Risk assessment method and device and electronic equipment

Abstract

The embodiment of the invention provides a risk assessment method and device and electronic equipment, which are applied to the field of data analysis, and the method can identify key fields from user information to generate a key field set after the user information is acquired. And querying a target assessment rule in a rule base based on the key field set, generating a risk assessment result according to the target assessment rule and the key field set, and outputting the risk assessment result, the risk assessment result comprising a risk type and a risk level. Wherein the rule base comprises a basic evaluation rule set based on the risk level and an extended evaluation rule generated through a machine learning algorithm based on the basic evaluation rule, and rule parameters in the evaluation rule can be dynamically adjusted according to a result modification instruction of the risk evaluation result. The method can automatically analyze the user portrait based on the user information and perform risk assessment, and can improve the efficiency and accuracy of risk assessment.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and in particular to a risk assessment method, apparatus, and electronic device. Background Technology

[0002] Risk assessment is a data analysis and processing method based on user information. It aims to determine the risk level of a user by analyzing the characteristics of that user information. Risk assessment can be applied in technical fields such as insurance, helping policyholders accurately identify various potential risks they may face by conducting a comprehensive risk assessment of user information.

[0003] For example, risk assessment can identify whether a user has potential health risks, accidents, or property losses, thereby recommending a suitable insurance plan to mitigate financial losses caused by insufficient or excessive insurance coverage. This ensures that the selected insurance products are highly compatible with the customer's actual needs, ultimately achieving personalized and adequate risk protection.

[0004] Risk assessment can be conducted by assessors or assessment agencies, which involves analyzing user information such as occupation and family circumstances to determine the user's risks and needs. For example, in the insurance industry, insurance agents can extract factors such as age, occupation, income, and health status from user information and assess the risks faced by the user based on these factors. However, because this risk assessment method relies heavily on the experience of assessors or assessment agencies and is performed manually, it can lead to low efficiency and poor accuracy. Summary of the Invention

[0005] In view of this, embodiments of this application provide a risk assessment method, apparatus, and electronic device to solve the problems of low efficiency and poor accuracy in risk assessment.

[0006] According to one aspect of this application, a risk assessment method is provided, the method comprising:

[0007] Obtain user information, wherein the user information includes at least one of first information and second information; the first information is information entered through an information input interface; and the second information is an uploaded information file.

[0008] Key fields are identified from the user information to generate a set of key fields, the set of key fields including at least one of the key fields; the key fields are used to characterize text information that is related to the risk level;

[0009] Based on the set of key fields, the target assessment rules are queried in the rule base, which includes a first type of rules and a second type of rules. The first type of rules are basic assessment rules based on risk level settings. The second type of rules are extended assessment rules generated by machine learning algorithms based on the first type of rules. The rule parameters in the second type of rules can be dynamically adjusted according to the modification instructions based on the risk assessment results.

[0010] The risk assessment results are generated and output based on the target assessment rules and the set of key fields. The risk assessment results include risk type and risk level.

[0011] According to another aspect of this application, a risk assessment apparatus is provided, the apparatus comprising:

[0012] The information acquisition module is used to acquire user information, which includes at least one of first information and second information; the first information is information input through an information input interface; the second information is an uploaded information file.

[0013] A field recognition module is used to identify key fields from the user information to generate a key field set, the key field set including at least one of the key fields; the key fields are used to characterize text information that is related to the risk level;

[0014] The rule query module is used to query target assessment rules in the rule base based on the set of key fields. The rule base includes a first type of rules and a second type of rules. The first type of rules are basic assessment rules based on risk level settings. The second type of rules are extended assessment rules generated by machine learning algorithms based on the first type of rules. The rule parameters in the second type of rules can be dynamically adjusted according to the modification instructions based on the risk assessment results.

[0015] The risk assessment module is used to generate risk assessment results based on the target assessment rules and the set of key fields, and to output the risk assessment results, which include risk type and risk level.

[0016] According to another aspect of this application, an electronic device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described risk assessment method.

[0017] According to another aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described risk assessment method.

