Risk early warning rule intelligent generation method and device

Abstract

The embodiment of the invention provides a risk early warning rule intelligent generation method and device, and the method comprises the steps: obtaining the credit data of a user from a preset credit database, extracting the index characteristics in the credit data, carrying out the index relation recognition and entity semantic disambiguation operation of the index characteristics, determining a corresponding risk control index, and carrying out the risk early warning rule generation. Determining a corresponding credit entity according to the risk control index and a preset entity extraction model, performing dependency syntactic analysis operation and relation mode matching operation on the risk control index, and determining a corresponding risk control rule; constructing a corresponding risk control rule logic expression according to the risk control intention, the credit and loan entity and the risk control rule, performing risk control early warning condition judgment according to the risk control rule logic expression, and generating an after-loan early warning signal when the risk control early warning condition is triggered; the risk control early warning rule generation efficiency and accuracy can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and specifically to a method and device for intelligently generating risk warning rules. Background Art

[0002] With the rapid development of information technology, the demand for intelligent and automated risk warning systems is increasing. Traditional risk warning systems usually rely on manual setting and maintenance of rules, which has problems such as untimely rule updates, complex rule design, and poor adaptability.

[0003] The generation of existing risk warning rules mainly relies on expert experience and historical data. Manually setting rules is not only time-consuming and labor-intensive, but also difficult to cover all possible risk scenarios. Due to the static nature of the rules, the system cannot respond to new risk types in a timely manner, resulting in a lag in warnings. In addition, the design and maintenance of rules requires professional knowledge, which increases the complexity and maintenance costs of the system.

[0004] In terms of rule optimization, existing systems often lack effective automation tools, making it difficult to dynamically adjust and optimize rules. The effectiveness and accuracy of rules rely on a large amount of historical data for verification, but the limitations of data quality and volume make the rule optimization process complex and unstable. Traditional methods are unable to fully utilize big data and machine learning technologies, resulting in a low degree of intelligence in rules and an inability to achieve accurate risk warnings.

[0005] In terms of rule management and execution, existing systems usually adopt a centralized architecture, and the updating and distribution efficiency of rules is low. When the system faces a changing market environment and a large amount of risk data, the execution efficiency of rules and the system response speed are difficult to guarantee, which affects the overall early warning effect. The scalability and flexibility of rules are also limited, making it difficult to meet the changes in different application scenarios and business needs.

[0006] In general, the existing risk warning rule generation methods and devices have many defects in rule generation, optimization, management and execution, etc. These problems limit the intelligence level and application effect of the risk warning system and pose a higher challenge to the risk management capabilities of financial institutions. Summary of the invention

[0007] In response to the problems in the prior art, the present application provides a method and device for intelligently generating risk warning rules, which can effectively improve the generation efficiency and accuracy of risk control warning rules.

[0008] In order to solve at least one of the above problems, the present application provides the following technical solutions:

[0009] In a first aspect, the present application provides a method for intelligently generating risk warning rules, comprising:

[0010] Acquire the user's credit data from a preset credit database, extract indicator features from the credit data, perform indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determine corresponding risk control indicators;

[0011] Inputting the text features extracted from the risk control indicator into a preset intent recognition long short-term memory network to determine the corresponding prediction result, performing data enhancement operation according to the prediction result and a preset synonym dictionary to determine the corresponding risk control intent, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing dependency syntax analysis operation and relational pattern matching operation on the risk control indicator to determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization;

[0012] A corresponding risk control rule logic expression is constructed according to the risk control intention, the credit entity and the risk control rule, a risk control warning condition is determined according to the risk control rule logic expression, and a post-loan warning signal is generated when the risk control warning condition is triggered.

[0013] Furthermore, the determining of corresponding risk control intentions, credit entities and risk control rules according to the risk control indicators includes:

[0014] Inputting the risk control indicator into a preset intention recognition algorithm, and determining the corresponding risk control intention according to the prediction result output by the preset intention recognition algorithm and a preset synonym dictionary;

[0015] The corresponding credit entity is determined according to the risk control indicator and the preset entity extraction model, and the risk control indicator is subjected to dependency syntax analysis and relational pattern matching operations to determine the corresponding risk control rules.

[0016] Further, the risk control warning condition determination according to the risk control rule logical expression includes:

[0017] Extracting the conditional variables in the risk control rule logic expression, and determining the corresponding conditional values ​​according to the extracted conditional variables;

[0018] A corresponding condition combination is determined according to the condition variable and the condition value, and a risk control warning condition is determined according to the condition combination.

[0019] Furthermore, the determining the corresponding credit entity according to the risk control indicator and the preset entity extraction model includes:

[0020] Inputting the text features extracted from the risk control indicators into a preset entity extraction conditional random field model, performing a feature capture operation on the text features according to a feature function layer of the preset entity extraction conditional random field model, and determining corresponding capture features;

[0021] The captured features are input into the potential energy function layer of the preset entity extraction conditional random field model to determine the corresponding normalized potential energy, and the normalized potential energy is input into the output layer of the preset entity extraction conditional random field model to determine the corresponding credit entity.

[0022] Furthermore, performing dependency syntax analysis and relational pattern matching operations on the risk control indicators to determine corresponding risk control rules includes:

[0023] Performing dependency syntactic analysis on the risk control indicator to determine the corresponding word structure, performing a relational pattern matching operation based on preset logical connectives and the word structure to determine the corresponding rule matching graph;

[0024] A rule classification operation is performed according to the rule matching graph to determine the corresponding risk control rules.

[0025] Furthermore, the step of constructing a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule includes:

[0026] Constructing a basic expression according to the risk control intention and the credit entity;

[0027] A risk control rule logical expression is constructed according to the basic expression, the risk control rule and the preset nested logical rule.

[0028] In a second aspect, the present application provides a risk warning rule intelligent generation device, comprising:

[0029] A data preprocessing module is used to obtain the user's credit data from a preset credit database, extract indicator features in the credit data, perform indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determine the corresponding risk control indicators;

[0030] A rule generation module, used for inputting the text features extracted from the risk control indicator into a preset intent recognition long short-term memory network, determining the corresponding prediction result, performing a data enhancement operation according to the prediction result and a preset synonym dictionary, determining the corresponding risk control intent, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing a dependency syntax analysis operation and a relational pattern matching operation on the risk control indicator, and determining the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement, and sentence reorganization;

[0031] The risk control module is used to construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

[0032] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for intelligently generating risk warning rules when executing the program.

[0033] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for intelligently generating risk warning rules.

[0034] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the risk warning rule intelligent generation method.

[0035] It can be seen from the above technical scheme that the present application provides a method and device for intelligently generating risk warning rules, which obtains the user's credit data from a preset credit database, extracts indicator features in the credit data, performs indicator relationship recognition and entity semantic disambiguation operations on the indicator features, determines the corresponding risk control indicators, determines the corresponding credit entities based on the risk control indicators and the preset entity extraction model, performs dependency syntax analysis operations and relational pattern matching operations on the risk control indicators, and determines the corresponding risk control rules; constructs corresponding risk control rule logical expressions based on risk control intentions, credit entities and risk control rules, determines risk control warning conditions based on the risk control rule logical expressions, and generates a post-loan warning signal when the risk control warning conditions are triggered, which can effectively improve the generation efficiency and accuracy of risk control warning rules. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0037] Figure 1 This is one of the flow charts of the method for intelligently generating risk warning rules in an embodiment of the present application;

[0038] Figure 2 This is the second flow chart of the method for intelligently generating risk warning rules in the embodiment of the present application;

[0039] Figure 3 This is the third flow chart of the method for intelligently generating risk warning rules in the embodiment of the present application;

[0040] Figure 4This is a fourth flow chart of the method for intelligently generating risk warning rules in an embodiment of the present application;

[0041] Figure 5 This is a fifth flow chart of the method for intelligently generating risk warning rules in an embodiment of the present application;

[0042] Figure 6 This is the sixth flow chart of the method for intelligently generating risk warning rules in the embodiment of the present application;

[0043] Figure 7 This is a structural diagram of a risk warning rule intelligent generation device in an embodiment of the present application;

[0044] Figure 8 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application.

