TWM574723U - Capital loan intelligent audit system - Google Patents

Abstract

This creation proposes an intelligent audit system for fund lending, including an operation interface module, a static audit module, a credit inquiry module and a dynamic audit module. The operation interface module receives static data and dynamic data input by the borrower. The static audit module calculates the static data to generate the first audit result to determine the preliminary rating of the borrower. When the static review module judges that the borrower ’s initial rating is passed, the credit questioning module calculates static data to generate a credit test question group for the borrower to answer and generates dynamic data in response to the borrower ’s physical response. The dynamic audit module cross-references static and dynamic data to generate a second audit result for presentation of the borrower's final rating. Among them, the dynamic audit module uses machine learning experience to determine the truth of the borrower's answer to the credit test questions. Falseness, so as to remotely and intelligently check the credit rating of the borrower.

Description

Fund lending intelligent review system

This creation is about a field of electronic finance, especially applied to the field of lending and matching, and the use of artificial intelligence to implement an intelligent credit and loan review system for reviewing credit ratings.

The development of the Internet has not only opened up many emerging industries, but has also driven innovation in many industries. In particular, the financial sector is more affected. For example, electronic financial innovation services such as online shopping and electronic payment have rapidly risen and quickly changed the consumption habits of users, and even the lending services for funds have begun to change. Among them, the Peer-to-peer lending (P2P Lending) service has been developed, which refers to the cash lending behavior between individuals and through the Internet platform, and removes the traditional role of bank intermediaries. Instead, the two individuals directly lend and borrow money.

The current online lending platform, where the credit risk assessment of the borrower will be reviewed by the staff of the online lending platform, which is to review the documents and materials submitted by the borrower when applying, or to further negotiate with the borrower's telecommunications And cross comparison. After the assessment, the credit rating of the lender is generated for subsequent borrowing behavior. However, the aforesaid operation takes too long, and even if the staff directly negotiates with the borrower, it is still not easy and effective to objectively and fairly evaluate their credit risk. On the other hand, the current labor cost is too high, resulting in a low operating efficiency of the overall borrowing.

In view of this, the author feels that he is not perfect and has exhausted his mental and painstaking research, and based on his accumulated experience in the industry for many years, he provides an intelligent review system for fund lending in order to improve the lack of the above-mentioned conventional technologies.

One of the purposes of this creation is to provide an intelligent review system for fund lending, which is applied to online lending platforms and uses artificial intelligence to provide remote intelligent review of credit ratings of borrowers.

In order to achieve the above purpose, this creation of an intelligent review system for fund lending for remote and intelligent review of a borrower's credit rating includes an operation interface module, a static audit module, a credit questioning module and a dynamic Review module. The operation interface module is used to prompt an input interface to the borrower, and to receive a static data and a dynamic data input by the borrower, and the static data includes at least one digital file or at least one text message. , The dynamic data includes at least one sound information or at least one image information. The static audit module telecommunication is connected to the operation interface module, and receives the static data transmitted by the operation interface module, and the static audit module calculates the static data to generate a first audit result to determine the borrower. The initial rating is pass or fail. The credit questioning module telecommunications connects the operation interface module and the static auditing module. When the static auditing module judges that the borrower's preliminary rating is passed, the credit questioning module receives and calculates the static data. A credit test question set is generated, and the credit test question set is transmitted to the operation interface module for the borrower to answer, and then the dynamic data is generated according to the physical response of the borrower. The dynamic auditing module is electrically connected to the operation interface module, the static auditing module and the credit questioning module, the dynamic auditing module receives the dynamic data transmitted by the operation interface module, and the dynamic auditing module It compares the static data with the dynamic data, and then generates a second audit result for presenting the final rating of the borrower. Among them, the dynamic audit module judges the borrower to answer based on machine learning experience. The authenticity of this credit test question group. In this way, this creation has multiple modules with intelligent calculations, and can use two stages of static and dynamic auditing modules and implement artificial intelligence using machine learning to intelligently review the credit rating of the borrower from a long distance. , Can improve the operational efficiency of subsequent borrowing behavior.

In addition, the fund lending intelligent audit system further includes a tracking fact module, which is connected to the static audit module, the credit questioning module and the dynamic audit module by telecommunications, and is tracked in one of the tracking fact modules. After time, the tracking fact module accesses the borrower's repayment record and generates a first correction information for amending the borrower's credit rating, so that the static review module, the credit questioning module and the The dynamic review module receives and calculates the first correction information to update through machine learning. The real credit of the borrower is learned through the tracking facts module, which can be updated by machine learning for each module with intelligent calculation of this creation, so as to improve the accuracy and efficiency of the intelligent calculation of this creation.