[0018] By employing the above technical solutions, embodiments of this application provide a risk assessment method, apparatus, and electronic device. The method, after acquiring user information, can identify key fields from the user information to generate a key field set. Then, based on the key field set, it queries a target assessment rule in a rule base, and generates and outputs a risk assessment result based on the target assessment rule and the key field set. The risk assessment result includes risk type and risk level. The rule base includes basic assessment rules set based on risk level and extended assessment rules generated by machine learning algorithms based on the basic assessment rules. Furthermore, the rule parameters in the assessment rules can be dynamically adjusted according to modification instructions based on the risk assessment result. This method can automatically perform risk assessment based on user information and assessment rules, accurately analyze user profiles, and assess risk types and risk levels, improving the efficiency and accuracy of risk assessment.

[0019] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0021] Figure 1 This is a schematic diagram of the risk assessment application environment provided for the embodiments of this application;

[0022] Figure 2 This is a schematic diagram of the risk assessment method provided in the embodiments of this application;

[0023] Figure 3 A flowchart illustrating the risk assessment process provided in this application embodiment;

[0024] Figure 4 This is a schematic diagram of the result display interface provided in the embodiments of this application;

[0025] Figure 5 This is a schematic diagram of the modification evaluation result process provided in the embodiments of this application;

[0026] Figure 6 This is a schematic diagram of the dynamic adjustment evaluation rule process provided in the embodiments of this application;

[0027] Figure 7 This is a schematic diagram of the risk dimension generation engine processing flow provided in the embodiments of this application;

[0028] Figure 8 This is a schematic diagram of the risk assessment device provided in an embodiment of this application. Detailed Implementation

[0029] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0030] In this embodiment, risk assessment is a data analysis and processing method based on user information. It aims to determine the risk level of a user by performing feature analysis on the user information. Risk assessment can be applied to technical fields such as insurance, helping to identify various potential risks that users may face through a comprehensive risk assessment of user information.

[0031] For example, risk assessment can identify whether a user has potential health risks, accidents, or property losses, thereby recommending a suitable insurance plan to mitigate financial losses caused by insufficient or excessive insurance coverage. This ensures that the selected insurance products are highly compatible with the customer's actual needs, ultimately achieving personalized and adequate risk protection.

[0032] In some embodiments, risk assessment can be conducted by assessors or assessment agencies, that is, by analyzing conditions such as occupation and family situation in user information to determine the user's risks and needs. For example, in the insurance industry, insurance agents can read factors such as age, occupation, income, and health status from user information, and assess the risks faced by the user based on these factors, thus achieving risk assessment.

[0033] However, because this risk assessment method relies heavily on the experience of assessors or assessment agencies for manual completion, it leads to low efficiency and poor accuracy. To address the problems of low efficiency and poor accuracy in risk assessment, some embodiments of this application provide a risk assessment method that can be applied to situations such as... Figure 1 In this technical scenario, the client communicates with the server via a network. The server can respond to assessment instructions sent by the client to perform risk assessments. The client can be, but is not limited to, various electronic devices, such as computers, servers, mobile terminals, smart wearable devices, industrial control machines, etc. The server can be a standalone server or a server cluster consisting of multiple servers.

[0034] For ease of description, electronic devices or risk assessment systems are used as the execution subject of the method in this application embodiment. It should be understood that the method can also be applied to other types of execution subjects, which will not be shown one by one in this application embodiment. Figure 2 As shown, the method includes:

[0035] S101. Obtain user information.

[0036] When conducting risk assessments for electronic devices, it is necessary to first obtain user information. User information may include personal information, family information, and business and asset information that can characterize the target user. For example, the personal information of a target customer includes, but is not limited to, at least one of the following: name, gender, mobile phone number, age, date of birth, occupation, annual income, health status (healthy, sub-healthy, suffering from diseases, etc.), past medical history, foreign nationality or immigration plans, marital status (single, married, divorced or widowed), relationship status (unknown, stable, unstable), second or subsequent marriages, family breadwinner, priority concerns (children's education, personal retirement, wealth transfer, wealth appreciation or preservation, tax planning, etc.), social security status, years of social security contributions, current monthly salary, expected monthly pension, frequently used means of transportation, and expected level of supplementary insurance (basic 200,000, mid-range 300,000, upper-middle 500,000, high-end 1,000,000), etc.

