Credit risk assessment system and method thereof

Abstract

The credit risk assessment system includes a data acquisition module and a processing module. The processing module controls the data acquisition module to retrieve the target company's ESG information from a remote database based on the target company information and the content of the acquisition task. The ESG information is then labeled with ESG issue items according to an ESG issue item classification. The processing module responds to the issue item labeling results and calculates the target company's ESG score based on the ESG issue item classification weights.

Description

Credit risk assessment system and method The present invention relates to a credit risk assessment technology, and more particularly to a credit risk assessment system and method based on three major aspects: environmental protection, social responsibility, and corporate governance. Generally speaking, after a financial institution lends money to a borrower, it must monitor the borrower's credit risk or repayment ability to prevent the borrower from being unable to fulfill the debt contract and endangering the financial institution's creditor rights. ESG, which stands for Environmental Protection, Social Responsibility, and Corporate Governance, is considered an indicator for evaluating a company's operations. In the past, companies focused solely on financial data. However, if operations violate human rights, discharge wastewater, infringe on consumer rights, and endanger the living environment of Earth's species, while maintaining impressive financial statements, this would violate international goals for global sustainability, prompting institutional investors to reduce investment and cooperation with such companies and even damaging the company's reputation. Today, companies that prioritize ESG not only have transparent financial statements but also stable, low-risk operating models, resulting in relatively robust long-term performance. Therefore, how to further more efficiently assess corporate credit risk based on corporate ESG is one of the topics of concern to those skilled in the art. One embodiment of the present application provides a credit risk assessment system for assessing the credit risk of a target company based on ESG indicators, including: a data acquisition module coupled to a remote database for data acquisition; and a processing module coupled to the data acquisition module, wherein the processing module controls the data acquisition module to acquire ESG information of the target company from the remote database based on the target company information and an acquisition task content; the processing module issues the ESG information based on an ESG issue classification; and the processing module calculates an ESG score of the target company based on the ESG issue classification weight in response to the issue marking result. In some embodiments, the credit risk assessment system further includes an interactive interface coupled to the processing module for receiving the target company information and transmitting the information to the processing module. In some embodiments, the content of the extraction task includes a positive ESG information extraction task, a negative ESG information extraction task, and a positive and negative ESG information extraction task. In some embodiments, the remote database includes a database of the Ministry of Labor, a database of the Environmental Protection Agency, an external sustainable database, a public information database of a stock exchange, a corporate website, and a job bank database. In some embodiments, when the acquisition task content is the negative ESG information acquisition task, the processing module controls the data acquisition module to acquire the ESG information of the target company from the database of the Ministry of Labor, the database of the Environmental Protection Agency, and the external sustainability database. In some embodiments, the processing module executes a natural language processing method to tag the ESG information with topic items. In some embodiments, the target company information includes the target company name, target company ID, target company industry, and company website. In some embodiments, calculating the ESG score of the target company further includes multiplying the issue labeling results by the ESG issue classification weights. In some embodiments, the credit risk assessment system further includes a storage unit coupled to the processing module, storing the ESG information after the processing module calculates the ESG score of the target company. Another embodiment of the present application provides a credit risk assessment method for assessing the credit risk of a target company based on ESG indicators, including: controlling a data acquisition module to acquire ESG information of the target company from a remote database based on target company information and an acquisition task content; having a processing module perform ESG issue tagging on the ESG information based on an ESG issue classification; and in response to the issue tagging result, the processing module calculates an ESG score of the target company based on the ESG issue classification weight, and the processing module multiplies the issue tagging result by the ESG issue classification weight to calculate the ESG score. Thus, the credit risk assessment system and method in this case can extract positive and negative ESG information by inputting basic information about the target company, and assign tasks for extracting