Preferably, the credit test question group has a plurality of questions, and the borrower answers each of the questions in order. The latter question is generated by the credit questioning module after receiving and calculating the previous question and its corresponding answer, and it is related to the previous question. In this way, the borrower can be reviewed step by step, and the fairness of the credit rating review of this creation can be further improved.

Further, the second audit result includes a plurality of answer question audit results, and the dynamic audit module calculates the static data and the dynamic data according to the answers corresponding to the questions, and then sequentially generates the corresponding answers to the questions. Answer review results. Therefore, each of these questions has a corresponding review result of each answer to further improve the fairness of the credit rating review of this creation.

And in this embodiment, the fund lending intelligent auditing system further includes a back-end processing module, which is a telecommunication connection to the operation interface module, the static auditing module, the credit questioning module, and the dynamic auditing module. And the tracking facts module. With this, maintenance personnel correct the first audit result or the answer audit results through the back-end processing module, and generate a second correction information for amending the borrower's credit rating, so that the static audit module The credit questioning module and the dynamic review module respectively receive and calculate the second correction information for updating through machine learning, thereby assisting the machine learning to improve the accuracy and efficiency of the intelligent calculation of this creation.

In addition, the dynamic audit module has a warning unit. When the borrower's preliminary or final rating is not passed, the warning unit generates a warning message, and the back-end processing module receives the warning message, so that the maintenance staff can immediately know the credit rating Wait. And when the credit rating results are obviously distorted, the maintenance staff can use the back-end processing module to correct the credit rating of the borrower to supplement the machine learning of each module created by this, thereby improving the accuracy of the intelligent calculation of this creation. Sex and effectiveness.

Continuing, the dynamic audit module has a contract unit. When the final audit result is passed, the contract unit generates a contract and provides a prompt for the operation interface module to provide the borrower to sign a contract to facilitate the completion of the borrowing behavior.

Preferably, the dynamic audit module further includes an environmental background unit, which calculates the dynamic data to generate environmental information, and calculates the environmental information and the first correction information to generate a third correction information for correction. Credit rating of the borrower. In this way, the true credit of the borrower is known through the first correction information to consider the correlation between the environmental information and the dynamic information, so that the static review module, the credit question module and the The dynamic review module receives and calculates the third correction information to update it through machine learning, thereby improving the fairness of the credit rating review of this creation.

Furthermore, the operation interface module provides the borrower to input the static data and the dynamic data through social media or email or communication software or a combination thereof. In this way, it can be combined with the media that has been widely used to improve the convenience of this creation.

In addition, the static audit module is connected to a government database by telecommunications, such as a court judgment query system, and the static audit module calculates the static data and the government database to generate the first audit result, and includes the borrower's court judgment record. To improve the fairness of the credit rating review of this creation.

To sum up, the intelligent fund borrowing and reviewing system of this creation can remotely review the borrower's credit rating, so as to quickly and intelligently review the borrower, thereby helping the fund lending behavior. Among them, through this creation, there are multiple modules with intelligent calculations, and the two-stage static and dynamic audit modules and machine learning can be used to implement artificial intelligence to review the credit rating of the borrower. Improve the operational efficiency of subsequent borrowing behavior.

In order for your reviewers to understand the content of this creation clearly, I would like to refer to the following descriptions and drawings.

Please refer to FIG. 1, which is a system block diagram of a preferred embodiment of the present invention. As shown in the figure, the creative fund borrowing and smart auditing system 1 is used to remotely audit a credit rating of a borrower (not shown), including an operation interface module 10, a static auditing module 11, and Credit question module 12 and a dynamic review module 13. Among them, the static auditing module 11 is telecommunication connected to the operation interface module 10, and the credit questioning module 12 is telecommunication connected to the operation interface module 10 and the static auditing module 11, and the dynamic auditing module 13 is telecommunication connected The operation interface module 10, the static review module 11 and the credit questioning module 12. In addition, this creation is deployed or integrated in a computer host (not shown), a mobile communication device (not shown), or a cloud (not shown). At the same time, this creation also connects the telecommunications input interface (not shown) of the computer host or mobile communication device of the borrower, and can communicate with each other through wired or wireless communication. Therefore, the borrower can use a tablet computer or a smart phone to use the fund lending intelligent review system 1 and use the input interface in this device to transfer data with the fund lending smart review system 1 to audit the borrower. Of credit can continue to borrow. The input interface includes a display, a microphone, a speaker, a camera, etc., for the borrower to view, receive, or input data into the operation interface module 10.