[0037] Family information includes, but is not limited to, at least one of the following: personal information of the spouse (age, date of birth, occupation, annual income, health status, past medical history, foreign nationality or immigration plans, family breadwinner, etc.), personal information of children (age, date of birth, gender, whether there are children, children's age, whether they are adopted or illegitimate children, whether they have foreign nationality, marital status, marital status, whether there are concerns about their financial situation, whether there are inheritance plans, direct or intergenerational inheritance, etc.), and personal information of parents (age, date of birth, gender, whether they have foreign nationality, marital status, etc.).

[0038] Enterprise and asset information includes, but is not limited to, at least one of the following: asset type, asset status, asset management information, insurance allocation information, and enterprise information. Among them, asset type may include at least one of the following: investments in equity funds and stocks and their percentage of total assets; investments in fixed-income wealth management trusts and their percentage of total assets; deposits; investments in government bonds and their percentage of total assets; the number of investment properties and their percentage of total assets; and whether there is equity in an enterprise.

[0039] Asset information may include at least one of the following: whether there are liabilities (loans) and guarantees, the percentage of liabilities (loans) and guarantees to total assets, the percentage of overseas assets to total assets, the percentage of assets held on behalf of others to total assets, and whether there are plans to transfer assets to children.

[0040] Asset management information may include annual family income, annual family expenditure, total assets, and potential large future expenditures. Information regarding children's education expenditures includes, but is not limited to, at least one of the following: annual expenditure on university education, annual growth rate of education costs, plans for further education, annual expenditure on postgraduate education, total amount of education savings, and expected annual return on education savings. If studying abroad is chosen, this information will also include: age of study abroad, duration of study abroad, and annual expenditures during the study abroad period.

[0041] Insurance coverage information may include at least one of the following types: health insurance (reimbursement / fixed benefit), annuity (monthly payout), life insurance coverage, accident insurance coverage, and critical illness insurance coverage. Company information may include company type, company prospects (development stage / mature stage / decline stage), and tax compliance status.

[0042] Depending on the source of the information, the user information may include at least one of a first information and a second information. The first information is information entered through an information input interface. For example, user information can be sent to the server via a client. That is, the target user can interactively control the client to run the Pocket e customer module to display the information input interface. Then, from the information input interface, the user can select "Create New Customer Analysis" (mainly for new customers), causing the client to pop up input controls for information entry. The target customer can sequentially enter personal information, family information, and business and asset information through the information input interface.

[0043] The second piece of information is the uploaded information file. For example, after the target user controls the client to display the information entry interface and pops up the input control, they can import existing customer information through the file upload option in the input control. The client can directly send the uploaded user information file to the server, or perform information recognition locally to extract the user information.

[0044] Since the content in the information file may differ from the required user information when inputting user information by uploading an information file, the customer information can be supplemented and improved through information entry after importing existing customer information.

[0045] In some embodiments, after obtaining user information, the client can also preview the user information through a preview interface. For example, after a target user imports existing customer information and completes the customer information through the information entry interface, the client can display an information confirmation interface. This interface allows previewing the imported existing customer information and the completed customer information. Furthermore, the information confirmation interface may include confirmation and modification controls. When the user clicks the confirmation control, the client can send the user information to the server, enabling the electronic device acting as the server to obtain the user information.

[0046] S102. Identify key fields from user information to generate a set of key fields.

[0047] After obtaining user information, the risk assessment system can perform text recognition on the user information, that is, identify key fields from the user information to generate a set of key fields. The set of key fields includes at least one of the key fields; the key fields are used to characterize textual information that is related to the risk level.

[0048] In some embodiments, to generate a key field set, when an electronic device identifies key fields from the user information to generate the key field set, it can first extract non-text data from the user information and, based on the data type of the non-text data, invoke a text conversion tool. Then, the text conversion tool is used to convert the non-text data into text data, and based on the text data, text to be identified is generated. Therefore, the text to be identified includes the text data and the text content in the user information. Then, referring to a preset thesaurus, the key fields are extracted from the text to be identified, that is, the field names and field values ​​corresponding to the key fields are extracted. Thus, the key fields are recorded based on a preset input template to generate the key field set. The preset input template is a template pre-built based on the layout of the input controls in the information input interface.

[0049] For example, when a user imports existing customer information through the information entry interface, they can upload non-text data such as images and audio files using a file upload control. When the electronic device detects that the uploaded user information includes non-text data, it can read the data type of the non-text data. That is, for image data, its data type can be determined by the file format name. For example, if the file extension is bmp, jpg, png, tif, gif, etc., the data type of the non-text data can be determined as image data. Similarly, for audio data, its data type can also be determined by the file format name. For example, if the file extension is CD, WAVE, AIFF, MPEG, MP3, MIDI, WMA, etc., the data type of the non-text data can be determined as audio data.

