which can be sent to the first mobile device in real time to notify the Therefore, it can effectively solve the problems encountered in the existing loan application.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are designated by the same reference numerals.

Referring to FIG. 1, an embodiment of the enterprise loan application evaluation system of the present invention is suitable for a business person, a credit investigation officer, a first mobile device 2, a second mobile device 3, and a joint credit investigation center computer system 4 , and contains a bank computer system 1 . Any one of the first mobile device 2 and the second mobile device 3 is, for example, a smart phone, a tablet computer, or other similar portable electronic devices with networking functions. The bank computer system 1, the business personnel, and the credit investigation personnel all belong to the same bank. The computer system 4 of the joint credit reporting center is, for example, one or more computer hosts or servers, to provide credit reporting services.

The first mobile device 2 is pre-installed with an application program (APP) provided by the bank and includes an image capturing module 21 . The image capturing module 21 is, for example, a camera lens and related sensors. The business person visits a customer, such as a person in charge of an enterprise, and after confirming that the enterprise is willing to apply for a loan to the bank, the first mobile device 2 is selected by the business person to execute the application to use The image capturing module 21 captures an application document image including a plurality of paper application documents, and establishes a connection with the bank computer system 1 to transmit the application document image to the bank computer system 1 . The paper application documents relate to the loan application made by the enterprise (ie the responsible person). For example, these paper documents include the relevant party's consent form, company registration information, various financial statements, guarantor documents and financial information, and other documents required for loan processing.

The bank computer system 1 is, for example, one or more computer hosts or servers of the bank, and includes a quota calculation and review module 11 and a database 12 . The quota calculation and review module 11 is, for example, a separate computer host or server, but not limited thereto. The database 12 stores a machine learning model and a customer profile related to the enterprise. In this embodiment, the machine learning model belongs to a supervised learning model, and is, for example, a gradient boosting decision tree (Gradient boosting decision tree, GBDT) model.

When the bank computer system 1 receives the image of the application document from the first mobile device 2, it first performs image recognition on the image of the application document to generate an application document including the contents of the paper application documents. The quota calculation and review module 11 then obtains a joint registration data related to the enterprise from the computer system 4 of the joint credit information center according to the application document data, and obtains the customer data related to the company from the database 12, Then, according to the joint data and the customer data, an input data is obtained. The amount calculation and review module 11 then inputs the input data into the machine learning model, and then calculates to generate a loanable amount. The bank computer system 1 then transmits the loanable amount to the first mobile device 2 for real-time loanable amount. notify the company.

In more detail, the quota calculation and review module 11 obtains the variable data of one industry category of the industry to which the enterprise belongs, the variable data of the first rating of the bank's rating, and related variables according to the joint survey data or the customer data. A plurality of revenue variable data related to the business' revenue, a plurality of asset variable data related to the business' assets, a plurality of liability variable data related to the business' liabilities, and a plurality of company seniority variable data.

In this embodiment, the rating variable data includes one of multiple different numerical values corresponding to multiple different ratings, for example, the five ratings of A, B, C, D, and E correspond to numerical values of 1 to 5, respectively. . The industry-specific variable data includes one of multiple different values corresponding to multiple different industries. For example, the values corresponding to manufacturing, wholesale, service, retail and catering, construction, and other industries are 1. ~6.

The revenue variable data includes a first variable to a sixth variable. The first variable is the revenue of the company in the last 12 months, and the unit is, for example, thousand yuan. The second variable is the monthly average amount of revenue that can be identified as the company's remitted passbook details in the past three months, the unit is, for example, thousand yuan. The third variable is the monthly average of the actual revenue of the company in the past six months, and the unit is, for example, thousand yuan. The fourth variable is the monthly average of the company's remitted passbook details that can be identified as revenue in the past three months divided by the monthly average of the company's actual revenue in the past six months. The fifth variable is the last three months' performance of the company and the person in charge and the spouse of the company divided by the company's revenue in the last 12 months. Deposits are the number of passbooks, and the average is calculated based on the balance of deposits on the 5th, 10th, 15th, 20th, 25th, and 30th of each month. The sixth variable is the revenue of the company in the last 12 months minus the revenue of the previous year, and the unit is, for example, thousand yuan.