Please also refer to FIG. 2A and FIG. 2B at the same time, which are flowcharts (1) and (2) of the system of the preferred embodiment of the creation, respectively. As shown in the figure, the operation interface module 10 is used to prompt the input interface to the borrower and to receive a static data and a dynamic data input by the borrower (step P1). The static audit module 11 receives the static data transmitted by the operation interface module 10 (step P2), and the static audit module 11 calculates the static data to generate a first audit result (step P3) to determine the Whether the borrower's preliminary rating is passed (step P4). And when the static review module 11 judges that the borrower's preliminary rating is not passed, the operation interface module 10 receives the first review result and prompts the input interface to display the result of its failure (step P5) . When the preliminary rating of the borrower is passed, the credit questioning module 12 receives and calculates the static data to generate a credit test question group (step P6). Then, the credit test question group is transmitted to the operation interface module 10 for the borrower to answer, and then the dynamic data is generated according to the physical response of the borrower (step P7). The dynamic auditing module 13 receives the dynamic data transmitted by the operation interface module 10 (step P8), and the dynamic auditing module 13 cross-references the static data with the dynamic data to generate a second audit result ( Step P9) for presenting the final rating of the borrower to determine whether the final rating of the borrower is passed (step P10). And when the final rating of the borrower is a failure, the operation interface module 10 receives the second audit result, and prompts the input interface to display the result of the failure (step P11). On the other hand, when the final rating of the borrower is passed, the operation interface module 10 receives the second audit result, and prompts the input interface to display the result of the pass (step P12), and then proceeds with the lending behavior (Step P13).

Furthermore, the static data includes at least one digital file or at least one text message, such as the borrower's identity card, negative photocopy, personal information and contact information, and financial status and other supporting documents, and is converted into a scanned image file or other electronic file The method exists, and the static data can be input to the operation interface module 10 through the input interface. Therefore, the static audit module 11 can calculate the static data in advance, and determine the credit rating of the borrower based on the preliminary audit.

Then, after the preliminary assessment is passed, the second stage of the review operation is continued, that is, the dynamic review module 13 calculates the dynamic data generated by the borrower to respond to the credit test question group, and generates the dynamic data to determine The borrower's final rating. For example, when the borrower uses the fund lending intelligent auditing system 1 through a tablet computer, the screen of the tablet computer will pop up a dialog box to prompt the borrower to start answering the credit test question group. The borrower can directly answer, and the camera of the tablet computer will record the image of the borrower, which is the dynamic data. The dynamic data includes at least one voice information or at least one image information, including the content of the borrower's response, its voice, and facial expressions, such as behavior patterns of the borrower's eyebrows, mouth opening, and talking. And the dynamic review module 13 can determine whether the borrower has lied or falsified information when answering the credit test question group through machine learning experience, as one of the borrower's credit judgment basis.

In this way, the author has a two-stage review operation, including preliminary static review and final dynamic review. And when the preliminary rating judged by the static review module 11 is passed, the party continues to perform the final rating operation of the dynamic review module 13. Therefore, the fund lending intelligent auditing system 1 can implement artificial intelligence through machine learning through multiple modules with intelligent calculations, and through multiple test calculations of known authenticity, it can audit the borrower's credit Rating. The better one, especially applied to online lending (Peer-to-peer lending, P2P Lending) and assessing its credit risk against the borrower, and the more people use it, the more intelligent calculation of each module can be improved The accuracy and efficiency of the borrower's credit rating to further improve the efficiency of subsequent fund lending.

In another embodiment of the present invention, the credit test question group has a plurality of questions, and the second audit result includes a plurality of answer question review results to judge the credit of the answers to the questions correspondingly. Preferably, the latter question is generated by the credit questioning module 12 after receiving and calculating the previous question and its corresponding answer, and the dynamic review module 13 calculates the answer based on the corresponding answers to these questions. Calculate the static data and the dynamic data, and then sequentially generate the answer review results corresponding to the questions. Therefore, when the borrower answers the credit test question group, he answers these questions one question at a time, and at the same time produces corresponding examination results of these answers. In addition, the latter question is related to the previous question and its corresponding answer, and the review result of each answer is a single credit rating showing each question and its answer. This can be used to check whether the borrower is lying or forging one by one step by step, and can gradually review the authenticity of each of the answers, thereby further increasing the fairness and accuracy of the final review stage.