[0050] After determining the data type of the non-text data, the electronic device can call the corresponding text conversion tool based on the data type. For image data, an Optical Character Recognition (OCR) tool can be used to recognize text content from the image data. For audio data, speech-to-text tools such as Buzz, Sonix, and Otter.ai can be used to convert the audio data into text data.

[0051] After converting non-text data into text data using a text conversion tool, the converted text data can be combined with the original text content in the user information to generate the text to be recognized based on the text data and user information. Therefore, the text to be recognized includes the text data and the text content in the user information.

[0052] After generating the text to be recognized, the electronic device can refer to a preset thesaurus to extract key fields from the text, that is, extract the field names and field values ​​corresponding to the key fields. For example, when extracting personal information of a target customer, keywords such as name, gender, mobile phone number, age, date of birth, occupation, annual income, and health status can be identified in the text to be recognized based on keyword matching, thus obtaining the field names. Then, based on the value range associated with the keywords, the field values ​​are extracted. For example, the last 11 digits of the mobile phone number field name are usually the mobile phone number, so the field value of the mobile phone number can be extracted from this range.

[0053] Multiple field names can be used to extract corresponding field values, thus obtaining multiple key fields. After obtaining multiple key fields, they can be combined to generate the key field set. To facilitate subsequent risk assessment and analysis, the key fields can be recorded based on a preset input template when combining multiple key fields.

[0054] The preset input template is a template pre-built based on the layout of the input controls in the information input interface. For example, if the text input boxes in the information input interface are arranged in the order of name, gender, mobile phone number, age, date of birth, occupation, annual income, and health status, a preset input template with the same field input order can be built based on this layout order. Thus, based on the layout order in the preset input template, key fields such as name, gender, mobile phone number, age, date of birth, occupation, annual income, and health status are recorded sequentially to obtain a set of key fields.

[0055] In some embodiments, the risk assessment system can also identify key fields based on semantic analysis. Therefore, when the electronic device performs the task of identifying key fields from the user information to generate a set of key fields, it can first call the semantic recognition model and then input the user information into the semantic recognition model so as to identify the semantic information of the user information through the semantic recognition model.

[0056] Semantic recognition models, in particular, can understand the meaning of text by analyzing its vocabulary, grammatical structure, and contextual information. They map natural language text to a semantic space, enabling semantic understanding, classification, and matching. Semantic recognition models can be rule-based, statistical, or deep learning-based.

[0057] Taking a deep learning-based semantic recognition model as an example, this model is a neural network model trained on sample business data. Specifically, it utilizes deep neural networks such as recurrent neural networks, long short-term memory networks, gated recurrent units, and the Transformer architecture to learn the semantic representation of text. The semantic recognition model can automatically extract features from the text and construct complex semantic mappings through the combination of multiple neural networks, enabling it to classify text into different categories based on its semantic content. For instance, when performing sentiment analysis on users, it can identify the text expressing emotional content within the business text and determine whether the user's emotional tendency is positive, negative, or neutral based on the semantic recognition results.

[0058] After recognizing the semantic information of user information using a semantic recognition model, the electronic device can perform text segmentation on the user information based on the semantic information to extract the key fields from the user information, and then combine multiple key fields into a key field set. For example, by performing semantic analysis on user consultation text and business text, the electronic device can determine semantic information from the semantic analysis results that represents whether the user is the breadwinner of the family, their priority concerns, social security status, years of social security contributions, current monthly salary, expected monthly pension, frequently used means of transportation, and expected level of social security coverage. Based on the semantic information, text segmentation and extraction are performed on consultation text and business text to obtain the text content corresponding to the semantic information, i.e., the key fields. After obtaining the key fields, multiple key fields are combined in the manner provided in the above embodiments to generate a key field set.

[0059] S103. Query the target evaluation rules in the rule base based on the key field set.