[0045] Reference numerals:

[0046] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0048] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0049] Taking into account the problem that there are many defects in the risk warning rule generation method and device in the prior art in terms of rule generation, optimization, management and execution, the present application provides a risk warning rule intelligent generation method and device, which obtains the user's credit data from a preset credit database, extracts indicator features in the credit data, performs indicator relationship identification and entity semantic disambiguation operations on the indicator features, determines the corresponding risk control indicator, determines the corresponding credit entity according to the risk control indicator and the preset entity extraction model, performs dependency syntax analysis operations and relational pattern matching operations on the risk control indicator, and determines the corresponding risk control rule; constructs the corresponding risk control rule logical expression according to the risk control intention, credit entity and risk control rule, determines the risk control warning condition according to the risk control rule logical expression, and generates a post-loan warning signal when the risk control warning condition is triggered, which can effectively improve the generation efficiency and accuracy of the risk control warning rule.

[0050] In order to effectively improve the generation efficiency and accuracy of risk control warning rules, this application provides an embodiment of a risk warning rule intelligent generation method, see Figure 1 The risk warning rule intelligent generation method specifically includes the following contents:

[0051] Step S101: obtaining the user's credit data from a preset credit database, extracting indicator features from the credit data, performing indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determining corresponding risk control indicators;

[0052] Optionally, in this embodiment, step S101 involves obtaining the user's credit data from a preset credit database, extracting the indicator features therein, and performing indicator relationship recognition and entity semantic disambiguation operations on these features, and finally determining the corresponding risk control indicators. The core of this step is to accurately extract and process the user's credit data in order to provide accurate basic data for subsequent risk control analysis. The following is a detailed description of step S101.

[0053] First, the system obtains the user's credit data from the preset credit database. The credit database usually contains a variety of information such as the user's credit report, loan record, repayment history, credit score, etc. These data are the key basis for assessing the user's credit risk. The system can efficiently extract the required credit data from the database through the preset interface and query statements. For example, the system extracts a user's credit report from the database, which contains "the user has applied for a credit card three times in the past two years, the current credit score is 700, and the repayment record is good."

[0054] Next, the system extracts indicator features from the extracted credit data. Indicator features refer to key data points that can reflect the user's credit status and risk level, such as the number of credit card applications, credit score, repayment record, etc. The system uses natural language processing (NLP) technology and text analysis methods to automatically identify and extract these indicator features from credit data. For example, in the above credit report, the system identifies that the "number of credit card applications" is 3 times, the "current credit score" is 700, and the "repayment record" is good. These feature data will serve as important inputs for subsequent risk control analysis.

[0055] After extracting the indicator features, the system performs indicator relationship identification and entity semantic disambiguation operations. The purpose of indicator relationship identification is to determine the relationship between different indicator features, such as temporal sequence, causal relationship, etc. For example, in the user's credit report, the system identifies that the time range in "applied for three credit cards in the past two years" is "the past two years", and associates it with the indicator "number of credit card applications" of 3 times. In addition, the system needs to perform entity semantic disambiguation operations to ensure that the identified features have clear semantics. For example, "700" in "credit score" refers to the user's credit score, not other meanings. Through semantic disambiguation, the system can accurately understand and process the meanings of different indicator features to avoid misjudgment and data confusion.

[0056] Through the above steps, the system finally determines the corresponding risk control indicators. These risk control indicators are key data that have been accurately extracted and processed, and can directly reflect the user's credit status and risk level. For example, the system determines that a user's risk control indicators are "3 credit card applications in the past two years", "Current credit score is 700", and "Good repayment record". These risk control indicators will serve as the basic data for subsequent risk control analysis to help the system conduct risk assessment and decision-making.

[0057] This embodiment achieves accurate processing and analysis of credit data by extracting user data from the credit database, performing indicator feature extraction, relationship identification, and semantic disambiguation. In the specific implementation process, the system uses NLP technology, text analysis algorithms, and semantic processing methods to ensure that the extracted indicator features are accurate, and the relationship identification and semantic disambiguation operations are effective and reliable. In this way, the system can efficiently process a large amount of user credit data, generate accurate risk control indicators, and provide a solid data foundation for subsequent risk assessment and decision-making.

[0058] In actual applications, the credit risk control system of a financial institution uses the above steps to perform refined processing on the user's credit data. For example, the system extracts the user's credit report from the database, automatically identifies and extracts key indicator features, such as the number of credit card applications, credit score, repayment record, etc. Next, the system identifies indicator relationships, determines the relationship between time ranges and features, and ensures that the semantics of each indicator feature is clear through semantic disambiguation operations. Finally, the system generates accurate risk control indicators, such as "3 credit card applications in the past two years", "Current credit score is 700", and "Good repayment record", providing reliable data support for subsequent risk assessment.

[0059] This embodiment not only improves the system's ability to process credit data, but also significantly enhances the accuracy and effectiveness of risk control analysis. Through automated and intelligent data processing, the system can analyze and process a large number of users' credit data in real time, ensuring the accuracy and timeliness of risk control decisions. Through continuous optimization and expansion, the system has good adaptability and scalability, and can process more complex and diverse credit data, providing reliable technical support for the risk control management of financial institutions.

[0060] Step S102: inputting the text features extracted from the risk control indicator into a preset intention recognition long short-term memory network to determine the corresponding prediction result, performing data enhancement operation according to the prediction result and a preset synonym dictionary to determine the corresponding risk control intention, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing dependency syntax analysis operation and relational pattern matching operation on the risk control indicator to determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization;

[0061] Optionally, in this embodiment, step S102 involves inputting the text features extracted from the risk control indicators into a preset intent recognition long short-term memory network (LSTM), determining the corresponding prediction results, and performing data enhancement operations based on the prediction results and the preset synonym dictionary to determine the corresponding risk control intent. Subsequently, the corresponding credit entity is determined based on the risk control indicators and the preset entity extraction model, and dependency syntax analysis and relational pattern matching operations are performed on the risk control indicators to finally determine the corresponding risk control rules. The following is a detailed description of step S102.

[0062] First, the system inputs the text features extracted from the risk control indicators into the preset intent recognition long short-term memory network (LSTM). LSTM is a special recursive neural network that is good at processing and predicting time series data. Its unique memory unit structure enables it to capture long-term dependencies. In this step, LSTM is used to identify the intent of text features. For example, the system extracts text features such as "3 credit card applications in the past two years", "Current credit score is 700", and "Good repayment record" from the user's credit report. By inputting these features, LSTM can predict the user's potential intentions, such as "high-risk credit behavior" or "low-risk credit behavior."

[0063] Next, the system performs data enhancement operations based on the prediction results and the preset synonym dictionary to determine the corresponding risk control intention. Data enhancement operations include synonym expansion, synonym replacement, and sentence reorganization, which can increase the diversity of data and improve the robustness and generalization ability of the model. For example, the system finds the synonym "credit score" for "credit score" in the synonym dictionary, and replaces "credit score" in the original text with "credit score" to generate a new data entry. In addition, the system can also reorganize sentence structures, such as reorganizing "the number of credit card applications in the past two years is 3 times" into "the number of credit card applications in the past two years is 3 times." These data enhancement operations help the system understand and process text features more comprehensively, and ultimately determine accurate risk control intentions.

