A financial business approval method and system based on multiple data sources

Abstract

The present invention discloses a financial business approval method and system based on multiple data sources, which relates to the field of business approval technology, including the system collecting and integrating all relevant data for initial screening of financial business data, and performing risk rating on the approval business after the initial screening; assigning the approval business to approval paths of different levels for preliminary approval based on the risk rating results, and performing secondary approval on the application that passes the preliminary approval; collecting data and decision results in the approval process, and constructing a decision approval library to identify approval cases and abnormal situations to adjust approval rules. The present invention can collect data from multiple data sources to achieve interconnection at the data level, improve the transparency and credibility of the approval process, and enhance the objectivity and accuracy of decision-making. Improve the overall approval efficiency, achieve the optimal use of approval resources, reduce approval time, and improve customer satisfaction. Improve the overall efficiency and market competitiveness of the approval system.

Description

Technical Field

[0001] The present invention relates to the technical field of business approval, and in particular to a financial business approval method and system based on multiple data sources. Background Art

[0002] Financial business approval has always been a key link in the financial market. With the rapid development of information technology, the approval process has gradually shifted from traditional paper-based and manual processes to digital and automated processes. In recent years, the widespread application of emerging technologies such as big data, artificial intelligence, and cloud computing has greatly promoted the efficiency and accuracy of financial approval. Big data analysis technology can extract valuable information from massive data and provide a basis for approval decisions; artificial intelligence, especially machine learning algorithms, can automatically identify potential risks and optimize approval strategies by learning from historical approval cases; cloud computing provides powerful computing resources and storage space for the approval process, supporting real-time, large-scale data processing.

[0003] However, despite significant technological advances, existing financial business approval methods still have some limitations. On the one hand, traditional approval processes often rely on a single data source and lack comprehensive consideration of multiple information, which can easily lead to one-sidedness and lag in approval decisions. On the other hand, the formulation and adjustment of approval rules are mostly based on empirical judgment, lacking systematicity and dynamism, and are difficult to adapt to the rapid changes in the market environment. In addition, the mechanism for identifying and handling abnormal situations in the approval process is not perfect, which affects the accuracy and efficiency of approval. For example, the accumulation and analysis of abnormal approval cases is insufficient, resulting in a lack of data support for the optimization of approval rules, and limited flexibility and adaptability of the approval path. Summary of the invention

[0004] In view of the problems existing in the existing financial business approval method based on multiple data sources, the present invention is proposed. Therefore, the problem to be solved by the present invention is how to provide a financial business approval method and system based on multiple data sources.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In the first aspect, the present invention provides a financial business approval method based on multiple data sources, which includes: systematically collecting and integrating all relevant data for initial screening of financial business data, and conducting risk rating on the approval business after the initial screening; allocating the approval business to different levels of approval paths for preliminary approval based on the risk rating results, and conducting secondary approval on applications that pass the preliminary approval; collecting data and decision results in the approval process, building a decision approval library to identify approval cases and adjust approval rules under abnormal circumstances.

[0007] As a preferred solution of the financial business approval method based on multiple data sources described in the present invention, the initial screening of financial business data includes the following steps: identifying data sources, including internal data sources and external data sources, connecting with external service providers through APIs, obtaining financial business data in real time or periodically, and cleaning the obtained data; ensuring that each record is calculated only once, standardizing the data format, manually verifying it, or using statistical methods to identify and process outliers; mapping data fields from different sources to a unified data model, merging data from different systems, and creating a data warehouse as a data storage center; performing an initial screening of financial business data according to screening rules, and applications with credit scores below the minimum score are directly transferred to a high-risk business approval process, and for situations that the rule engine cannot determine, the screening results are transferred to manual review and analysis.