[0060] After generating a set of key fields, electronic devices can perform risk assessments based on this set. During the risk assessment, the system can comprehensively evaluate potential risks according to assessment rules based on mapping relationships. These assessment rules can be pre-built according to the classification methods of business domains. Each assessment rule can include multiple assessment intervals, and each interval can be associated with a corresponding assessment result. Multiple assessment rules can be categorized and stored according to factors such as type, purpose, and applicable key fields to form a rule base.

[0061] In some embodiments, the rule base includes a first type of rules and a second type of rules. The first type of rules are basic assessment rules based on risk level settings; the second type of rules are extended assessment rules generated by machine learning algorithms based on the first type of rules, and the rule parameters in the second type of rules can be dynamically adjusted according to modification instructions based on the risk assessment results.

[0062] For the first type of rule, rule data can be constructed based on a pre-defined "information-risk" mapping relationship. For example, "information-risk" mapping relationship data can include the correspondence between personal information, family information, business and asset information, and risk levels across various dimensions. Different mapping relationships can be included for different user information. For instance, "information-risk" mapping relationship data could include: "information-family risk" mapping relationship data, "information-marital risk" mapping relationship data, "information-wealth inheritance risk" mapping relationship data, "information-tax risk" mapping relationship data, and "information-nominee shareholding risk" mapping relationship data, etc.

[0063] The "information-risk" mapping data can be set according to actual needs, and this embodiment does not limit it. For example, the "information-nominee shareholding risk" mapping data can be as follows: if the percentage of the nominee shareholding assets is less than 5% of the total assets and the tax is compliant, the nominee shareholding risk assessment result is determined to be low risk; if the percentage of the nominee shareholding assets is between 5% and 10% of the total assets and the tax is compliant, the nominee shareholding risk assessment result is determined to be medium risk; if the percentage of the nominee shareholding assets is greater than 10% of the total assets and the tax is non-compliant, the nominee shareholding risk assessment result is determined to be high risk.

[0064] By pre-setting the first type of rule, namely the "information-risk" mapping relationship data, automated risk assessment can be achieved, effectively solving the inefficiency problem caused by relying on the experience of agents in existing technologies; comprehensive analysis based on multi-dimensional information such as personal, family and corporate assets can significantly improve the accuracy of risk assessment; at the same time, it covers a variety of risk types such as family, marriage, property inheritance, taxation and nominee shareholding, making the assessment results more comprehensive and systematic.

[0065] For the second type of rules, the risk assessment system can run an Artificial Intelligence (AI) model. The AI ​​model can build upon the first type of rules (i.e., basic assessment rules), combining business data learning and user feedback to further expand the rules and generate the second type of rules (i.e., extended assessment rules). For example, the AI ​​model can collect data on external factors affecting risk in real time, such as market information and policy information. By analyzing this data, it can determine the current market environment and policy factors, and adjust the pre-set basic assessment rules accordingly to generate extended assessment rules. This allows the basic assessment rules to adapt to complex and ever-changing market environments and other external factors.

[0066] Since the key field set can include multiple key fields, and different key fields represent different contents, the corresponding assessment rules used in risk assessment will also differ. Therefore, after generating the key field set, the electronic device can perform rule matching based on each key field in the key field set, that is, query the target assessment rule in the rule base according to the key field.

[0067] In some embodiments, when querying target assessment rules, the risk assessment system can determine the target assessment rules based on the risk assessment intent. That is, when an electronic device queries the rule base for target assessment rules based on the set of key fields, it can first obtain the risk assessment intent. The risk assessment intent characterizes the application purpose of the current risk assessment result; for example, when the risk assessment is used to recommend insurance products to a user, the risk assessment intent can be product recommendation.

[0068] The risk assessment intent can be specified by the user; that is, the user or customer service personnel can specify the risk assessment intent through the information entry interface when entering user information. The risk assessment system then directly obtains the risk assessment intent from the entered user information after the user submits the data. Alternatively, the risk assessment intent can be indirectly obtained by detecting the method through which user information was entered. After obtaining user information, the risk assessment system can detect the method through which the information was obtained. When the method is an application (or a specific interactive interface), the system can determine the application type and thus determine the risk assessment intent based on the application type. For example, when a user enters user information through an insurance purchase service app, the risk assessment system can determine the risk assessment intent as product recommendation based on the insurance purchase service app.