[0064] After determining the risk control intent, the system determines the corresponding credit entity based on the risk control indicators and the preset entity extraction model. The entity extraction model can be trained to identify key entities in the text, such as "credit card", "loan", "repayment record", etc. For example, the system identifies credit entities such as "credit card", "credit score", "repayment record" from text features, and associates these entities with the corresponding risk control intent to form a complete risk control analysis framework.

[0065] Subsequently, the system performs dependency syntactic analysis and relational pattern matching operations on the risk control indicators to determine the corresponding risk control rules. Dependency syntactic analysis is used to reveal the grammatical relationships between features in the text, such as subject-predicate, object-complement, and modification. For example, in the sentence "The user's credit score is 700, and the repayment record is good", the system identifies "credit score" as the subject, "700" as the complement, "repayment record" as the subject, and "good" as the complement through dependency syntactic analysis. Based on these grammatical relationships, the system performs relational pattern matching operations to identify the logical relationships and patterns between features, such as parallel relationships, conditional relationships, etc. Finally, the system determines the corresponding risk control rules based on these relationships and patterns. For example, the system can formulate risk control rules for "low-risk users" based on the parallel relationship of "credit score is 700 and repayment record is good".

[0066] This embodiment achieves comprehensive analysis and processing of risk control indicators through intent recognition LSTM, data enhancement operation, entity extraction, dependency syntax analysis and relational pattern matching. In the specific implementation process, the system uses advanced deep learning algorithms and natural language processing technology to ensure the accuracy and efficiency of intent recognition and relational matching. Data enhancement operation improves the robustness and generalization ability of the model by increasing data diversity, thereby improving the reliability of the overall risk control analysis.

[0067] In actual applications, a financial institution conducts in-depth analysis and processing of users' credit data through the above steps. For example, the system extracts key text features from the user's credit report, inputs them into the LSTM model for intent recognition, and predicts the user's credit risk level. Based on the prediction results, the system further determines the user's risk control intention through data enhancement operations such as synonym expansion, synonym replacement, and sentence reorganization. Next, the system uses the entity extraction model to identify key credit entities in the text, such as "credit card", "loan", etc., and determines the logical relationship and matching pattern between these entities through dependency syntactic analysis and relational pattern matching, and finally generates specific risk control rules, such as "high-risk users should be further reviewed" or "low-risk users can be approved for loans."

[0068] This embodiment solves the problem of insufficient processing capabilities of complex text data in traditional risk control systems. By introducing the LSTM model for intent recognition, the system can capture the deep intentions of user behavior, and combined with data enhancement operations, the model's adaptability to diversified data is enhanced. Entity extraction and dependency syntactic analysis ensure accurate recognition and understanding of key information in the text, and relational pattern matching further strengthens the grasp of the logical relationship between features. The combination of this series of technical processes enables the system to generate accurate and reliable risk control rules, significantly improving the scientificity and effectiveness of risk control decisions.

[0069] In actual cases, for example, a user's credit record shows that he frequently applied for multiple loan products in the past year and had several overdue repayments. The system first extracts these text features from the database and inputs them into the LSTM model for intent recognition. The model predicts that the user has high-risk credit behavior. In order to further enhance the diversity of the data, the system expands and replaces "loan products" with "borrowing products" and "overdue repayments" with "delayed payments". At the same time, the system reorganizes the sentences of these texts, such as reorganizing "frequently applying for multiple loan products in the past year" into "frequently applying for multiple loan products in the past year". These enhanced data enable the system to have a more comprehensive understanding of user behavior.

[0070] Subsequently, the system uses the entity extraction model to identify key credit entities such as "loan products" and "overdue repayments" in the text, and determines the relationship between these entities through dependency syntactic analysis, such as the causal relationship between "frequent applications" and "overdue repayments". Through relational pattern matching, the system compares these relationships with the preset risk control rules, and ultimately determines that the user is a high-risk user and generates corresponding risk control decisions, such as recommending a stricter credit review of the user or directly rejecting his loan application.

[0071] Through this comprehensive technical solution, the system not only solves the problem of processing complex text data, but also ensures the accuracy of risk control rules and the reliability of decision-making through multi-level data enhancement and relationship matching. This not only improves the risk control efficiency of financial institutions, but also reduces potential credit risks and ensures the stable operation of credit business.

[0072] In summary, step S102 generates accurate risk control rules by inputting text features into LSTM for intent recognition, data enhancement, entity extraction, dependency syntax analysis, and relational pattern matching. Practical applications have shown that this solution significantly improves the risk control system's ability to process complex text data and its decision-making accuracy, providing solid technical support and guarantee for the credit business of financial institutions.

[0073] Step S103: construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

[0074] Optionally, in this embodiment, the main task of step S103 is to construct corresponding risk control rule logic expressions according to risk control intentions, credit entities and risk control rules, and to determine risk control warning conditions according to these logic expressions, so as to generate a post-loan warning signal when the warning conditions are triggered. The implementation of this step is crucial for timely discovery and prevention of potential credit risks. The following is a detailed description of step S103.

[0075] First, the system constructs the corresponding risk control rule logic expression based on the risk control intention, credit entity and risk control rules determined in the previous steps. The risk control rule logic expression is a formal rule representation method used to describe the relationship and conditions between different risk control elements. For example, suppose the risk control intention identified by the system is "high-risk credit behavior", the credit entities include "number of credit card applications", "credit score" and "repayment record", and the risk control rule is "If the user has applied for a credit card more than 3 times in the past year and the credit score is less than 650, and there is a record of overdue repayment, then he is judged as a high-risk user." Based on these elements, the risk control rule logic expression constructed by the system may be: IF (number of credit card applications>3AND credit score<650AND overdue repayment record exists) THEN high-risk user.

[0076] In the process of constructing logical expressions, the system needs to fully consider the specific values ​​and relationships of various risk control factors. For example, the system will obtain the user's actual credit card application times, credit score, and repayment record from the database, and substitute these values ​​into the logical expression for conditional judgment. This process requires the system to have strong data integration and processing capabilities to ensure the accuracy and effectiveness of the logical expression.

[0077] Next, the system determines the risk control warning conditions based on the constructed risk control rule logic expression. Risk control warning condition determination means that the system monitors and analyzes the user's credit behavior in real time based on the conditions of the logic expression. For example, the system determines whether a user has applied for a credit card more than 3 times in the past year, whether the credit score is less than 650, and whether there is a record of overdue repayment based on the logic expression. If all conditions are met, the system determines that the user has triggered the warning conditions for high-risk credit behavior.

[0078] When the risk control warning condition is triggered, the system generates a post-loan warning signal. The post-loan warning signal is a prompt message that reminds relevant personnel or system users that there is a potential credit risk and that corresponding measures need to be taken. For example, the warning signal generated by the system may include "User A has applied for a credit card more than 3 times in the past year, his credit score is less than 650, and there is a record of overdue repayments. It is recommended to conduct further credit assessment or take risk control measures." This warning signal will be sent to relevant risk control personnel or system managers so that they can take timely measures to prevent potential credit risks.

[0079] Through this embodiment, the system can achieve real-time monitoring and early warning of user credit behavior, thereby effectively preventing potential credit risks. In the specific implementation process, the system uses advanced logical expression construction technology and real-time data analysis capabilities to ensure the accuracy and timeliness of risk control early warning condition determination. The construction of risk control rule logical expressions and the early warning condition determination process require not only efficient data processing capabilities, but also powerful algorithm support to ensure accurate analysis and processing of complex credit data.