[0008] As a preferred solution of the financial business approval method based on multiple data sources described in the present invention, wherein: the risk rating of the approval business after preliminary screening includes the following steps: collecting all past approval business cases, given a document set D and a vocabulary table V, each document d∈D is represented as a vector Vocabulary V = [v1, v2, …, v n ] contains all the different words that appear in the documents; is the vector representation of document d, f n Represents vocabulary v n Frequency in document d; for document d and vocabulary v, calculate the term frequency and inverse document frequency, the calculation formula is:

[0009]

[0010] tfidf(v,d,D)=f(v,d)×idf(v,D)

[0011] In the formula, f(v,d) represents the frequency of occurrence of word v in document d, that is, word frequency, d t represents the number of occurrences of word t in document d, d z represents the total number of words in document d, idf(v,D) represents the inverse document frequency, D is the size of the document set, df(v,D) represents the number of documents in the document set D that contain the word v, and tfidf(v,d,D) represents the TF value of the word v; extract keywords and classify business types for each case, and establish a historical database; set risk weights for different keywords and business types based on historical data, and conduct risk assessment for new business approval; extract keywords and identify business types from the description of the business to be approved, use the text similarity algorithm to compare the similarity between the new business and historical cases, and calculate the similarity between the word frequency of the new approved business and the historical approved business. The calculation formula is:

[0012]

[0013] In the formula, S i represents the similarity score between the i-th new approval business and the historical case set; is the vector representation of document a; perform risk score calculation, calculate the weighted similarity score of each historical case, calculate the average risk level of cases with similar characteristics based on historical risk events and approval results, and weighted sum the average risk levels to obtain the final risk score. The calculation formula is:

[0014]

[0015] In the formula, R n Represents the final risk score of the newly approved business; r i Indicates the risk score of each approval case; conducts business risk rating, calculates risk threshold based on final risk score, divides approval business risk level, determines approval decision based on risk rating, and allocates approval paths of different levels for preliminary approval.

[0016] As a preferred solution of the financial business approval method based on multiple data sources described in the present invention, the calculation formula of the risk threshold is:

[0017]

[0018] In the formula, R t is the risk threshold, w j is the final risk score weight factor for the j-th approval business, and N is the number of business approvals.

[0019] As a preferred solution of the financial business approval method based on multiple data sources described in the present invention, wherein: the allocation of the approval business to different levels of approval paths for preliminary approval based on the risk rating results includes: when the final risk score R of the new approval business n At the risk threshold R t If the final risk score R of the new approval business is less than 10%, the new approval business is judged as a low-risk approval business, and the most similar approval case is found in the approval business database to determine whether the most similar approval case is a low-risk approval business. If the approval results of the historical cases are all low-risk approval businesses, the new business is automatically approved, the approval decision and the basis case are updated in the database, and the approval result notification is sent to the applicant. When the final risk score R of the new approval business is less than 10%, the new business is automatically approved, and the approval decision and the basis case are updated in the database, and the approval result notification is sent to the applicant. n Greater than the risk threshold R t, then the current new approval business is judged to be a high-risk approval business, and similar approval cases in the approval business database are searched to determine whether similar approval cases are all high-risk approval businesses. If similar approval cases are all high-risk approval businesses, the approval business will be transferred to manual approval. After the approval is completed, the approval event and results will be added to the approval business database. If there is a low-risk approval business in similar approval cases, the risk threshold will be recalculated, the final risk score weight factor of the approval business will be adjusted, and the new approval business will be divided into approval business risk levels for the second time.

[0020] As a preferred solution of the financial business approval method based on multiple data sources described in the present invention, the secondary approval of the application that has passed the preliminary approval includes the following steps: recalculating the initial risk score and risk threshold of the new approval business; for the approval business that is initially classified as low-risk, the system automatically searches for the most similar approval cases in the database based on keywords, business types and historical scoring data, and evaluates the approval results of the matching cases to confirm whether they are all low-risk approval businesses; if all similar cases are low-risk approval businesses, the system automatically approves the new business and records the decision basis; for the approval business that is initially classified as high-risk, the system continues to analyze market intelligence, credit reports and financial status, identifies cases with similar risk characteristics to the current business in the approval database, and analyzes the approval process and results; enables the expert system to provide rule-based decision support to identify potential risk points and compliance issues, and selects the most appropriate approval path based on the in-depth analysis results and expert system recommendations; adjusts the approval rules and decision-making process based on the approval results to update the approval case database.