[0069] In some embodiments, the risk assessment intent is also obtained by performing intent recognition on the user information. That is, after obtaining user information, the risk assessment system can invoke an intent recognition model. This intent recognition model can be a natural language processing model trained based on intent classification sample data. The intent recognition model can determine the risk assessment intent by analyzing the semantics and contextual information in the user information. For example, when the user-input information may include consultation text content for inquiring about insurance products, the intent recognition model can perform semantic recognition on the consultation text content and, combined with the contextual information of the consultation text, determine that the risk assessment intent is product recommendation.

[0070] After obtaining the risk assessment intent, the electronic device then calls the target rule base according to the risk assessment intent, calculates the risk scanning dimensions based on the key field set, and extracts the target assessment rules associated with the risk scanning dimensions from the target rule base. For example, when the risk assessment intent is to recommend insurance products, the target rule base can be called according to the risk assessment intent; that is, the rules in the target rule base are rules set for product recommendations. Furthermore, the assessment rules in the target rule base can be categorized according to the risk scanning dimensions.

[0071] Risk scanning dimensions can include risks associated with the family's breadwinner and family cash flow risks. The impact of the family's breadwinner extends beyond cash flow to potentially cause inheritance disputes; the solution is leveraged life insurance combined with an insurance proceeds trust plan. Cash flow risk not only leads to a decline in the family's quality of life and asset damage, but the solution is annuity insurance to guarantee continuous cash flow and the ability to weather difficulties through policy loans.

[0072] Therefore, electronic devices can calculate risk scanning dimensions based on a set of key fields, that is, determine the risk scanning dimension to which the set of key fields belongs, and extract target assessment rules associated with the risk scanning dimension from the target rule base. For example, if the proportion of key fields representing family information in the set of key fields is greater than or equal to a preset proportion threshold, and the number of key fields representing the economic pillar in the family information is the largest, then the risk scanning dimension can be calculated as the risk of the family's economic pillar based on the set of key fields. Then, assessment rules for evaluating the risk of the family's economic pillar, such as user health status assessment rules and user income stability assessment rules, can be extracted from the target rule base based on the risk scanning dimension.

[0073] In some embodiments, when querying target assessment rules in the rule base based on a set of key fields, the needs, types, and output methods of the risk assessment results can also be considered. To this end, the electronic device can identify result format information from the risk assessment intent, wherein the result format information is used to characterize the needs, types, and output methods of the risk assessment results. Then, the rule classification is determined based on the result format information, and target assessment rules are extracted from the rule base according to the rule classification.

[0074] For example, such as Figure 3As shown, when user information includes personal information, family information, business and asset information, and the system requires results in the form of risk level, customer needs analysis results, and comprehensive assessment report, the risk assessment system can query three types of mapping rules in the rule base: Know Your Customer (KYC) field risk mapping rules, risk needs block mapping rules, and risk summary assessment mapping rules. Then, using these three types of mapping rules, the risk assessment results, customer needs analysis results, and comprehensive assessment report are obtained, respectively.

[0075] S104. Generate risk assessment results based on the target assessment rules and key field set, and output the risk assessment results.

[0076] After obtaining the target assessment rules through the rule base query, the risk assessment system can perform risk assessment on the user information corresponding to the key field set according to the target assessment rules, so as to generate risk assessment results based on the target assessment rules and the key field set.

[0077] In some embodiments, during the risk assessment process, the electronic device can set the field values ​​of key fields into the script engine template by executing a script engine interface of a specific format, and then execute the script to determine whether the key field meets the assessment rules, that is, to determine whether the field value of the key field meets the corresponding risk item.

[0078] For example, the Java language provides the ScriptEngine interface, which can be used to execute JavaScript expressions, thereby evaluating string expressions. Electronic devices can configure risk assessment rules into templates, such as "High risk for the breadwinner of the family = "n8=true&&n55=false&&n5=suffering from a disease". Here, n8 represents the 8th item in the KYC field. By setting the corresponding KYC field value into the template and then executing the script, it can be determined whether the key field meets the risk requirements.

[0079] In some embodiments, during the risk assessment process, the risk assessment system may first receive a risk assessment instruction sent by the user, and in response to the risk assessment instruction sent by the user, generate family risk assessment results, marital risk assessment results, property inheritance risk assessment results, tax risk assessment results, and nominee shareholding risk assessment results for the target customer based on preset "information-risk" mapping relationship data, according to the personal information, family information, and enterprise and asset information.