[0080] In actual applications, a financial institution uses the above steps to conduct real-time monitoring and early warning of users' credit behaviors. For example, the system constructs risk control rule logic expressions based on users' credit reports and transaction records, and makes real-time condition determinations. When a user's credit card application times, credit score, and repayment records meet the early warning conditions for high-risk behaviors, the system immediately generates a post-loan early warning signal and notifies relevant risk control personnel to take measures. In this way, financial institutions can promptly detect and prevent potential credit risks and ensure the safe and sound operation of credit business.

[0081] This embodiment not only solves the problem of insufficient processing capacity of complex credit data in traditional risk control systems, but also significantly improves the response speed and accuracy of risk control systems through real-time monitoring and early warning mechanisms. Through comprehensive analysis of risk control intentions, credit entities, and risk control rules, the system can construct accurate logical expressions and make effective early warning condition determinations, thereby providing reliable risk control decision support for financial institutions. Practical applications have proved that this solution has a significant effect in improving risk control efficiency and reducing credit risks, and provides a solid technical guarantee for the credit business of financial institutions.

[0082] From the above description, it can be seen that the intelligent generation method of risk warning rules provided in the embodiment of the present application can obtain the user's credit data from a preset credit database, extract indicator features in the credit data, perform indicator relationship identification and entity semantic disambiguation operations on the indicator features, determine the corresponding risk control indicators, determine the corresponding credit entities based on the risk control indicators and the preset entity extraction model, perform dependency syntax analysis operations and relational pattern matching operations on the risk control indicators, and determine the corresponding risk control rules; construct the corresponding risk control rule logical expression according to the risk control intention, credit entity and risk control rules, determine the risk control warning conditions according to the risk control rule logical expression, and generate a post-loan warning signal when the risk control warning conditions are triggered, which can effectively improve the generation efficiency and accuracy of risk control warning rules.

[0083] In one embodiment of the risk warning rule intelligent generation method of the present application, see Figure 2 , and can also include the following:

[0084] Step S201: inputting the risk control indicator into a preset intention recognition algorithm, and determining the corresponding risk control intention according to the prediction result output by the preset intention recognition algorithm and a preset synonym dictionary;

[0085] Step S202: Determine the corresponding credit entity according to the risk control indicator and the preset entity extraction model, perform dependency syntax analysis and relational pattern matching operations on the risk control indicator, and determine the corresponding risk control rule.

[0086] Optionally, in this embodiment, steps S201 and S202 are key links for realizing in-depth analysis and processing of risk control indicators. Step S201 involves inputting risk control indicators into a preset intention recognition algorithm to determine the corresponding risk control intention; step S202 further determines the credit entity based on these risk control indicators and the preset entity extraction model, and determines the corresponding risk control rules through dependency syntax analysis and relational pattern matching operations. The linkage of these two steps realizes the comprehensive assessment and management of user credit risk. The following is a detailed description of steps S201 and S202.

[0087] In step S201, the system first inputs the risk control indicators into the preset intent recognition algorithm. The intent recognition algorithm is usually based on deep learning models, such as long short-term memory networks (LSTM) or transformer models, which can effectively capture semantic information and potential intentions in text. For example, when the system receives a user's credit report, which includes risk control indicators such as "5 loan applications in the past year", "Current credit score is 600", and "2 overdue payments", the intent recognition algorithm will analyze these data and predict that the user's credit intention may be "high-risk lending behavior".

[0088] In order to improve the accuracy of intent recognition, the system also combines it with a preset synonym dictionary for processing. The synonym dictionary can expand and replace the vocabulary in the risk control indicators, allowing the intent recognition algorithm to recognize more diverse expressions. For example, the system replaces "loan" with "borrowing" and "credit score" with "credit score" to generate diverse input data. This data enhancement operation helps the model understand the user's credit behavior more comprehensively and ensure the accuracy of intent recognition.

[0089] In step S202, the system determines the corresponding credit entity based on the identified risk control intention and risk control indicators using a preset entity extraction model. The entity extraction model uses natural language processing technology to identify key credit entities in the text, such as "number of loans", "credit score", "repayment record", etc. For example, the system extracts key information such as "number of loans" of 5, "credit score" of 600, and "number of overdue repayments" of 2 from the user's credit report.

[0090] Next, the system performs dependency syntactic analysis and relational pattern matching operations on these risk control indicators. Dependency syntactic analysis is used to reveal the grammatical relations between features in the text, such as subject-predicate, object-complement, and modification. For example, in the sentence "The user's credit score is 600, and the number of overdue repayments is 2 times", the system identifies "credit score" as the subject, "600" as the complement, "number of overdue repayments" as the subject, and "2 times" as the complement through dependency syntactic analysis. Based on these grammatical relations, the system performs relational pattern matching operations to identify the logical relations and patterns between features, such as parallel relations, conditional relations, etc. These relational patterns help the system accurately understand and analyze the credit behavior of users.

[0091] Through dependency syntactic analysis and relational pattern matching, the system ultimately determines the corresponding risk control rules. For example, the system can formulate risk control rules for "high-risk users" based on the parallel relationship of "credit score is 600 and the number of overdue payments is 2 times". These risk control rules will be used for subsequent credit risk assessment and management to ensure that the system can timely detect and prevent potential credit risks.

[0092] This embodiment achieves comprehensive analysis and processing of risk control indicators through intent recognition algorithms, data enhancement operations, entity extraction, dependency syntax analysis, and relational pattern matching. In the specific implementation process, the system uses advanced deep learning algorithms and natural language processing technologies to ensure the accuracy and efficiency of intent recognition and relational matching. Data enhancement operations increase the robustness and generalization ability of the model by increasing data diversity, thereby improving the reliability of the overall risk control analysis.

[0093] In actual applications, a financial institution conducts in-depth analysis and processing of the user's credit data through the above steps. For example, the system extracts key text features from the user's credit report, inputs the intent recognition algorithm for intent recognition, and predicts the user's credit risk level. Based on the prediction results, the system further determines the user's risk control intention through data enhancement operations such as synonym expansion and synonym replacement. Then, the system uses the entity extraction model to identify the key credit entities in the text, and through dependency syntax analysis and relational pattern matching, it determines the logical relationship and matching pattern between these entities, and finally generates specific risk control rules, such as "high-risk users should be further reviewed" or "low-risk users can be approved for loans."

[0094] This embodiment not only solves the problem of insufficient processing capability of traditional risk control systems for complex text data, but also ensures the accuracy of risk control rules and the reliability of decision-making through multi-level data analysis and relationship matching. This not only improves the risk control efficiency of financial institutions, but also reduces potential credit risks and ensures the stable operation of credit business.

[0095] Through the above two steps, the system can achieve a comprehensive assessment and management of user credit risk. For example, suppose a user has frequently applied for loans in the past year and has a low credit score. The system first determines the user's behavioral intention as "high-risk borrowing behavior" through the intention recognition algorithm, and then expands the data through the synonym dictionary to generate diversified input data to ensure the accuracy of intention recognition.

[0096] Next, the system uses the entity extraction model to identify the user's credit entities such as "number of loans", "credit score", and "repayment record", and uses dependency syntax analysis to identify the grammatical relationship between these entities, such as the parallel relationship between "number of loans" and "number of overdue repayments". Finally, the system determines the logical relationship and matching pattern between these entities through relational pattern matching, and generates corresponding risk control rules. For example, the system determines that "if a user has applied for a loan more than 3 times in the past year and has a credit score below 650 and has a record of overdue repayments, he or she is judged to be a high-risk user", and further reviews the user or takes other risk control measures.