[0021] As a preferred solution of the financial business approval method based on multiple data sources described in the present invention, the construction of a decision-making approval library to identify approval cases and abnormal situations to adjust approval rules includes the following steps: collecting data from various links of the approval process, integrating external data sources to build a database; importing the approved business data into the database, adding indexes and labels to the data; classifying approval cases according to approval results and business types; using statistical methods and machine learning algorithms to identify abnormal situations in the approval process, analyzing the connections between cases to identify common patterns and trends; conducting compliance reviews of approval businesses to ensure that the approval process complies with relevant laws, regulations and internal policies, and identifying potential compliance risks; adjusting approval rules, collecting feedback after trying out the adjusted approval rules, formally deploying the verified approval rules into the approval process, collecting feedback from approvers and applicants, and improving approval rules and processes.

[0022] In the second aspect, the present invention provides a financial business approval system based on multiple data sources, which includes: a screening module, which is used to systematically collect and integrate all relevant data for initial screening of financial business data, and to perform risk rating on the approval business after the initial screening; an approval module, which is used to assign the approval business to different levels of approval paths for preliminary approval based on the risk rating results, and to conduct a secondary approval on applications that pass the preliminary approval; an update module, which is used to collect data and decision results in the approval process, build a decision approval library to identify approval cases and adjust approval rules in abnormal situations.

[0023] In a third aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, the steps of a financial business approval method based on multiple data sources are implemented.

[0024] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, the steps of a financial business approval method based on multiple data sources are implemented.

[0025] The beneficial effects of the present invention are that data can be collected from multiple data sources to achieve data-level interconnection, improve the transparency and credibility of the approval process, and enhance the objectivity and accuracy of decision-making. It can improve the overall approval efficiency, achieve optimal utilization of approval resources, reduce approval time, and improve customer satisfaction. It can also improve the overall efficiency and market competitiveness of the approval system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0027] Figure 1 The figure is a flowchart of a financial business approval method based on multiple data sources. DETAILED DESCRIPTION

[0028] In order to make the above-mentioned purposes, features and advantages of the present invention more understandable, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0029] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0030] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0031] Example 1

[0032] Reference Figure 1 , which is the first embodiment of the present invention, and provides a financial business approval method based on multiple data sources, including:

[0033] S1: The system collects and integrates all relevant data for initial screening of financial business data, and conducts risk rating for the approved business after the initial screening;

[0034] Specifically, it conducts initial intelligent screening and risk assessment of financial business data. Once the business data application is submitted, the system collects and integrates all relevant data, including but not limited to the applicant's financial statements, credit scores, industry trends, market position, historical transaction records, etc.

[0035] Identify data sources, including internal data sources and external data sources. Internal data sources include customer information, historical transaction records, account status, credit scores, financial statements, etc. External data sources may include credit reports, industry analysis reports, macroeconomic data, market trends, legal proceedings records, etc.

[0036] Connect with external service providers through APIs to obtain financial business data in real time or periodically, and clean the acquired data;

[0037] Ensure that each record is counted only once, standardize the data format, and ensure that all data can be understood on the same platform. Outlier handling: Identify and handle outliers, which may require manual verification or use of statistical methods to correct. Missing value filling: Fill missing data using predictive models or based on business rules.

[0038] Map data fields from different sources to a unified data model. Merge data from different systems; create a data warehouse or data lake as a center for data storage and analysis.

[0039] Perform initial screening of financial business data, set screening rules, and set credit score thresholds: set minimum credit score requirements, and applications below this threshold will be automatically rejected. Financial indicators: check financial ratios, such as debt-to-income ratio, current ratio, etc., and set restrictions. Transaction patterns: analyze transaction history and identify abnormal patterns, such as a large number of small transactions in a short period of time may indicate money laundering risks.

[0040] Financial business data is initially screened based on the screening rules. Applications with credit scores below the minimum are directly transferred to the high-risk business approval process. For situations that the rule engine cannot judge, the screening results are transferred for manual review and analysis.