[0080] By leveraging pre-defined "information-risk" mapping data, automated risk assessment can be achieved, improving its efficiency and accuracy. Furthermore, comprehensive analysis based on multi-dimensional information such as personal, family, and corporate assets significantly enhances the accuracy of risk assessments, covering diverse risk types including family, marriage, wealth transfer, taxation, and nominee shareholding, resulting in more comprehensive and systematic assessment outcomes.

[0081] In some embodiments, the risk assessment result includes a risk type and a risk level. To obtain the risk assessment result, when the risk assessment system generates the result based on the target assessment rule and the set of key fields, it can also read the field names and values ​​of the key fields from the set of key fields, and extract sub-rules from the target assessment rule based on the field names. The sub-rules include multiple assessment intervals, each of which is assigned an associated risk type and assessment score.

[0082] For example, the "information-cash flow risk" mapping data is as follows: If all of the following conditions are met, it is judged as low risk: the family's economic pillar is healthy, the family's economic pillar has social security, there is no debt (loan) or guarantee, there is no possible large expenditure in the future, and the family's annual income is more than twice the family's annual expenditure; If any of the following conditions are met, it is judged as medium risk: the family's economic pillar is in a sub-healthy state, and the difference between the family's annual income and the family's annual expenditure is less than 50,000; If any of the following conditions are met, it is judged as high risk: the family's economic pillar is ill, the family's annual income is less than the family's annual expenditure, there is debt (loan) and guarantee, and there is a possible large expenditure in the future.

[0083] For example, the "information-health and medical risk" mapping data is as follows: If all of the following conditions are met, it is judged as low risk: all family members are healthy, have social security, have no debt (loan) or guarantee, and have no possible large expenditures in the future; If any of the following conditions are met, it is judged as medium risk: the individual or spouse is in a sub-healthy state and has possible large expenditures in the future; If the following conditions are met, it is judged as high risk: the individual or spouse is ill and has no social security, no insurance, and has debt (loan) or guarantee.

[0084] For example, the "Information-Education Planning Risk" mapping data is as follows: if the following conditions are met simultaneously, it is judged as low risk: the total amount of education savings is greater than 1 million, and the family's economic pillar is in good health; if the following conditions are met, it is judged as medium risk: the total amount of education savings is greater than 100,000, there are no debts (loans) or guarantees, and the family's economic pillar is in good health; if any of the following conditions are met, it is judged as high risk: the total amount of education savings is less than 50,000, the family's economic pillar is ill, or there are debts (loans) or guarantees.

[0085] For example, the "Information-Retirement Planning Risk" mapping data is as follows: if at least 3 risk points are met, it is judged as high risk; if at least 1 risk point is met, it is judged as medium risk; if no risk points are met, it is judged as low risk. The risk points include the following four types: 1. Income Gap Risk Point: Annual income < 100,000, no social security or expected monthly pension > 80% of current monthly salary. 2. Asset Pressure Risk Point: Existing debt (loans) and guarantees. 3. Health Expenditure Risk Point: Health status is "suffering from an illness". 4. Family Burden Risk Point: Potential large future expenditures.

[0086] For example, the mapping relationship data of "information-special member care risk" is as follows: if all family members are healthy and the critical illness insurance coverage is ≥500,000, it is judged as low risk; if any of the following exist, it is judged as medium risk: the individual or spouse is in a sub-healthy state or has an illness, the critical illness insurance coverage is <300,000; the family's annual income is ≈ the family's annual expenditure (surplus <20%); if at least two of the following exist, it is judged as high risk: the individual or spouse is "having an illness", there is no health insurance and no critical illness insurance coverage <100,000, the family's annual income is < the family's annual expenditure.

[0087] Based on the extracted sub-rules, the target interval to which the field value belongs can be queried among multiple assessment intervals, and the assessment score associated with the target interval can be obtained. Then, based on the assessment scores corresponding to multiple key fields in the key field set, a total assessment score is calculated, and the risk level is calculated according to the total assessment score.