[0097] This comprehensive risk control assessment and management process not only improves the scientificity and effectiveness of risk control decisions, but also enhances the system's adaptability to diverse credit data. By introducing advanced deep learning algorithms and natural language processing technology, the system can accurately identify users' credit intentions and behavior patterns, thereby formulating more precise risk control rules and improving the overall risk control level.

[0098] In practical applications, financial institutions can use this technical solution to timely discover and prevent potential credit risks and ensure the safe and stable operation of credit business. For example, after introducing this system, a financial institution can monitor users' credit behavior in real time and generate early warning signals in time according to risk control rules to remind relevant personnel to take necessary risk control measures. This not only effectively reduces credit risks, but also improves the management efficiency and service quality of credit business.

[0099] In general, steps S201 and S202 achieve comprehensive assessment and management of user credit risk through technical means such as intent recognition, data enhancement, entity extraction, dependency syntax analysis, and relational pattern matching. Through this comprehensive technical solution, the system can accurately identify the user's credit intention and behavior pattern, ensure the accuracy of risk control rules and the reliability of decision-making, and thus provide solid technical support and guarantee for the credit business of financial institutions.

[0100] In one embodiment of the risk warning rule intelligent generation method of the present application, see Figure 3 , and can also include the following:

[0101] Step S301: extracting conditional variables in the risk control rule logic expression, and determining corresponding conditional values ​​according to the extracted conditional variables;

[0102] Step S302: Determine a corresponding condition combination according to the condition variable and the condition value, and make a risk control warning condition determination according to the condition combination.

[0103] Optionally, in this embodiment, steps S301 and S302 are key links, and these two steps work together to ensure accurate determination and early warning of user credit risk, thereby effectively preventing potential credit risks. The following is a detailed description of steps S301 and S302.

[0104] In step S301, the system first extracts conditional variables from the risk control rule logic expression. The risk control rule logic expression is a formal rule representation method used to describe the relationship and conditions between different risk control elements. For example, suppose there is a risk control rule logic expression in the system: "IF(credit score <650AND number of overdue payments>2AND income level <3000) THEN high-risk users". In this expression, "credit score", "number of overdue payments" and "income level" are conditional variables.

[0105] The system extracts these conditional variables one by one by parsing the logical expression. Then, the system needs to obtain the specific conditional value of each conditional variable. For example, a user's credit score is 620, the number of overdue payments is 3, and the income level is 2,500 yuan. The system will correspond these specific values ​​to the conditional variables, thus forming a conditional variable-conditional value pair: credit score-620, number of overdue payments-3, income level-2,500.

[0106] In step S302, the system determines a condition combination based on the extracted condition variables and the corresponding condition values. A condition combination refers to the joint expression of multiple condition variables and their condition values, which is used to determine the risk control warning conditions. For example, based on the above condition variable-condition value pairs, the system combines the following condition combinations: credit score 620, number of overdue payments 3 times, and income level 2,500 yuan.

[0107] Then, the system determines the risk control warning conditions based on these combinations of conditions. Risk control warning condition determination means that the system determines whether the user meets certain warning criteria based on pre-set risk control rules. For example, the system will check whether the user's credit score is less than 650, whether the number of overdue payments exceeds 2 times, and whether the income level is less than 3,000 yuan. If the user's condition combination meets all of the above conditions, the system determines that the user has triggered a high-risk warning condition.

[0108] In this way, the system can monitor the user's credit behavior in real time and generate early warning signals in a timely manner. For example, when a user's credit score is 620, the number of overdue payments is 3 times, and the income level is 2,500 yuan, the system will determine that the user is a high-risk user according to the risk control rules and generate an early warning signal, prompting relevant risk control personnel to take further review or other risk control measures.

[0109] This embodiment achieves accurate monitoring of user credit risk through condition variable extraction and condition combination determination. In the specific implementation process, the system uses advanced data analysis and processing technology to ensure the accurate extraction of condition variables and condition values. Through logical expression analysis, the system can identify the key variables in the risk control rules and obtain the corresponding condition values ​​to form a complete condition combination. In this process, the system needs to have strong data processing and logical analysis capabilities to ensure the accuracy of the condition combination and the timeliness of the risk control warning determination.

[0110] In actual applications, a financial institution uses the above steps to conduct real-time monitoring and early warning of users' credit behaviors. For example, the system extracts key credit data from the user's credit report, inputs the risk control rule logic expression for parsing, and extracts the condition variables and condition values. Then, the system forms a condition combination based on the condition variables and condition values, and determines the early warning conditions based on the preset risk control rules. When a user's condition combination meets the high-risk early warning standard, the system immediately generates an early warning signal and notifies the relevant risk control personnel to take necessary risk control measures.

[0111] This refined risk control assessment and early warning mechanism not only improves the risk control efficiency of financial institutions, but also effectively reduces potential credit risks. By introducing advanced logical expression parsing and conditional combination judgment technology, the system can accurately identify the user's credit risk and generate early warning signals in a timely manner to ensure the safe and stable operation of credit business.

[0112] In general, steps S301 and S302 achieve accurate monitoring and early warning of user credit risk through condition variable extraction, condition value determination and condition combination judgment. This not only solves the problem of insufficient complex data processing capabilities of traditional risk control systems, but also ensures the accuracy and timeliness of risk control early warnings through multi-level data analysis and logical analysis, thereby providing reliable technical support and guarantee for the credit business of financial institutions.

[0113] In one embodiment of the risk warning rule intelligent generation method of the present application, see Figure 4 , and can also include the following:

[0114] Step S401: inputting the text features extracted from the risk control indicators into a preset entity extraction conditional random field model, performing a feature capture operation on the text features according to the feature function layer of the preset entity extraction conditional random field model, and determining corresponding capture features;

[0115] Step S402: input the captured features into the potential energy function layer of the preset entity extraction conditional random field model to determine the corresponding normalized potential energy, input the normalized potential energy into the output layer of the preset entity extraction conditional random field model to determine the corresponding credit entity.

[0116] Optionally, in this embodiment, steps S401 and S402 use complex natural language processing technology to accurately analyze and process risk control indicators. These two steps use the conditional random field model (CRF) to extract entities to ensure that credit entities are accurately captured and identified from risk control indicators. The following is a detailed description of steps S401 and S402.

[0117] In step S401, the system first inputs the text features extracted from the risk control indicators into the preset entity extraction conditional random field model. The conditional random field model is a probabilistic graph model that is commonly used in sequence labeling tasks. It captures and represents the dependencies in sequence data by defining feature functions and potential energy functions. In this step, the system predefines a set of feature functions to identify and capture key information in text features.

[0118] For example, suppose the risk control indicators include "the user applied for loans 5 times in the past year, the credit score is 600, and the number of overdue payments is 2 times". These indicators are processed into text features and input into the conditional random field model. The feature function layer of the model scans these text features and identifies the key elements, such as "number of loan applications", "credit score", "number of overdue payments", etc. The feature capture operation extracts these key elements through the operation of the feature function layer and generates the corresponding feature representation.

[0119] Next, in step S402, the system inputs the captured features into the potential energy function layer of the conditional random field model. The potential energy function layer is used to calculate the potential energy value of each feature and perform normalization. Normalized potential energy represents the relative importance and relevance of the feature in the entire feature space. For example, based on the feature capture results, the system calculates the potential energy value of each feature, such as the potential energy value of "number of loan applications" is 0.8, the potential energy value of "credit score" is 0.6, and the potential energy value of "number of overdue payments" is 0.7.