[0041] Collect all past approval business cases, including business descriptions, keywords, approval results, risk events and other information.

[0042] Given a document set D and a vocabulary V, each document d∈D is represented as a vector Vocabulary V = [v1, v2, …, v n ] contains all the different words that appear in the documents; is the vector representation of document d, f n Represents vocabulary v n frequency in document d;

[0043] For document d and vocabulary v, calculate the term frequency and inverse document frequency, the calculation formula is:

[0044]

[0045] tfidf(v,d,D)=f(v,d)×idf(v,D)

[0046] In the formula, f(v,d) represents the frequency of occurrence of word v in document d, that is, word frequency, d t represents the number of occurrences of word t in document d, d z represents the total number of words in document d, idf(v,D) represents the inverse document frequency, D is the size of the document set, df(v,D) represents the number of documents in document set D that contain word v, and tfidf(v,d,D) represents the TF value of word v;

[0047] Keyword and business type labeling: Extract keywords and classify business types for each case. This can be done automatically through natural language processing technology to build a historical database.

[0048] Set risk weights for different keywords and business types based on historical data, and review and adjust weights regularly to reflect the latest market conditions and internal risk management strategies.

[0049] Conduct risk assessment for new business approval, extract keywords and identify business types from the business description to be approved, and use text similarity algorithm to compare the similarity between new business and historical cases. The formula for calculating the similarity between the word frequency of new approval business and historical approval business is:

[0050]

[0051] In the formula, S i represents the similarity score between the i-th new approval business and the historical case set; is the vector representation of document a;

[0052] Perform risk scoring calculations and calculate the weighted similarity score for each historical case. For each case with similar characteristics, calculate the average risk level based on its historical risk events and approval results, and add up these average risk levels to get the final risk score.

[0053]

[0054] In the formula, R n Represents the final risk score of the newly approved business; r i Indicates the risk score of each approval case;

[0055] Carry out business risk rating, calculate risk threshold according to final risk score, divide approval business risk level, decide approval decision according to risk rating, and allocate approval paths of different levels for preliminary approval; the calculation formula of risk threshold is:

[0056]

[0057] In the formula, R t is the risk threshold, w j is the final risk score weight factor for the j-th approval business, and N is the number of business approvals;

[0058] Regularly review approval results, update the weights of keywords and business types, and monitor system performance to ensure the accuracy of risk ratings.

[0059] S2: Based on the risk rating results, the approval business is allocated to different levels of approval paths for preliminary approval, and the applications that pass the preliminary approval are subject to secondary approval;

[0060] Specifically, based on risk rating, applications are automatically assigned to different levels of approval paths. Low-risk applications directly enter the automated decision-making process, and the system automatically generates approval results based on preset rules. For medium- and high-risk applications, the system will generate detailed approval documents and risk analysis reports, which will be sent to the corresponding level of approvers for manual review. After receiving the application, the approver will make a comprehensive judgment based on the approval guidelines and risk report, and may request additional materials, arrange interviews, or conduct on-site investigations.

[0061] The authority management of the reviewers ensures that only authorized personnel can access sensitive information and make decisions. The transparency of the review decisions records the basis and reasons for each review to facilitate subsequent audits and appeals. The accuracy and explainability of automated decisions ensure that the system decisions can be understood and verified.

[0062] When the final risk score of the newly approved business is R n At the risk threshold R t When the following conditions are met, the current new approval business is judged to be a low-risk approval business, and the most similar approval case is found in the approval business database to determine whether the most similar approval case is a low-risk approval business. If the approval results of the historical cases are all low-risk approval businesses, the new business is automatically approved, the approval decision and the basis case are updated in the database, and the approval result notification is sent to the applicant at the same time;

[0063] When the final risk score of the newly approved business is R n Greater than the risk threshold R t , then the current new approval business is judged to be a high-risk approval business, and similar approval cases in the approval business database are searched to determine whether similar approval cases are all high-risk approval businesses. If similar approval cases are all high-risk approval businesses, the approval business will be transferred to manual approval. After the approval is completed, the approval event and results will be added to the approval business database. If there is a low-risk approval business in similar approval cases, the risk threshold will be recalculated, the final risk score weight factor of the approval business will be adjusted, and the new approval business will be divided into approval business risk levels for the second time.