positive and negative ESG information. ESG information is then extracted based on the assigned tasks, expanding the scope of ESG scoring. Furthermore, through labeling training based on the classification criteria for ESG issues, automatic labeling and scoring are performed, significantly reducing labor costs. Furthermore, by assigning different weights to the same issue for different industries, the system can better reflect the current state of the industry and increase the accuracy of ESG scoring. The following disclosure provides a number of different embodiments or illustrations for implementing various features of the present invention. The components and configurations of particular illustrations are used to simplify the following discussion. Any illustrations discussed are for illustrative purposes only and are not intended to limit the scope or significance of the present invention or its illustrations in any way. Furthermore, the disclosure may repeat references to numerical symbols and / or letters in different illustrations. This repetition is for simplicity and clarification purposes and does not, in itself, specify the relationship between the different embodiments and / or configurations discussed below. Unless otherwise noted, terms used throughout the specification and claims generally have their ordinary meanings in the art, within the context of this disclosure, and in the specific context. Certain terms used to describe the present disclosure are discussed below or elsewhere in this specification to provide additional guidance to those skilled in the art regarding the description of the present disclosure. As used herein, "coupling" or "connection" may refer to two or more elements making direct physical or electrical contact with each other, or making indirect physical or electrical contact with each other, and "coupling" or "connection" may also refer to two or more elements operating or moving with each other. In this document, it is understood that the terms first, second, third, etc. are used to describe various elements, components, regions, layers and / or blocks. However, these elements, components, regions, layers and / or blocks should not be limited by these terms. These terms are limited to identifying single elements, components, regions, layers and / or blocks. Therefore, a first element, component, region, layer and / or block in the following text may also be referred to as a second element, component, region, layer and / or block without departing from the spirit of the present invention. As used herein, the term "and / or" includes any combination of one or more of the listed associated items. The "and / or" mentioned in this document refers to any combination of any one, all or at least one of the listed elements. FIG1 is a block diagram of an evaluation system for evaluating corporate credit risk based on ESG indicators according to an embodiment of the present case. ESG stands for Environmental Protection (Environment), Social Responsibility (Social) and Corporate Governance (Governance). Referring to FIG1, the credit risk evaluation system 100 provided in this embodiment includes a data acquisition module 102 and a processing module 104. In a preferred embodiment, the processing module 104 is coupled to the data acquisition module 102 to perform scoring based on the ESG content captured by the data acquisition module 102. In a preferred embodiment, the data acquisition module 102 can be connected to a remote database 110 in a wired or wireless manner for data acquisition. In a preferred embodiment, the evaluation system 100 of this case can be built on a host device such as a server, a personal computer, a notebook computer, an industrial computer, etc., but this case is not limited to this. In addition, the processing module 104 is at least one of a central processing unit, a graphics processing unit, an embedded system, a microsystem, a single chip, a special purpose chip, etc., but the present invention is not limited thereto. In a preferred embodiment, remote database 110 may be a database of the Ministry of Labor, the Environmental Protection Administration, an external sustainability database, a public information database of a stock exchange, a corporate database, a job bank database, etc. However, it is worth noting that this invention is not limited to these databases. In a preferred embodiment, data capture module 102 primarily captures recorded ESG penalty information from the Ministry of Labor and the Environmental Protection Administration databases, including regulations, penalty dates, company names, penalty details, penalty reasons, and penalty amounts. Data capture module 102 primarily captures recorded corporate governance information from external sustainability databases, including penalties against executives and senior management for insider trading, internal audit and internal control, money laundering, embezzlement, and breach of trust. Data capture module 102 primarily captures information related to a company's employee benefits from a job bank database. Data capture module 102 primarily captures a company's ESG implementation records and predetermined implementation targets from corporate databases, such as official websites. Generally speaking, companies listed on the stock exchange, over-the-counter (OTC), emerging stock exchanges, or publicly offered companies are obligated to disclose information such as financial reports, annual shareholder meeting reports, and perpetual sustainability reports, and upload them to public information databases for investor evaluation. Therefore, data