[0120] After the potential energy function layer completes the calculation, the system inputs the normalized potential energy into the output layer of the conditional random field model. The output layer determines the corresponding credit entity based on the distribution of the normalized potential energy. A credit entity refers to a specific credit factor related to risk control indicators, such as "high-risk borrowing behavior" and "low credit score users". For example, based on the calculation results of the normalized potential energy, the system determines that a user with a "number of loan applications" of 5 times is a high-risk borrowing behavior, and a user with a "credit score" of 600 is a low credit score user.

[0121] In this way, the system uses the feature function layer and potential energy function layer of the conditional random field model to accurately capture text features and accurately identify credit entities. This embodiment ensures comprehensive analysis and processing of risk control indicators through progressive feature processing and potential energy calculation.

[0122] In the specific implementation process, the system uses advanced natural language processing algorithms and probabilistic graphical model technology to ensure the accuracy of feature capture and entity extraction. For example, when processing complex risk control indicator texts, the system can accurately identify key credit factors and generate accurate credit entity representations through potential energy calculation and normalization. In this process, the system needs to have strong data processing and computing capabilities to ensure the efficiency and accuracy of feature capture and entity extraction.

[0123] In actual applications, a financial institution conducts in-depth analysis and processing of users' credit data through the above steps. For example, the system extracts key information from the user's credit report and inputs it into the conditional random field model for feature capture and potential energy calculation. In this way, the system can accurately identify the user's credit behavior pattern and generate the corresponding credit entity representation. When a user's behavior pattern conforms to high-risk lending behavior, the system can generate early warning signals in a timely manner to remind relevant personnel to take necessary risk control measures.

[0124] This embodiment not only solves the problem of insufficient processing capacity of complex text data by traditional risk control systems, but also ensures the accuracy and timeliness of credit entity extraction through multi-level feature analysis and potential energy calculation. This not only improves the risk control efficiency of financial institutions, but also effectively reduces potential credit risks and ensures the stable operation of credit business.

[0125] In general, steps S401 and S402 achieve accurate analysis of risk control indicators and accurate extraction of credit entities through the characteristic function layer and potential energy function layer of the conditional random field model. Through this comprehensive technical solution, the system can accurately capture text features and ensure the efficiency and accuracy of credit entity extraction through potential energy calculation and normalization processing, thereby providing reliable technical support and guarantee for the credit business of financial institutions.

[0126] In one embodiment of the risk warning rule intelligent generation method of the present application, see Figure 5 , and can also include the following:

[0127] Step S501: performing dependency syntactic analysis on the risk control indicator to determine a corresponding word structure, performing a relational pattern matching operation on the preset logical connectives and the word structure to determine a corresponding rule matching graph;

[0128] Step S502: Perform rule classification operations according to the rule matching graph to determine corresponding risk control rules.

[0129] Optionally, in this embodiment, steps S501 and S502 are key steps for in-depth analysis and processing of risk control indicators using natural language processing technology. These two steps ultimately determine the classification and application of risk control rules through dependency syntax analysis and relational pattern matching, ensuring that the system can accurately assess and warn of credit risks. The following is a detailed description of steps S501 and S502.

[0130] In step S501, the system first performs dependency syntactic analysis on the risk control indicators to determine the corresponding word structure. Dependency syntactic analysis is a technology used to parse the internal structure of a sentence. It forms a dependency tree or dependency graph by identifying the dependency relationship between words in a sentence. Dependency syntactic analysis can reveal key grammatical structures such as subject-predicate relationship and modification relationship in a sentence. For example, suppose that the risk control indicators include "the user has applied for a loan 5 times in the past year, the credit score is 600, and the number of overdue repayments is 2 times." The system identifies dependency relationships such as "user-apply for a loan", "credit score-600", "number of overdue repayments-2 times" through dependency syntactic analysis to form the corresponding word structure.

[0131] Next, the system performs relational pattern matching operations based on preset logical connectives and word structures. Logical connectives are words used to connect different grammatical structures, such as "and", "or", "if", etc. In the relational pattern matching process, the system uses these logical connectives to match and combine relevant grammatical relationships in the word structure to form a rule matching graph. For example, based on the dependency relationship between "user applies for a loan 5 times" and "credit score 600", the system matches through the logical connective "and" to generate a rule matching graph: "user applies for a loan 5 times" and "credit score 600" are associated together to form a comprehensive risk control rule.

[0132] In step S502, the system performs rule classification operations based on the rule matching graph. Rule classification refers to classifying the generated rule matching graph into a specific risk control rule type. For example, based on the rule matching graph, the system classifies the combination of "user applies for loan 5 times" and "credit score 600" as a "high-risk lending behavior" rule. In this way, the system can classify and integrate different risk control indicators to form specific risk control rules.

[0133] In the specific implementation process, the system uses advanced natural language processing and graph model technology to ensure the accuracy of dependency syntactic analysis and relational pattern matching. For example, when processing complex risk control indicator texts, the system can accurately identify key dependency relationships and generate accurate rule matching graphs by matching logical connectives. In this process, the system needs to have strong data parsing and processing capabilities to ensure the efficiency and accuracy of dependency syntactic analysis and relational pattern matching.

[0134] In actual applications, a financial institution conducts in-depth analysis and processing of the user's credit data through the above steps. For example, the system extracts key information from the user's credit report, forms a word structure through dependency syntax analysis, and performs relational pattern matching based on logical connectives to generate a rule matching graph. Then, the system classifies the rules based on the rule matching graph and classifies the user's credit behavior pattern into specific risk control rules. When a user's behavior pattern meets the high-risk lending behavior rules, the system can generate early warning signals in a timely manner to remind relevant personnel to take necessary risk control measures.

[0135] This embodiment not only solves the problem of insufficient processing capability of traditional risk control systems for complex text data, but also ensures the accuracy and timeliness of risk control rule classification through multi-level grammatical analysis and relationship matching. This not only improves the risk control efficiency of financial institutions, but also effectively reduces potential credit risks and ensures the stable operation of credit business.

[0136] In general, steps S501 and S502 achieve in-depth analysis of risk control indicators and accurate classification of risk control rules through dependency syntax analysis and relational pattern matching. Through this comprehensive technical solution, the system can accurately identify and match key grammatical relationships in risk control indicators, and through rule classification operations, ensure the efficiency and accuracy of risk control rules, thereby providing reliable technical support and guarantee for the credit business of financial institutions.

[0137] In one embodiment of the risk warning rule intelligent generation method of the present application, see Figure 6 , and can also include the following:

[0138] Step S601: constructing a basic expression according to the risk control intention and the credit entity;

[0139] Step S602: construct a risk control rule logic expression according to the basic expression, the risk control rule and the preset nested logic rule.

[0140] Optionally, in this embodiment, steps S601 and S602 are key links. By constructing basic expressions and risk control rule logic expressions, the system can transform complex risk control intentions and credit entities into operational rules to ensure accurate assessment and early warning of user credit risks. The following is a detailed description of steps S601 and S602.

[0141] In step S601, the system constructs a basic expression based on the risk control intention and the credit entity. The risk control intention refers to the goal or purpose that the system or risk control personnel hopes to achieve, such as "identifying high-risk lending behavior" or "detecting users with low credit scores." Credit entities are specific credit factors extracted from user data, such as "number of loan applications," "credit score," "number of overdue repayments," etc. The system first combines these risk control intentions and credit entities to form a basic logical expression. For example, for the risk control intention of "identifying high-risk lending behavior," the system may combine the two credit entities of "number of loan applications greater than 5 times" and "credit score less than 600" to construct a basic expression: "number of loan applications>5AND credit score<600."