[0064] Recalculate the initial risk score and risk threshold for newly approved businesses;

[0065] For approval businesses that are initially classified as low-risk, the system automatically searches for the most similar approval cases in the database based on keywords, business types, and historical scoring data, and evaluates the approval results of matching cases to confirm whether they are all low-risk approval businesses;

[0066] If all similar cases are low-risk approval businesses, the system automatically approves the new business and records the basis for the decision;

[0067] For approval businesses that are initially classified as high-risk, the system continues to analyze market intelligence, credit reports, and financial status, identifies cases with similar risk characteristics to the current business in the approval database, and analyzes the approval process and results;

[0068] Enable expert systems to provide rule-based decision support, identify potential risk points and compliance issues, and automatically select the most appropriate approval path based on in-depth analysis results and expert system recommendations, including fast approval, standard approval, or senior approval committee review;

[0069] Adjust approval rules and decision-making processes based on approval results, and continuously update the approval case database, including approval decisions, analysis of basis, and feedback from manual approval.

[0070] S3: Collect data and decision results in the approval process, build a decision-making approval library to identify approval cases and abnormal situations to adjust approval rules.

[0071] Specifically, data is collected from all aspects of the approval process, including application information, approval decisions, approval time, feedback from approvers, approval results, etc. External data integration: Integrate external data sources, such as credit rating agencies, market trend analysis, industry reports, etc., to provide a more comprehensive approval perspective.

[0072] Build a database containing detailed information on approval cases, approval rules, decision results, and related analysis. Import approved business data into the database to ensure data integrity, consistency, and accessibility. Data indexing and labeling: Add indexes and labels to data for quick retrieval and classification queries.

[0073] According to the approval results and business types, the approval cases are classified, such as low-risk approval, high-risk approval, fast approval cases, etc. Statistical methods and machine learning algorithms, such as cluster analysis and outlier detection, are used to identify abnormalities in the approval process, such as long approval time, large deviation in approval decisions, etc. Case association analysis: Analyze the connection between cases and identify common patterns and trends, such as specific types of business are often rejected or fast approval cases under specific conditions.

[0074] Evaluate the efficiency of the approval process, identify bottlenecks and areas for improvement. Conduct compliance reviews of approval operations to ensure that the approval process complies with relevant laws, regulations and internal policies, and identify potential compliance risks.

[0075] Adjust approval rules, such as modifying risk scoring models, updating approval thresholds, optimizing approval paths, etc. Rule verification: Apply the adjusted approval rules in a small range, collect feedback, and verify their effectiveness and applicability. Rule deployment: Formal deployment of verified approval rules into the approval process to ensure that all approvers and systems follow the updated rules. Collect feedback from approvers and applicants to further improve approval rules and processes.

[0076] Furthermore, this embodiment also provides a financial business approval system based on multiple data sources, including: a screening module, which is used to systematically collect and integrate all relevant data for initial screening of financial business data, and to perform risk rating on the approval business after the initial screening; an approval module, which is used to assign the approval business to different levels of approval paths for preliminary approval based on the risk rating results, and to conduct a secondary approval on applications that pass the preliminary approval; an update module, which is used to collect data and decision results in the approval process, build a decision approval library to identify approval cases and adjust approval rules under abnormal circumstances.

[0077] This embodiment also provides a computer device, which is applicable to the case of a financial business approval method based on multiple data sources, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement all or part of the steps of the method described in the embodiment of the present invention as proposed in the above embodiment.

[0078] This embodiment also provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method in any optional implementation of the above embodiment is executed. Wherein, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk.