capture module 102 primarily retrieves the perpetual sustainability reports uploaded by publicly offered companies from the public information databases of the stock exchanges, extracting important ESG keyword contexts. In a preferred embodiment, after extracting this data, data capture module 102 transmits the captured data back to processing module 104, which extracts the ESG keyword contexts through text mining. An ESG rating model then tags the captured ESG keyword contexts to determine the company's ESG rating based on the corresponding ESG issue topics. In a preferred embodiment, data is categorized by tagging the paragraphs surrounding an ESG keyword, as shown in the table below. If the paragraphs meet the criteria for the ESG issue, they are assigned a score of 0.5-1 based on the degree of achievement. Otherwise, they are assigned a score of 0. In a preferred embodiment, a score of 0 indicates non-achievement, 0.5 indicates qualitative policy and disclosure compliance, 0.8 indicates quantitative policy and disclosure compliance, 0.9 indicates certification, and 1 indicates an anticipated implementation of international SBTi reduction targets. In a preferred embodiment, the ESG rating model's criteria for categorizing ESG issues include renewable energy management, greenhouse gas management, water management, air management, sustainable supply chains, and information security. However, it is worth noting that this application is not limited to these criteria. Tagging and categorization involves tagging the paragraphs surrounding an ESG keyword based on the aforementioned ESG issue definitions to identify the ESG issue to which the paragraph relates: In this example, the use of renewable electricity and the pursuit of carbon neutrality and net-zero emissions are discussed. Therefore, automated extraction technology identifies the ESG keyword context related to renewable energy management and greenhouse gas management. Therefore, these two topics are assigned a score of 0.5-1 based on the degree of achievement. However, the ESG keyword context is unrelated to water management, air management, sustainable supply chains, and information security, so these topics are assigned a score of 0. Consequently, the ESG rating model uses the weighted summation of the ESG keyword context to determine the company's ESG score. In one embodiment, processing module 104 can control data acquisition module 102 to retrieve data from remote database 110 based on user input. In a preferred embodiment, the user input information includes a company name, company ID, company industry, and company website. Based on this information, data acquisition module 102 retrieves multiple ESG public information about the target company from remote database 110, the Ministry of Labor database, the Environmental Protection Administration database, external sustainability databases, public information databases of stock exchanges, corporate websites, and job bank databases. This information includes, but is not limited to, ESG penalty information about the target company, including laws and regulations, penalty date, penalty company name, penalty content, penalty reason, penalty amount, employee benefits information, ESG implementation records, planned implementation targets, and sustainability reports. After the data acquisition module 102 obtains multiple ESG public information about the target company, the processing module 104 extracts the paragraphs before and after the ESG keyword from the multiple ESG public information through text mining and performs an ESG issue classification. In a preferred embodiment, the processing module 104 marks the paragraphs before and after the ESG keyword according to the ESG rating model's ESG issue classification criteria to identify the ESG issues related to the paragraph and assign a score accordingly. In a preferred embodiment, the ESG rating model is used to assign scores to the various issue categories corresponding to the paragraphs before and after the ESG keyword, and the weighted sum of the various issue categories is used to obtain the corresponding company's ESG score. In a preferred embodiment, the processing module 104 can multiply each issue category by its corresponding weight and add the sum to obtain an ESG score. In this way, the evaluation system 100 can evaluate the ESG performance of the corresponding company based on this ESG score and assign a corresponding ESG rating. In some embodiments, this ESG score can be divided into four levels, but this is not limited to this. It is worth noting that the corresponding weighting values ​​can be adjusted based on the nature of the company. In other preferred embodiments, this system can score not only positive ESG information but also negative ESG information. In a preferred embodiment, before performing ESG scoring based on the ESG rating model, processing module 104 first performs ESG rating model tagging training to automatically tag the paragraphs surrounding ESG keywords. In one embodiment, processing module 104 first defines ESG issue criteria based on the ESG rating model. Using automated extraction technology, it extracts the paragraphs surrounding ESG keywords and performs multi-label classification on the data surrounding the ESG keyword paragraphs. The data is then labeled and classified as ESG issues using one or more labels to create a training dataset. An artificial intelligence network is trained based on this training dataset to perform natural language processing on the paragraphs surrounding ESG keywords and tag them according to the ESG issue