As shown in FIG. 3, the fund lending intelligent auditing system 1 further includes a tracking fact module 14 and a back-end processing module 15. As shown in the figure, the tracking fact module 14 is connected to the static auditing module 11, the credit questioning module 12 and the dynamic auditing module 13, and the back-end processing module 15 is connected to the operation interface module. 10. The static review module 11, the credit questioning module 12, the dynamic review module 13, and the tracking fact module 14. In addition, the dynamic review module 13 further includes an alert unit 131, a contract unit 132, and an environmental background unit 133. The detailed description of the tracking fact module 14, the back-end processing module 15, the warning unit 131, the contract unit 132, and the environmental background unit 133 are as follows.

Please refer to FIGS. 2A and 2B and FIGS. 3 to 5 again. FIGS. 4 and 5 are part of a system flowchart (a) and (b) of another embodiment of the creation, and both continue from FIG. 2A or FIG. 2B . As shown in FIG. 4, it is the continuation of step P5 in FIG. 2A and step P11 in FIG. 2B, respectively, when the preliminary rating of the first review result and the final rating of the second review result of the borrower fail. At this time, the warning unit 131 generates a warning message (step P14). The back-end processing module 15 receives the warning message for the maintenance personnel (not shown) to immediately know the credit rating result of the borrower (step P15), and the maintenance personnel judge the first audit result or Whether the result of the second review is obviously distorted (step P16), that is, the maintenance operator reviews and judges whether the result of the intelligent calculation of the creation is significantly distorted. When the calculation result is not significantly distorted, the maintenance personnel ignore the warning message and do nothing (step P17).

On the other hand, when the calculation result is obviously distorted, for example, the dynamic data shows that the borrower sneezes, resulting in its final rating as untrustworthy, but the static data can prove that the borrower did not lie. Or, for example, the static information shows that the borrower has a judgment on non-performance of damages and other events, which led to its initial rating as failure, but it is verified that the party involved is not the borrower by comparing with other information. The maintenance personnel can directly modify the judgment result through the back-end processing module 15 to return that the borrower has not lied or has not acted in bad faith. Among them, the back-end processing module 15 is used by the maintenance staff to correct the first review result (step P18), or the back-end processing module 15 is used by the maintenance staff to correct the answer review results (step P19), and thereafter The end processing module 15 generates a second correction information (step P20) for correcting the credit rating of the borrower. Then, the static review module 11, the credit questioning module 12, and the dynamic review module 13 respectively receive and calculate the second correction information for updating through machine learning (step P21). Therefore, the maintenance staff will use human assistance to determine whether the borrower is lying or falsifying the data, and feedback the real situation to each intelligent calculation module, thereby assisting the machine learning of each module, and further improving the accuracy of its intelligent calculation. And effectiveness to get more fair and accurate credit rating results.

In addition, following step P13 in FIG. 2B, when the borrower finally approves and passes the lending behavior, part of the process is shown in FIG. A tracking time is set in the tracking fact module 14 and after the tracking time has passed (step P22), the tracking fact module 14 accesses the borrower's repayment record (step P23), and the tracking fact module 14 generates a first correction information (step P24). The tracking time can be 30 days or the 5th of each month as an example. The borrower's repayment status can be continuously tracked to identify whether the borrower has the ability to repay and infer the credit of each module to the borrower. The rating judgement is correct and knows its true credit. For example, the next month after the borrower receives the funds, if the borrower does not pay back, the first correction information shows that the borrower is actually not creditworthy. Conversely, if the borrower repays the money on time, the first correction information is displayed as credit information. Therefore, the first correction information returns the true status of the borrower's credit for correction of the credit evaluation results and calculation methods of each module. Then, the static auditing module 11, the credit questioning module 12, and the dynamic auditing module 13 respectively receive and calculate the first correction information to be updated through machine learning (step P25). Therefore, the tracking fact module 14 can learn the real credit of the borrower and give it back to other intelligent modules, thereby training its machine learning to improve the accuracy and efficiency of its intelligent calculations.