[0079] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0080] As can be seen from the above, this method can collect data from multiple data sources to achieve data-level interconnection and interoperability. Improve the transparency and credibility of the approval process and enhance the objectivity and accuracy of decision-making. It can quickly filter out applications that obviously do not meet the conditions, save resources in subsequent approval links, and improve overall approval efficiency. It enables high-risk businesses to be handled more prudently, while low-risk businesses can be approved through fast channels, balancing the needs of risk control and approval efficiency. It can achieve optimal utilization of approval resources, reduce approval time and improve customer satisfaction. Reducing the possibility of approval errors plays an important role in protecting financial institutions from potential risks. By identifying approval cases and abnormal situations, approval rules can be updated in a timely manner to ensure that approval decisions keep pace with the times, thereby improving the overall efficiency and market competitiveness of the approval system.

[0081] Example 2

[0082] Referring to Table 1, which is a second embodiment of the present invention, this embodiment provides a financial business approval method based on multiple data sources. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0083] Table 1 Technical characteristics comparison table

[0084]

[0085] As can be seen from the above, this method can collect data from multiple data sources to achieve data-level interconnection and interoperability. Improve the transparency and credibility of the approval process and enhance the objectivity and accuracy of decision-making. It can quickly filter out applications that obviously do not meet the conditions, save resources in subsequent approval links, and improve overall approval efficiency. It enables high-risk businesses to be handled more prudently, while low-risk businesses can be approved through fast channels, balancing the needs of risk control and approval efficiency. It can achieve optimal utilization of approval resources, reduce approval time and improve customer satisfaction. Reducing the possibility of approval errors plays an important role in protecting financial institutions from potential risks. By identifying approval cases and abnormal situations, approval rules can be updated in a timely manner to ensure that approval decisions keep pace with the times, thereby improving the overall efficiency and market competitiveness of the approval system.

[0086] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A financial business approval method based on multiple data sources, characterized by: include, The system collects and integrates all relevant data for initial screening of financial business data, and conducts risk rating on the approved businesses after the initial screening; Based on the risk rating results, the approval business is allocated to different levels of approval paths for preliminary approval, and the applications that pass the preliminary approval are subject to secondary approval; Collect data and decision results in the approval process, build a decision-making approval library to identify approval cases and abnormal situations to adjust approval rules; The risk rating of the approved business after the initial screening The following steps are included: Collect all past approval business cases, given a document set D and a vocabulary V, each document d∈D is represented as a vector Vocabulary V = [v1, v2, ..., v n ] contains all the different words that appear in the documents; is the vector representation of document d, f n Words to express v n frequency in document d; For document d and vocabulary v, calculate the term frequency and inverse document frequency, the calculation formula is: tfidf(v,d,D)=f(v,d)×idf(v,D) In the formula, f(v, d) represents the frequency of occurrence of word v in document d, that is, word frequency, d t represents the number of occurrences of word t in document d, d z represents the total number of words in document d, idf(v, D) represents the inverse document frequency, D is the size of the document set, df(v, D) represents the number of documents containing word v in document set D, and tfidf(v, d, D) represents the TF value of word v; Extract keywords and classify business types for each case and establish a historical database; Set risk weights for different keywords and business types based on historical data, and conduct risk assessment for new business approval; Extract keywords and identify business types from the description of the business to be approved, use the text similarity algorithm to compare the similarity between the new business and historical cases, and calculate the similarity between the word frequency of the new approval business and the historical approval business. The calculation formula is: In the formula, S i represents the similarity score between the i-th new approval business and the historical case set; is the vector representation of document a; Perform risk scoring calculations, calculate the weighted similarity score for each historical case, calculate the average risk level for cases with similar characteristics based on historical risk events and approval results, and weighted sum the average risk levels to obtain the final risk score. The calculation formula is: In the formula, R n Represents the final risk score of the newly approved business; r i Indicates the risk score of each approval case; Carry out business risk rating, calculate risk thresholds based on the final risk score, divide the approval business risk level, determine the approval decision based on the risk rating, and assign approval paths of different levels for preliminary approval; The calculation formula of the risk threshold is: In the formula, R t is the risk threshold, w j is the final risk score weight factor for the j-th approved business, and N is the number of business approvals; The allocation of approval business to different levels of approval paths for preliminary approval based on risk rating results includes: When the final risk score of the newly approved business is R n At the risk threshold R t When the following conditions are met, the current new approval business is judged to be a low-risk approval business, and the most similar approval case is found in the approval business database to determine whether the most similar approval case is a low-risk approval business. If the approval results of the historical cases are all low-risk approval businesses, the new business is automatically approved, the approval decision and the basis case are updated in the database, and the approval result notification is sent to the applicant at the same time; When the final risk score of the newly approved business is R n Greater than the risk threshold R t When the new approval business is judged as a high-risk approval business, similar approval cases in the approval business database are searched to determine whether the similar approval cases are all high-risk approval businesses. If the similar approval cases are all high-risk approval businesses, the approval business is transferred to manual approval. After the approval is completed, the approval event and results are added to the approval business database. If there is a low-risk approval business in the similar approval cases, the risk threshold is recalculated, the final risk score weight factor of the approval business is adjusted, and the new approval business is divided into the approval business risk level for the second time. The secondary review of the application that has passed the preliminary review includes the following steps: Recalculate the initial risk score and risk threshold for newly approved businesses; For approval businesses that are initially classified as low-risk, the system automatically searches for the most similar approval cases in the database based on keywords, business types, and historical scoring data, and evaluates the approval results of matching cases to confirm whether they are all low-risk approval businesses; If all similar cases are low-risk approval businesses, the system automatically approves the new business and records the basis for the decision; For approval businesses that are initially classified as high-risk, the system continues to analyze market intelligence, credit reports, and financial status, identifies cases with similar risk characteristics to the current business in the approval database, and analyzes the approval process and results; Enable expert systems to provide rule-based decision support to identify potential risk points and compliance issues, and select the most appropriate approval path based on in-depth analysis results and expert system recommendations; Adjust approval rules and decision-making processes based on approval results and update the approval case database.