The asset variable data includes a seventh variable to an eighth variable. The seventh variable is, for example, equal to 0 or 1, indicating that the enterprise and the person in charge of the enterprise and the spouse and children do not hold or hold real estate, respectively. The eighth variable is the record of the company and the person in charge and the spouse of the company in the past three months, and the unit is, for example, thousand yuan.

The liability variable information includes a ninth variable to a fifteenth variable. The ninth variable is the latest total credit balance of the enterprise, which is the total credit balance updated by the latest joint collection center. For example, the joint collection center updates information on the 15th of each month, and the unit is, for example, thousand yuan . The tenth variable is the total credit balance of the husband and wife of the person in charge of the enterprise, and the unit is, for example, thousand yuan. The 11th variable is the monthly principal and interest expenses of the enterprise and the person in charge of the enterprise and the spouse, and the unit is, for example, thousand yuan. The 12th variable is the company's most recent total credit balance divided by the most recent 12-month revenue. The 13th variable is the sum of the recent total credit balance of the company and the total credit balance of the person in charge of the company divided by the company's revenue in the last 12 months. The 14th variable is the business record of the enterprise and the person in charge and the spouse of the enterprise in the past three months divided by the monthly principal and interest expenses of the enterprise and the person in charge of the enterprise and the spouse. The 15th variable is the latest total credit balance between the enterprise and the person in charge of the enterprise and the spouse minus the total credit balance between the enterprise and the person in charge and the spouse of the enterprise in the same period last year, in units of thousand yuan, for example.

The variable information of company qualifications includes a 16th variable to a 20th variable and a prosperity index. The 16th variable is, for example, equal to 0 or 1, respectively indicating that the company does not or has contact with the leasing company. The 17th variable is how many banks the company has been jointly inquired about in the past three months. The 18th variable is how many banks the company has dealings with that includes the bank. The 19th variable is how many banks the enterprise has dealings with that do not include the bank. The 20th variable is how many consecutive months the business has been in business. For example, if the prosperity index is equal to 1 to 5, it indicates that the prosperity is extremely poor, slightly poor, moderate, slightly better, and excellent, respectively.

In addition, one of the revenue variable data is the company's revenue in a revenue range within a predetermined time range, such as annual revenue. The quota calculation and review module 11 determines an industry note, a rating note, and a revenue range note respectively according to the industry-specific variable data, the rating variable data, and the revenue in the revenue range. In this embodiment, each of the industry note, the rating note, and the revenue range note includes two types of values, and the quota calculation and review module determines the industry category to which the company belongs according to the industry-specific variable data When they are respectively equal to and not equal to a set industry, it is determined that the industry note is equal to a first value and a second value respectively, and according to the rating variable data, it is determined that the rating of the enterprise in the bank is higher than or equal to and lower than a respectively. When setting the rating, it is determined that the rating note is respectively equal to a third value and a fourth value, and the business is determined according to when the revenue of the enterprise in the revenue range in the predetermined time interval is greater than or equal to and less than a set amount, respectively. The closing interval notes are respectively equal to a fifth value and a sixth value.

For example, if the set industry is a construction industry and the industry is the construction industry, the industry mark is equal to 0 (ie the first value), otherwise, it is equal to 1 (ie the second value). The set rating is level B. When the rating is level A or B, the rating note is equal to 1 (ie the third value), otherwise, it is equal to 0 (ie the fourth value). The set amount is 40 million yuan, and when the annual revenue is greater than or equal to 40 million yuan, the revenue range note is equal to 1, (that is, the fifth value), otherwise, it is equal to 0 (that is, the sixth value).

The quota calculation and review module 11 uses the industry variable data, the rating variable data, the revenue variable data, the asset variable data, the liability variable data, the company qualification variable data, and the industry note , the rating note, and the revenue range note are used as the input data to input the machine learning model, and then calculate to generate the loanable amount. In addition, it should be specially added that: in other embodiments, the input data can also omit the industry note, the rating note, and the revenue range note, but the effect of the calculated creditable amount will be slightly Almost.