Following the above description, when the tracking fact module 14 generates the first correction information (step P24), the environmental background unit 133 calculates the dynamic data to generate environmental information (step P26), and then the environmental background unit 133 calculates the The environmental information and the first correction information generate a third correction information (step P27). The environmental information includes the environment or state of the borrower when entering the dynamic data, such as a coffee shop, a home with construction noise, and the like. In this way, the correlation between the environmental information and the dynamic information is taken into consideration by the first correction information before the real credit of the borrower is known. For example, in the system's machine learning process, if the real situation is when the environmental information is a coffee shop, and the information returned by the first correction information is credited as expected, the environmental information is When at home, the first correction information is that the money was not repaid as scheduled without credit. Therefore, when the credit rating of the borrower is reviewed, the dynamic review module 13 can reasonably infer that the environmental information is a coffee shop, and the borrower will repay the loan as scheduled in the future and should have a higher credit rating. Or in another example, during the system's machine learning process, if the environmental information includes the sound of turning the paper, and when the first correction information is returned, the borrower has no credit but does not have credit; otherwise, it has credit. With this, when the borrower uses the system, the dynamic auditing module 13 can learn and recognize whether there is a sound of turning paper in the dynamic information, so as to further judge whether the borrower forged self-credit because of looking at written false information. Rating. Therefore, the dynamic audit module 13 considers the environmental information separately as a parameter of another intelligent calculation, and can generate a more rigorous credit rating. Then, the static auditing module 11, the credit questioning module 12, and the dynamic auditing module 13 respectively receive and calculate the third correction information for updating through machine learning (step P28). In this way, the artificial intelligence of this creation can consider more factors to obtain a fairer and more accurate credit rating.

Furthermore, when performing a borrowing behavior (step P13), the contract unit 132 calculates the static data to generate a contract, and provides a prompt to the operation interface module 10, provides the borrower and can sign the contract through the input interface. . In this way, the borrower can quickly sign a contract to complete its credit protection procedures, and can continue to carry out transactions of funds to facilitate the completion of borrowing.

In addition, in various embodiments, the borrower can use the fund lending intelligent review system 1 through a variety of software media, especially because the software media has registered and used the borrower. The software media includes social media such as Facebook Messenger and other communication software or web pages for the borrower to enter the static data and the dynamic data by filling out a form or having a conversation. For example, the borrower can log in to his Facebook through a smart phone, and transmit the static data and the dynamic data through Facebook's interface to conduct credit rating review and sign a contract without entering a new Internet Create a new account on the web page. In this way, the fund lending intelligent auditing system 1 can also combine software media that has been commonly used, thereby improving its convenience.

Preferably, the static review module 11 is further connected to a government database (not shown in the figure) by telecommunications, such as a government information disclosure platform and a court judgment inquiry system. Therefore, the static review module 11 can access information in the government database, including the court's judgment record of the borrower. In this way, the static audit module 11 calculates the static data and the data of the government database to generate the first audit result as the basis for the borrower's credit rating, thereby improving the accuracy of the intelligent audit to obtain a more fair And accurate credit rating.

To sum up, the intelligent review system for fund lending proposed by this creation can remotely review the credit rating of the borrower to quickly and intelligently determine whether the borrower has good credit, which will help subsequent borrowing. Trading behavior. Among them, through the static auditing module 11 and the dynamic auditing module 13, it is a two-stage audit judgment operation to evaluate the credit rating of the borrower. In addition, this creation uses machine learning to implement artificial intelligence, calculate the static data and dynamic data input by the borrower, and subsequent repayment records of the borrower, which can further improve the accuracy of the intelligent calculation of each module. And efficiency, thereby improving the accuracy of the borrower's credit rating.

However, the above are only the preferred embodiments of this creation, and are not intended to limit the scope of implementation of this creation; therefore, all equal changes and modifications made without departing from the spirit and scope of this creation should be covered in Within the scope of the patent of this creation.

1‧‧‧Fund Loan Intelligent Audit System

10‧‧‧ Operation interface module

11‧‧‧Static Audit Module

12‧‧‧ Credit Request Question Module

13‧‧‧Dynamic Audit Module

131‧‧‧Warning unit

132‧‧‧ contract unit

133‧‧‧Environmental background unit

14‧‧‧Track Facts Module

15‧‧‧back-end processing module

P1 ~ P28‧‧‧step

FIG. 1 is a system block diagram of a preferred embodiment of the present invention. FIG. 2A is a system flowchart (1) of a preferred embodiment of the present invention. FIG. 2B is a system flowchart (2) of a preferred embodiment of the present invention. FIG. 3 is a block diagram of a system according to another embodiment of the present invention. FIG. 4 is a partial system flowchart (1) of another embodiment of the authoring. FIG. 5 is a partial system flowchart (2) of another embodiment of the authoring.