definition criteria to create labeled data for the ESG rating model in this case. In a preferred embodiment, automated ESG keyword extraction techniques can include a lexicon comparison method, which uses an established lexicon to compare input documents and extract phrases that appear in the lexicon. A grammar analysis method uses a grammar analysis program based on natural language processing technology to analyze noun phrases in a document, then uses an established dictionary or corpus to filter out unsuitable words to extract the desired vocabulary. Alternatively, a statistical analysis method analyzes the document, counts the frequency of word occurrence, and extracts words whose frequency falls within a certain range. However, it is worth noting that this invention is not limited to the above methods. Continuing with FIG. 1 , in a preferred embodiment, the evaluation system 100 further includes an interactive interface 106 that is communicatively or electrically coupled to the processing module 104. In some embodiments, the interactive interface 106 may include an input interface and an output interface (not shown), such as a keyboard, microphone, mouse, or touchscreen display, for a user to input information, such as a company name. The interactive interface 106 converts the company name information entered by the user into an electrical signal and transmits the signal to the processing module 104. The processing module 104 controls the data acquisition module 102 to retrieve data from the remote database 110 based on the company name. In some embodiments, the interactive interface 106 may include an output interface (not shown), such as a display, speaker, or printer. The interactive interface 106 can receive the electrical signal from the processing module 104 and present the information corresponding to the electrical signal in the form of images, sounds, texts or physical forms through the output interface. In a preferred embodiment, the evaluation system 100 further includes a storage unit 108. The storage unit 108 can be internal or external memory, including, but not limited to, flash memory, EEPROM, a hard drive, or other storage devices. The storage unit 108 can be coupled to the processing module 104 and store a plurality of program instructions. The processing module 104 can access at least one instruction from the storage unit 108 and execute the at least one instruction to further implement the application defined by the at least one instruction. In this embodiment, the program defined by the at least one instruction is the process of executing an ESG rating model to perform a company assessment. Figure 2 shows a flowchart for executing an ESG rating model to conduct a company assessment according to one embodiment of this case. Please refer to Figures 1 and 2 simultaneously for the credit risk assessment process 200 of this case. When the assessment system 100 receives an assessment instruction, it first creates a target company task in step 201 based on the company information entered by the user in the interactive interface 106. The company information includes the target company name, target company identification number, target company industry, and company website. In a preferred embodiment, the target company tasks include tasks for scoring the target company's positive ESG information, tasks for scoring negative ESG information, and tasks for scoring both positive and negative ESG information. In this manner, the ESG rating model of this case can extract and score the target company's ESG data for the assigned tasks. Next, in step 202, for the assigned task, data acquisition module 102 is controlled to retrieve data on the target company from remote database 110. In a preferred embodiment, data acquisition module 102 obtains multiple public ESG information on the target company from remote database 110, the Ministry of Labor database, the Environmental Protection Administration database, external sustainability databases, public information databases of stock exchanges, corporate websites, and job bank databases. In a preferred embodiment, if the task involves scoring negative ESG information, since negative ESG information is not recorded in the public information databases of stock exchanges or corporate websites, it will only appear in penalty information in the Ministry of Labor database, the Environmental Protection Administration database, and external sustainability databases. Therefore, in order to improve the processing speed of data acquisition, this case further provides a selection step 203 to determine whether it is a negative ESG information scoring task. If it is a negative ESG information scoring task, this case will, in step 205, have the data acquisition module 102 acquire the target company's negative ESG information from the Ministry of Labor database, the Environmental Protection Administration database, and the external sustainability database, such as ESG penalty information, including laws and regulations, penalty date, name of the penalized company, penalty content, penalty reason, and penalty amount. Conversely, if the task is not a negative ESG rating task, data acquisition module 102 will first retrieve corresponding company information from the job bank database in step 206. Then, in step 207, it will retrieve positive ESG information, such as ESG performance records, planned performance targets, and ESG sustainability reports, from public information databases on stock exchanges and corporate websites. In a preferred embodiment, if the task is to score both positive and negative