The machine learning model is trained with a training input data and a training target data in advance, and the training input data includes a first time belonging to the bank and within a training time interval and related to a revenue and credit ratio. The industry-specific variable data, the rating variable data, the revenue variable data, the asset variable data, the liability variable data, the Company seniority variable information, such industry notes, such rating notes, and such revenue range notes. The training target data is a plurality of actual loan amounts respectively corresponding to the enterprise customers whose input data should be trained.

For example, the revenue to credit ratio = (the company's financial liabilities - the company's long-term guarantees and long-term loans + the financial liabilities of the person in charge of the company - the long-term guarantee and long-term loans of the company's responsible person + The revolving card fee of the credit card of the person in charge of the enterprise + the loan balance of the cash card of the person in charge of the enterprise + the financial liabilities of the spouse of the person in charge of the enterprise - the long-term guarantee and long-term loan loan of the spouse of the person in charge of the enterprise + The revolving card fee of the credit card of the spouse of the person in charge of the enterprise + the loan balance of the cash card of the spouse of the person in charge of the enterprise)/the revenue of the 401 form of the enterprise in the past 12 months, the first ratio is 6.3%, The second ratio is 92.5%, the training time interval is the first 11 months among the 12 months before the current month, and the corresponding data of the previous month before the current month is used as the test data. In the testing process after training, when each loanable amount calculated by the machine learning model is greater than or equal to 80% of the corresponding actual loan amount, it is regarded as a hit, that is, the estimated result is regarded as a hit. to be valid. In addition, it should be noted that: in this embodiment, the calculation result of the loanable amount is, for example, the amount corresponding to one type of repayment loan, but it can also be converted to the amount of another type of revolving loan through a predetermined formula.

After the business person transmits the image of the application document to the bank computer system 1 through the first mobile device 2, the bank computer system 1 can immediately transmit the loanable amount back to the first mobile device 2, so that The business person can advise the customer (ie the person in charge) on the product mix and quota on the spot. When the enterprise decides to apply for the loan application of the loanable amount, the first mobile device 2 sends a confirmation loan application instruction to the bank computer system 1 through the business personnel, so that the bank computer system 1 generates a corresponding loan application. A credit question group in another industry to which the enterprise belongs, and transmits the credit question group to the second mobile device 3 .

Afterwards, the credit investigator brings the second mobile device 3 to the enterprise to have an interview with the person in charge, asks the person in charge the question of the credit investigation group, and inputs and generates a corresponding response through the second mobile device 3 A credit report of the credit investigation team is to write and complete the credit report. The second mobile device 3 then transmits the credit report to the bank computer system 1, so that the bank computer system 1 decides whether to approve the loan application corresponding to the loanable amount according to the credit report. For example, the credit report is a 5P questionnaire, and includes: 1. People, the history of the company, experience of the person in charge, etc.; 2. Purpose, the purpose of the loan and the rationality of the current borrowing; 3. Payment, the source of repayment ; 4.Protection, relevant qualifications of the insurer; and 5.Perspective, the company's future development and prospects. In addition, the correctness and reasonableness of the relevant information are reconfirmed by the authorized personnel, or the content is automatically identified and interpreted by the bank's computer system 1 to decide whether to approve the loan application. Therefore, the credit investigator can simultaneously complete the factory visit report according to the credit investigation group of the second mobile device 3 at the end of the visit, thereby greatly reducing the time required for the verification of the loan.

To sum up, after receiving the image of the application document from the first mobile device 2, the bank computer system 1 first obtains the data of the application document according to the image of the application document, and then the computer system 4 of the Joint Credit Information Center respectively obtains the data of the application document according to the image of the application document. And the database 12 obtains the joint data and the customer data. The limit calculation and review module 11 obtains the input data according to the joint data and the customer data, and inputs the machine learning model to calculate the loanable limit, which can be sent to the first mobile device 2 in real time for Notify the business. Furthermore, the credit investigation personnel can use the second mobile device 3 to complete the factory visit report while performing the credit investigation interview, so the problems encountered in the existing loan application can be effectively solved, and the object of the present invention.