[0142] Next, in step S602, the system constructs a risk control rule logic expression based on the basic expression, risk control rules and preset nested logic rules. Risk control rules are pre-set standards and conditions for evaluating and determining credit risk. For example, a financial institution may set multiple risk control rules, such as "high-frequency borrowing users", "low credit score users", "high overdue risk users", etc. Nested logic rules are logic rules used to combine and nest multiple basic expressions, such as "AND", "OR", "IF-THEN", etc.

[0143] The system first matches basic expressions with risk control rules to determine which basic expressions meet specific risk control rules. For example, the basic expression "Number of loan applications > 5 AND credit score < 600" meets the risk control rule of "high-frequency borrowing users". Then, the system combines and nests multiple basic expressions according to the preset nested logic rules to form a complex risk control rule logic expression. For example, the system can combine "Number of loan applications > 5 AND credit score < 600" with "Number of overdue repayments > 3" to form a nested logic expression: "(Number of loan applications > 5 AND credit score < 600) OR Number of overdue repayments > 3" to identify higher-risk lending behaviors.

[0144] This embodiment realizes the accurate expression and evaluation of complex risk control intentions and credit entities through the construction of basic expressions and nested logical rules. In the specific implementation process, the system adopts advanced logic processing and expression parsing technology to ensure the accuracy of basic expressions and risk control rule logical expressions. For example, when processing complex risk control intentions and credit entities, the system can accurately identify and combine related logical conditions to generate accurate risk control rule logical expressions. In this process, the system needs to have powerful logic processing and data parsing capabilities to ensure the efficiency and accuracy of risk control rule logical expressions.

[0145] In actual applications, a financial institution conducts in-depth analysis and processing of users' credit data through the above steps. For example, the system first constructs basic expressions based on risk control intentions and credit entities, and matches these expressions with preset risk control rules. Then, the system combines and nests multiple basic expressions according to nested logic rules to generate complex risk control rule logic expressions. When a user's behavior pattern meets these logical expressions, the system can generate early warning signals in a timely manner to remind relevant personnel to take necessary risk control measures.

[0146] This embodiment not only solves the problem of insufficient processing capacity of complex logic in traditional risk control systems, but also ensures the accuracy and timeliness of risk control rule logic expressions through multi-level logic expression and nesting. This not only improves the risk control efficiency of financial institutions, but also effectively reduces potential credit risks and ensures the stable operation of credit business.

[0147] In general, steps S601 and S602 achieve accurate expression and evaluation of risk control intentions and credit entities by constructing basic expressions and risk control rule logic expressions. Through this comprehensive technical solution, the system can accurately identify and combine risk control intentions and credit entities, and through nested logic rules, ensure the efficiency and accuracy of risk control rule logic expressions, thereby providing reliable technical support and guarantee for the credit business of financial institutions.

[0148] In order to effectively improve the generation efficiency and accuracy of risk control warning rules, the present application provides an embodiment of a risk warning rule intelligent generation device for realizing all or part of the content of the risk warning rule intelligent generation method, see Figure 7 The risk warning rule intelligent generation device specifically includes the following contents:

[0149] The data preprocessing module 10 is used to obtain the user's credit data from a preset credit database, extract the indicator features in the credit data, perform indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determine the corresponding risk control indicators;

[0150] A rule generation module 20 is used to input the text features extracted from the risk control indicator into a preset intention recognition long short-term memory network, determine the corresponding prediction result, perform data enhancement operation according to the prediction result and a preset synonym dictionary, determine the corresponding risk control intention, determine the corresponding credit entity according to the risk control indicator and a preset entity extraction model, perform dependency syntax analysis operation and relational pattern matching operation on the risk control indicator, and determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization;

[0151] The risk control module 30 is used to construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

[0152] From the above description, it can be seen that the intelligent generation device of risk warning rules provided in the embodiment of the present application can obtain the user's credit data from a preset credit database, extract indicator features in the credit data, perform indicator relationship identification and entity semantic disambiguation operations on the indicator features, determine the corresponding risk control indicators, determine the corresponding credit entities according to the risk control indicators and the preset entity extraction model, perform dependency syntax analysis operations and relational pattern matching operations on the risk control indicators, and determine the corresponding risk control rules; construct the corresponding risk control rule logical expression according to the risk control intention, credit entity and risk control rules, determine the risk control warning conditions according to the risk control rule logical expression, and generate a post-loan warning signal when the risk control warning conditions are triggered, which can effectively improve the generation efficiency and accuracy of risk control warning rules.

[0153] From the hardware level, in order to effectively improve the generation efficiency and accuracy of risk control warning rules, the present application provides an embodiment of an electronic device for implementing all or part of the contents of the risk warning rule intelligent generation method, and the electronic device specifically includes the following contents:

[0154] Processor, memory, communication interface and bus; wherein the processor, memory and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between the risk warning rule intelligent generation device and related devices such as core business systems, user terminals and related databases; the logic controller can be a desktop computer, a tablet computer and a mobile terminal, etc., and the present embodiment is not limited thereto. In the present embodiment, the logic controller can be implemented with reference to the embodiment of the risk warning rule intelligent generation method and the embodiment of the risk warning rule intelligent generation device in the embodiment, and the contents thereof are incorporated herein, and the repeated parts are not repeated.

[0155] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0156] In practical applications, part of the intelligent generation method of risk warning rules can be executed on the electronic device side as described above, or all operations can be completed in the client device. The selection can be made based on the processing capability of the client device and the limitations of the user's usage scenario. This application does not limit this. If all operations are completed in the client device, the client device may also include a processor.

[0157] The client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and other implementation scenarios may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0158] Figure 8 FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Figure 8 As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Figure 8 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0159] In one embodiment, the function of the intelligent generation method of risk warning rules can be integrated into the central processing unit 9100. The central processing unit 9100 can be configured to perform the following control:

[0160] Step S101: obtaining the user's credit data from a preset credit database, extracting indicator features from the credit data, performing indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determining corresponding risk control indicators;

[0161] Step S102: inputting the text features extracted from the risk control indicator into a preset intention recognition long short-term memory network to determine the corresponding prediction result, performing data enhancement operation according to the prediction result and a preset synonym dictionary to determine the corresponding risk control intention, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing dependency syntax analysis operation and relational pattern matching operation on the risk control indicator to determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization;

[0162] Step S103: construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

[0163] From the above description, it can be seen that the electronic device provided in the embodiment of the present application obtains the user's credit data from a preset credit database, extracts indicator features in the credit data, performs indicator relationship recognition and entity semantic disambiguation operations on the indicator features, determines the corresponding risk control indicators, determines the corresponding credit entities according to the risk control indicators and the preset entity extraction model, performs dependency syntax analysis operations and relational pattern matching operations on the risk control indicators, and determines the corresponding risk control rules; constructs corresponding risk control rule logical expressions according to risk control intentions, credit entities and risk control rules, determines risk control warning conditions according to the risk control rule logical expressions, and generates a post-loan warning signal when the risk control warning conditions are triggered, which can effectively improve the generation efficiency and accuracy of risk control warning rules.

[0164] In another embodiment, the risk warning rule intelligent generation device can be configured separately from the central processing unit 9100. For example, the risk warning rule intelligent generation device can be configured as a chip connected to the central processing unit 9100, and the function of the risk warning rule intelligent generation method is realized through the control of the central processing unit.