ESG information for the target company, data acquisition module 102 will simultaneously execute the data acquisition processes of steps 205 and 207, simultaneously capturing both positive and negative ESG information. In this way, the ESG rating model can perform ESG data acquisition and ESG rating of the target company for the assigned task. Next, in step 208, a tagging process is executed to classify ESG issues based on the ESG information obtained by the data acquisition module 102. In a preferred embodiment, after the data acquisition module 102 obtains multiple ESG information for the target company, the processing module 104 extracts the paragraphs before and after the ESG keywords in the multiple ESG information through text mining and performs an ESG issue classification. In a preferred embodiment, because negative ESG information is not scored for specific issues, the tagging process in step 208 is only performed on positive ESG information. Accordingly, the processing module 104 tags the paragraphs before and after an ESG keyword in the positive ESG information according to the ESG issue classification criteria based on the ESG rating model to identify the ESG issue related to the paragraph and to assign a score accordingly. Finally, in step 209, the processing module 104 executes a scoring process. In a preferred embodiment, the scores for the positive and negative ESG information are separately calculated and then summed to obtain the ESG score for the corresponding company. In a preferred embodiment, when processing module 104 scores negative ESG information, it assigns different weights to the penalties in the company's negative ESG information to calculate the negative ESG information score. When scoring the company's positive ESG information, it multiplies each ESG issue in the company's positive ESG information by its corresponding weight and adds them together to obtain the positive ESG information score. The negative ESG information score for the company is then multiplied by the corresponding industry weight, and the positive ESG information score is multiplied by the corresponding industry weight, and the sum is calculated to obtain the company's ESG score. In a preferred embodiment, if the data acquisition process in step 205 and step 207 fails, the data acquisition module 102 will return a acquisition failure message to the processing module 104, and the processing module 104 will reallocate tasks based on the failure message. In a preferred embodiment, after the processing module 104 executes a scoring procedure in step 209 , the processing module 104 further stores the corresponding ESG keyword context data captured by the data capture module 102 in the storage unit 108 , so as to provide the user with the ability to query the original ESG keyword context data. Another embodiment of the present application provides a computer storage medium storing computer executable instructions, wherein the computer executable instructions are used to implement the method of the above-mentioned multiple processing steps when executed by the processing module 104. Another embodiment of the present application provides a computer program, including computer-readable codes. When the computer-readable codes are executed in an electronic device, a processor in the electronic device executes a method for implementing the above-mentioned processing steps. In summary, this credit risk assessment system and method differs from traditional systems that only access client ESG penalty information through external data open platforms, the Ministry of Labor database, and the Environmental Protection Administration database. This system allows for the extraction of positive and negative ESG information, as well as the assignment of tasks for each, by inputting basic target company information. ESG information is then extracted based on the assigned tasks, expanding the scope of ESG scoring. Furthermore, through labeling training based on ESG issue classification criteria, automated labeling and scoring is performed, significantly reducing labor costs. Furthermore, by adjusting the scores of different issue items using weights, different weights are assigned to the same issue item for different industries. This allows for a more accurate interpretation of industry realities and enhances the accuracy of ESG scoring. Although the present invention has been disclosed in the form of an implementation method as described above, it is not intended to limit the present invention. Anyone skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the scope of the patent application attached hereto. 100: Credit risk assessment system 102: Data acquisition module 104: Processing module 106: Interactive interface 108: Storage unit 110: Remote database 200: Credit risk assessment process 201-209: Steps The accompanying drawings herein are incorporated into and constitute a part of this specification. These drawings illustrate embodiments consistent with the present invention and, together with the specification, serve to illustrate the technical solutions of the embodiments of the present invention. FIG1 is a block diagram of an assessment system for assessing corporate credit risk based on ESG indicators according to one embodiment of the present invention. FIG2 is a flow chart of an assessment method for assessing corporate credit risk based on ESG indicators according to one embodiment of the present invention. 100: Credit risk assessment system 102:Data Acquisition Module 104:Processing module 106:Interactive Interface 108: Storage unit 110: Remote Database