Claims (10)

An intelligent review system for fund lending for remotely reviewing a credit rating of a borrower includes: an operation interface module for prompting an input interface to the borrower, and for receiving one of the inputs from the borrower Static data and dynamic data, and the static data includes at least one digital file or at least one text message, the dynamic data includes at least one sound information or at least one image information; a static audit module, which is a telecommunication connected to the operation interface module , The static audit module receives the static data transmitted by the operation interface module, and the static audit module calculates the static data to generate a first audit result to determine whether the preliminary rating of the borrower is passed or Failed; a credit questioning module, which is the telecommunications connection to the operation interface module and the static audit module, when the static audit module judges that the borrower's preliminary rating is passed, the credit questioning module is Receiving and calculating the static data to generate a credit test question group, which is transmitted to the operation interface module for the borrower to answer, The dynamic data is generated in response to the physical response of the borrower; and a dynamic audit module, which is a telecommunications connection to the operation interface module, the static audit module, and the credit questioning module, the dynamic audit module receives The dynamic data transmitted by the operation interface module, and the dynamic audit module cross-references the static data with the dynamic data, and then generates a second audit result for presenting the final rating of the borrower; among them, The dynamic review module judges the authenticity of the credit test group based on the borrower's experience through machine learning experience. As described in item 1 of the scope of patent application, the intelligent review system for fund borrowing and lending further includes a tracking fact module, which is connected to the static audit module, the credit questioning module and the dynamic audit module by telecommunications, and After one of the tracking time settings of the tracking fact module, the tracking fact module accesses the borrower's repayment record and generates a first correction information for correcting the borrower's credit rating, so that the static review module The credit questioning module and the dynamic review module respectively receive and calculate the first correction information for updating through machine learning. As described in item 2 of the scope of patent application for the intelligent review of funds borrowing and lending system, wherein the credit test question group has a plurality of questions, and the borrower answers each of the questions in sequence; of which, the latter question is determined by the The credit questioning module receives and calculates the previous question and its corresponding answer and then generates it. According to the intelligent review system for fund borrowing and lending as described in item 3 of the scope of patent application, wherein the second audit result includes a plurality of answers to the audit results, and the dynamic audit module calculates the static data according to the answers corresponding to the questions and This dynamic data, in turn, produces the answer review results corresponding to the questions. As described in item 4 of the scope of patent application, the intelligent review system for fund borrowing and lending also includes a back-end processing module, which is connected to the operation interface module, the static audit module, the credit questioning module, and the dynamics by telecommunications. The audit module and the tracking fact module are used by maintenance personnel to correct the first audit result or the answer audit results through the back-end processing module, and generate a second correction information for amending the borrower's credit The rating enables the static audit module, the credit questioning module, and the dynamic audit module to receive and calculate the second correction information to be updated through machine learning. For example, the intelligent review system for fund borrowing as described in item 5 of the scope of patent application, wherein the dynamic review module has a warning unit. When the preliminary or final rating of the borrower fails, the warning unit is A warning message is generated, and the back-end processing module receives the warning message for maintenance personnel to know. For example, the intelligent review system for fund lending as described in item 6 of the patent application scope, wherein the dynamic review module has a contract unit, and when the final review result is passed, the borrower can sign the contract. As described in item 7 of the patent application scope, the intelligent review system for fund borrowing and lending, wherein the dynamic audit module further includes an environmental background unit that calculates the dynamic data to generate environmental information, and calculates the environmental information and the first A correction information generates a third correction information for amending the borrower's credit rating, so that the static audit module, the credit questioning module and the dynamic audit module respectively receive and calculate the third correction information To update through machine learning. As described in any one of claims 1 to 8 of the patent application scope, the intelligent review system for fund borrowing and lending, wherein the operation interface module provides the borrower input through social media or email or communication software or a combination thereof The static data and the dynamic data. According to the smart audit system for fund borrowing as described in item 9 of the scope of patent application, wherein the static audit module is connected to a government database by telecommunications, and the static audit module calculates the static data and the government database to generate the first audit result. .