2. The financial business approval method based on multiple data sources as claimed in claim 1, characterized in that: The initial screening of financial business data includes the following steps: Identify data sources, including internal and external data sources, connect with external service providers through APIs, obtain financial business data in real time or on a regular basis, and clean the obtained data; Ensure that each record is counted only once, standardize the data format, manually verify or use statistical methods to identify and deal with outliers; Map data fields from different sources to a unified data model, merge data from different systems, and create a data warehouse as a data storage center; Financial business data is initially screened based on the screening rules. Applications with credit scores below the minimum are directly transferred to the high-risk business approval process. For situations that the rule engine cannot judge, the screening results are transferred for manual review and analysis.

3. The financial business approval method based on multiple data sources as claimed in claim 2, characterized in that: The construction of the decision-making approval library to identify approval cases and adjust approval rules in abnormal situations includes the following steps: Collect data from all aspects of the approval process and integrate external data sources to build a database; Import the approved business data into the database and add indexes and tags to the data; Classify approval cases according to approval results and business types; Use statistical methods and machine learning algorithms to identify anomalies in the approval process and analyze connections between cases to identify common patterns and trends; Conduct compliance review of approval business to ensure that the approval process complies with relevant laws, regulations and internal policies, and identify potential compliance risks; Adjust the approval rules, collect feedback after trial use of the adjusted approval rules, formally deploy the verified approval rules into the approval process, collect feedback from approvers and applicants, and improve the approval rules and processes.

4. A financial business approval system based on multiple data sources, based on the financial business approval method based on multiple data sources according to any one of claims 1 to 3, characterized in that: include, The screening module is used to collect and integrate all relevant data for initial screening of financial business data and to conduct risk rating on the approved business after the initial screening; The approval module is used to allocate approval businesses to different levels of approval paths for preliminary approval based on risk rating results, and conduct secondary approval for applications that pass the preliminary approval; The update module is used to collect data and decision results in the approval process, build a decision-making approval library to identify approval cases and adjust approval rules in abnormal situations.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the financial business approval method based on multiple data sources described in any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the financial business approval method based on multiple data sources described in any one of claims 1 to 3 are implemented.

Images