Claims

1. A credit risk assessment system for assessing the credit risk of a target company based on ESG indicators, comprising: A data acquisition module is coupled to a plurality of remote databases for data acquisition, wherein the plurality of remote databases are divided into a first database group that stores only negative ESG information and a second database group. The system also includes a processing module coupled to the data acquisition module. The processing module controls the data acquisition module to acquire ESG information about the target company from a plurality of remote databases based on the target company information and an acquisition task. The processing module further: determines whether the acquisition task content is a negative ESG information score; if the acquisition task content is a negative ESG information score, controls the data acquisition module to acquire negative ESG information about the target company only from the first database group; and if the acquisition task content is not a negative ESG information score, controls the data acquisition module to acquire positive ESG information about the target company only from the second database group; marks the positive ESG information as an ESG issue item according to an ESG issue item classification; and generates a positive ESG score for the target company based on the classification weight of the ESG issue item, in response to the marking result of the ESG issue item. A negative ESG score is generated for the target company based on one of the penalty information in the negative ESG information; and an ESG score is generated for the target company based on the positive ESG score and the negative ESG score.

2. The credit risk assessment system as described in claim 1 further includes an interactive interface coupled to the processing module for receiving information about the target company and transmitting it to the processing module.

3. The credit risk assessment system as described in claim 1, wherein the first database group includes at least the database of the Ministry of Labor, the database of the Environmental Protection Administration, and an external sustainability database.

4. The credit risk assessment system as described in claim 3, wherein the second database group includes at least a publicly available information database of a securities exchange, a corporate website, and a human resources bank database.

5. The credit risk assessment system as described in claim 1, wherein the processing module performs a natural language processing method to label the positive ESG information with topic tags.

6. The credit risk assessment system as described in claim 1, wherein the target company information includes the target company name, the target company's unified identification number, the industry to which the target company belongs, and the company's website.

7. The credit risk assessment system as described in claim 1, wherein generating the positive ESG score for the target company further includes multiplying the issue item labeling result by the ESG issue item classification weight.

8. The credit risk assessment system as described in claim 1, further comprising a storage unit coupled to the processing module, which stores the ESG information after the processing module generates the ESG score of the target company.

9. A credit risk assessment method for evaluating the credit risk of a target company based on ESG indicators, comprising: Based on target company information and extraction task content, a processing module controls a data extraction module to extract ESG information of the target company from a plurality of remote databases. The plurality of remote databases are divided into a first database group storing only negative ESG information and a second database group. The process further includes: determining whether the extraction task content is a negative ESG information rating; responding that the extraction task content is the negative ESG information rating, controlling the data extraction module to extract only the target company's negative ESG information from the first database group; and responding that the extraction task content is not the negative ESG information rating, controlling the data extraction module to extract only the target company's positive ESG information from the second database group; and, based on an ESG issue category, the processing module marks the positive ESG information with ESG issue categories. In response to the labeling results of the ESG issue item, the processing module generates a positive ESG score for the target company based on the classification weight of the ESG issue item; generates a negative ESG score for the target company based on one of the penalty information in the negative ESG information; and generates an ESG score for the target company based on the positive ESG score and the negative ESG score.

Images