However, the above are only examples of the present invention, and should not limit the scope of implementation of the present invention. Any simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the contents of the patent specification are still included in the scope of the present invention. within the scope of the invention patent.

1: Bank computer system 11: Quota calculation and review module 12:Database 2: The first mobile device 21: Image capture module 3: Second mobile device 4: Computer System of Joint Credit Information Center

Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, wherein: FIG. 1 is a block diagram illustrating an embodiment of the enterprise loan application evaluation system of the present invention.

1: Bank computer system

11: Quota calculation and review module

12:Database

2: The first mobile device

21: Image capture module

3: Second mobile device

4: Computer System of Joint Credit Information Center

Claims (10)

An enterprise loan application evaluation system is suitable for a first mobile device and a computer system of a joint credit information center, and includes a bank computer system, the bank computer system belongs to a bank and can establish a connection with the first mobile device , and includes a quota calculation and review module and a database, the database stores a machine learning model and a customer data related to an enterprise, wherein, When the bank computer system receives an application document image from the first mobile device, it performs image recognition on the application document image to generate an application document containing the contents of a plurality of paper application documents included in the application document image information, such paper application documents relate to a loan application made by the enterprise, The quota calculation and review module obtains a joint information about the enterprise from the computer system of the joint credit information center according to the application documents, and obtains the customer information of the enterprise from the database, and then according to the association collect data and the customer data, obtain an input data to input the machine learning model, and then calculate to generate a loanable amount, the bank computer system then transmits the loanable amount to the first mobile device for real-time notification the enterprise. The enterprise loan application evaluation system according to claim 1, wherein the machine learning model is a gradient boosting decision tree (Gradient boosting decision tree, GBDT) model. The enterprise loan application evaluation system according to claim 1, wherein the quota calculation and review module obtains the variable data of an industry of the industry to which the company belongs, and the evaluation module of the bank to which the bank belongs according to the joint collection data or the customer data. etc., multiple revenue variable information related to the company's revenue, multiple asset variable information related to the company's assets, multiple liability variable data related to the company's liabilities, and Multiple company seniority variable data as the input data. The enterprise loan application evaluation system according to claim 3, wherein: The rating variable data includes one of a plurality of different values corresponding to a plurality of different ratings, and the industry-specific variable data includes one of a plurality of different values corresponding to a plurality of different industries, Such revenue variable data include the first variable of the company's revenue in the last 12 months, the first variable of the company's remittance passbook details in the last three months that can be identified as the monthly average of revenue, the company's recent The first and third variable of the monthly average amount of actual revenue in half a year, and the details of the passbook remitted by the company in the past three months can be identified as the first and third variable of the monthly average amount of revenue divided by the monthly average amount of actual revenue of the company in the past half year The 4th variable, the last three months’ performance of the company and the person in charge and the spouse of the company divided by the company’s revenue in the last 12 months. The fifth variable, and the company’s revenue in the last 12 months minus last year’s revenue At least one of the sixth variables of , The asset variable information includes a seventh variable of whether the enterprise and the person in charge of the enterprise and his spouse and children have real estate, and an eighth variable of the record of the enterprise and the person in charge of the enterprise and his spouse in the past three months. at least one of them, The debt variable data includes the ninth variable of the latest total credit balance of the enterprise, the tenth variable of the total credit balance of the person in charge of the enterprise, and the monthly principal and interest expenses of the enterprise and the person in charge and the spouse of the enterprise. The 11th variable, the latest total credit balance of the company divided by the revenue in the last 12 months. The 12th variable, the sum of the company's latest total credit balance and the total credit balance of the person in charge of the company divided by the company's latest 12 months The 13th variable of revenue, the record of the company and the person in charge of the company in the past three months divided by the monthly principal and interest expenses of the company and the person in charge of the company and the spouse, and the company and the company. The 14th variable The latest total credit balance of the person in charge of the