[0165] like Figure 8 As shown, the electronic device 9600 may also include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Figure 8 In addition, the electronic device 9600 may also include Figure 8 For components not shown, reference may be made to the prior art.

[0166] like Figure 8As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0167] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0168] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0169] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0170] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0171] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0172] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module, etc. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0173] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps of the method for intelligently generating risk warning rules in the above-mentioned embodiments, where the execution subject is a server or a client. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, all the steps of the method for intelligently generating risk warning rules in the above-mentioned embodiments are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0174] Step S101: obtaining the user's credit data from a preset credit database, extracting indicator features from the credit data, performing indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determining corresponding risk control indicators;

[0175] Step S102: inputting the text features extracted from the risk control indicator into a preset intention recognition long short-term memory network to determine the corresponding prediction result, performing data enhancement operation according to the prediction result and a preset synonym dictionary to determine the corresponding risk control intention, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing dependency syntax analysis operation and relational pattern matching operation on the risk control indicator to determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization;

[0176] Step S103: construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

[0177] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application obtains the user's credit data from a preset credit database, extracts indicator features in the credit data, performs indicator relationship identification and entity semantic disambiguation operations on the indicator features, determines the corresponding risk control indicators, determines the corresponding credit entities according to the risk control indicators and the preset entity extraction model, performs dependency syntax analysis operations and relational pattern matching operations on the risk control indicators, and determines the corresponding risk control rules; constructs the corresponding risk control rule logical expressions according to the risk control intentions, credit entities and risk control rules, determines the risk control warning conditions according to the risk control rule logical expressions, and generates a post-loan warning signal when the risk control warning conditions are triggered, which can effectively improve the generation efficiency and accuracy of risk control warning rules.

[0178] The embodiments of the present application also provide a computer program product capable of implementing all the steps of the risk warning rule intelligent generation method in the above-mentioned embodiment, where the execution subject is a server or a client. When the computer program / instruction is executed by a processor, the steps of the risk warning rule intelligent generation method are implemented. For example, the computer program / instruction implements the following steps:

[0179] Step S101: obtaining the user's credit data from a preset credit database, extracting indicator features from the credit data, performing indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determining corresponding risk control indicators;

[0180] Step S102: inputting the text features extracted from the risk control indicator into a preset intention recognition long short-term memory network to determine the corresponding prediction result, performing data enhancement operation according to the prediction result and a preset synonym dictionary to determine the corresponding risk control intention, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing dependency syntax analysis operation and relational pattern matching operation on the risk control indicator to determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization;

[0181] Step S103: construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

[0182] From the above description, it can be seen that the computer program product provided in the embodiment of the present application obtains the user's credit data from a preset credit database, extracts indicator features in the credit data, performs indicator relationship identification and entity semantic disambiguation operations on the indicator features, determines the corresponding risk control indicators, determines the corresponding credit entities based on the risk control indicators and the preset entity extraction model, performs dependency syntax analysis operations and relational pattern matching operations on the risk control indicators, and determines the corresponding risk control rules; constructs the corresponding risk control rule logical expressions based on the risk control intentions, credit entities and risk control rules, determines the risk control warning conditions based on the risk control rule logical expressions, and generates a post-loan warning signal when the risk control warning conditions are triggered, which can effectively improve the generation efficiency and accuracy of risk control warning rules.

[0183] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0184] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0185] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0186] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0187] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for intelligently generating risk warning rules, characterized in that: The method comprises: Acquire the user's credit data from a preset credit database, extract indicator features from the credit data, perform indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determine corresponding risk control indicators; Inputting the text features extracted from the risk control indicator into a preset intent recognition long short-term memory network to determine the corresponding prediction result, performing data enhancement operation according to the prediction result and a preset synonym dictionary to determine the corresponding risk control intent, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing dependency syntax analysis operation and relational pattern matching operation on the risk control indicator to determine the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement and sentence reorganization; A corresponding risk control rule logic expression is constructed according to the risk control intention, the credit entity and the risk control rule, a risk control warning condition is determined according to the risk control rule logic expression, and a post-loan warning signal is generated when the risk control warning condition is triggered.

2. The method for intelligently generating risk warning rules according to claim 1 is characterized in that: Determining the corresponding risk control intention, credit entity and risk control rules according to the risk control indicator includes: Inputting the risk control indicator into a preset intention recognition algorithm, and determining the corresponding risk control intention according to the prediction result output by the preset intention recognition algorithm and a preset synonym dictionary; The corresponding credit entity is determined according to the risk control indicator and the preset entity extraction model, and the risk control indicator is subjected to dependency syntax analysis and relational pattern matching operations to determine the corresponding risk control rules.

3. The method for intelligently generating risk warning rules according to claim 1 is characterized in that: The determining of the risk control warning condition according to the risk control rule logical expression includes: Extracting the conditional variables in the risk control rule logic expression, and determining the corresponding conditional values ​​according to the extracted conditional variables; A corresponding condition combination is determined according to the condition variable and the condition value, and a risk control warning condition is determined according to the condition combination.

4. The method for intelligently generating risk warning rules according to claim 2 is characterized in that: The determining the corresponding credit entity according to the risk control indicator and the preset entity extraction model includes: Inputting the text features extracted from the risk control indicators into a preset entity extraction conditional random field model, performing a feature capture operation on the text features according to a feature function layer of the preset entity extraction conditional random field model, and determining corresponding capture features; The captured features are input into the potential energy function layer of the preset entity extraction conditional random field model to determine the corresponding normalized potential energy, and the normalized potential energy is input into the output layer of the preset entity extraction conditional random field model to determine the corresponding credit entity.

5. The method for intelligently generating risk warning rules according to claim 2 is characterized in that: The performing dependency syntax analysis and relational pattern matching operations on the risk control indicators to determine corresponding risk control rules includes: Performing dependency syntactic analysis on the risk control indicator to determine the corresponding word structure, performing a relational pattern matching operation based on preset logical connectives and the word structure to determine the corresponding rule matching graph; A rule classification operation is performed according to the rule matching graph to determine the corresponding risk control rules.

6. The method for intelligently generating risk warning rules according to claim 1 is characterized in that: The step of constructing a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule includes: Constructing a basic expression according to the risk control intention and the credit entity; A risk control rule logical expression is constructed according to the basic expression, the risk control rule and the preset nested logical rule.

7. A risk warning rule intelligent generation device, characterized in that: The device comprises: A data preprocessing module is used to obtain the user's credit data from a preset credit database, extract indicator features in the credit data, perform indicator relationship recognition and entity semantic disambiguation operations on the indicator features, and determine the corresponding risk control indicators; A rule generation module, used for inputting the text features extracted from the risk control indicator into a preset intent recognition long short-term memory network, determining the corresponding prediction result, performing a data enhancement operation according to the prediction result and a preset synonym dictionary, determining the corresponding risk control intent, determining the corresponding credit entity according to the risk control indicator and a preset entity extraction model, performing a dependency syntax analysis operation and a relational pattern matching operation on the risk control indicator, and determining the corresponding risk control rule, wherein the data enhancement operation includes at least one of synonym expansion, synonym replacement, and sentence reorganization; The risk control module is used to construct a corresponding risk control rule logic expression according to the risk control intention, the credit entity and the risk control rule, determine the risk control warning condition according to the risk control rule logic expression, and generate a post-loan warning signal when the risk control warning condition is triggered.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the risk warning rule intelligent generation method described in any one of claims 1 to 6 are implemented.

9. 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 risk warning rule intelligent generation method described in any one of claims 1 to 6 are implemented.

Images