enterprise and the spouse minus at least one of the 15th variables of the total credit balance of the enterprise and the person in charge of the enterprise and the spouse of the same period last year, The variable data of company qualifications include a 16th variable of whether the company has contacts with a leasing company, a 17th variable of how many banks the company has been jointly inquiring about in the past three months, and how many banks the company has contacted with the bank. at least one of an 18th variable of the enterprise and the number of correspondent banks that do not include the bank, a 20th variable of how many consecutive months the enterprise has been operating, and an economic indicator, the economic indicator Contains one of a number of different values corresponding to a number of different prosperity levels. The enterprise loan application evaluation system according to claim 3, wherein the machine learning model is firstly trained with a training input data and a training target data, and the training input data includes belonging to the bank and at a training time The industry-specific variable data, the rating variable data, and the revenue variable data for a plurality of corporate customers in the interval and between a first ratio and a second ratio related to a revenue-to-credit ratio , the asset variable data, the liability variable data, and the company seniority variable data, and the training target data is a plurality of actual loan amounts respectively corresponding to the corporate customers whose data should be input for training. The enterprise loan application evaluation system according to claim 3, wherein one of the revenue variable data is the revenue of the company in a revenue range within a predetermined time interval, and the quota calculation and review module is based on the industry At least one of the variable data, the variable data of the rating, and the revenue of the revenue range, determine at least one of an industry note, a rating note, and a revenue range note, so as to register the industry , the rating note, and at least the counterpart of the revenue range note as part of the input data. The enterprise loan application evaluation system according to claim 6, wherein each of the industry note, the rating note, and the revenue range note includes two values, and the quota calculation and review module is based on the industry When it is judged from the variable data that the industry to which the company belongs is equal to or not equal to a set industry, it is determined that the industry note is equal to a first value and a second value, and the company is judged to be in the bank's rating based on the variable data of the rating. When the rating is higher than or equal to and lower than a set rating, it is determined that the rating note is respectively equal to a third value and a fourth value, and the revenue of the enterprise in the predetermined time interval is greater than or equal to and respectively. When the amount is less than a set amount, it is determined that the camp award interval notes are respectively equal to a fifth value and a sixth value. The enterprise loan application evaluation system according to claim 6, wherein the machine learning model is firstly trained with a training input data and a training target data, and the training input data includes belonging to the bank and at a training time The industry-specific variable data, the rating variable data, and the revenue variable data for a plurality of corporate customers in the interval and between a first ratio and a second ratio related to a revenue-to-credit ratio , the asset variable data, the liability variable data, the company seniority variable data, the industry notes, the rating notes, and the revenue range notes, the training target data is the input data corresponding to the training multiple actual loan limits for those corporate customers. The enterprise loan application evaluation system according to claim 5 or 8, wherein, the ratio of revenue to credit grant = (the financial liabilities of the enterprise - the long-term guarantees and long-term loans of the enterprise + the financial liabilities of the person in charge of the enterprise - The long-term guarantee and long-term loan loan of the person in charge of the enterprise + the revolving card fee of the credit card of the person in charge of the enterprise + the loan balance of the cash card of the person in charge of the enterprise + the financial liabilities of the spouse of the person in charge of the enterprise - the enterprise The long-term guarantee and long-term loan loan of the spouse of the person in charge of the enterprise + the revolving card fee of the credit card of the spouse of the person in charge of the enterprise + the loan balance of the cash card of the spouse of the person in charge of the enterprise) / 401 of the enterprise in the past 12 months table revenue. The enterprise loan application evaluation system as described in claim 5 or 8 is also applicable to a credit investigator and a second mobile device. After the first mobile device notifies the enterprise of the loanable amount, a confirmation of the loan application is sent to the enterprise. After the instruction is sent to the bank computer system, the bank computer system generates a credit report question group corresponding to the industry to which the enterprise belongs, and transmits the credit report question group to the second mobile device, and the second mobile device borrows The credit investigator obtains a credit report corresponding to the credit investigation group, and completes the writing of the credit report and transmits it to the bank's computer system. The loan application for the loan amount.

Images