[0088] For example, the extracted sub-rule R1 can include [L1, L2], [L2, L3], ..., [L... n-1 L n The evaluation intervals are defined as follows. Among them, [L1, L2] correspond to an associated evaluation score of S. 12 The evaluation score for the associated settings of [L2, L3] is S. 23 ...Therefore, for the key field K1, when it is determined that it satisfies the evaluation interval [L2, L3] in sub-rule R1, the evaluation score corresponding to the key field can be determined to be S. 23 Similarly, for other key fields, the corresponding evaluation scores can be determined using the interval mapping method described above. After determining the evaluation scores for multiple key fields, the total evaluation score, S, can be calculated using a weighted summation method. s = w1S1 + w2S2 + ... + w n S n Among them, S s Indicates the total assessment score; w1, w2, ..., w nRepresenting rules R1, R2, ..., R respectively n The weights; S1, S2, ..., S n These represent the key fields K1, K2, ..., K respectively. n The assessment score is then used to calculate the risk level. This involves determining the risk level based on the interval mapping relationship between the risk level and the total assessment score, thus generating the risk assessment result.

[0089] After generating risk assessment results based on target assessment rules and key field sets, electronic devices can output the risk assessment results, that is, display the risk assessment results through input methods such as information sending and information display.

[0090] In some embodiments, when outputting the risk assessment result, the electronic device may first obtain the application scope of the user information, and then generate a demand analysis result based on the application scope and the risk assessment result. The demand analysis result includes a set of recommended information that meets the risk type and risk level within the application scope. A summary assessment report is then generated and output based on the demand analysis result and the risk assessment result.

[0091] For example, when the user input includes information such as people or things of priority, a summary evaluation report can be generated based on the preset "risk-summary" mapping data and according to each risk assessment dimension, determining the results of needs and risk analysis, risk impact analysis, and improvement method analysis. Then, based on the people or things of priority, the customer's needs and risk analysis, the potential impact and losses of the risks, and the improvement measures and suggestions, a summary evaluation report can be generated.

[0092] After generating the risk assessment results or a summary assessment report, the electronic device can visualize the risk assessment results through a results display interface. For example... Figure 4 As shown, the results display interface can include multiple visual display items, such as the people and things that users care about most, a brief analysis of customer needs and risks, the potential impact and losses of risks, and measures and suggestions for improvement.

[0093] By applying the technical solutions of the above embodiments, the risk assessment method provided can identify key fields from user information after obtaining it, thereby generating a key field set. Then, based on the key field set, it queries a target assessment rule in a rule base, and generates and outputs a risk assessment result based on the target assessment rule and the key field set. The risk assessment result includes risk type and risk level. The rule base includes basic assessment rules based on risk level settings and extended assessment rules generated by machine learning algorithms based on the basic assessment rules. Furthermore, the rule parameters in the assessment rules can be dynamically adjusted according to modification instructions based on the risk assessment result. This method can automatically analyze user profiles and perform risk assessments based on user information, improving the efficiency and accuracy of risk assessment.

[0094] In some embodiments, as a refinement and extension of the specific implementation of the above embodiments, and in order to fully illustrate the specific implementation process of this embodiment, some embodiments of this application also provide a risk assessment method, such as... Figure 5 As shown, the method includes:

[0095] S201, Receive result modification instruction;

[0096] S202. In response to the result modification instruction, obtain the user permissions corresponding to the result modification instruction;

[0097] S203. If the user's permissions are higher than or equal to the preset modification permissions, extract the specified modification type and specified modification level from the result modification instruction.

[0098] S204. Modify the risk type in the risk assessment results to the specified modification type, and modify the risk level in the risk assessment results to the specified modification level.

[0099] The result modification instruction is used to modify the risk assessment result. In some application areas, due to the specific nature of the business content, some users (such as senior business personnel) may need to modify the risk assessment result. Therefore, the risk assessment system can also support proactive modification of the risk assessment result. That is, the risk assessment system can receive result modification instructions input by users to modify the risk assessment result. Depending on the specific modification content, the result modification instruction may include at least one of specifying a modification type and a modification level.

[0100] Upon receiving a result modification instruction, the risk assessment system can respond by determining the permissions of the user who input the instruction. These user permissions characterize the modification authority of the user who sent the instruction. Therefore, the user's permissions can be determined by reading user identification information and combining it with the user database. The user's permissions are then compared with preset modification permissions. If the user's permissions are lower than the preset permissions, it means the user currently inputting the result modification instruction does not have the authority to modify the risk assessment result. Therefore, the result modification instruction can be either not executed, or a prompt message can be displayed indicating that the user lacks modification authority.

[0101] If the user's permissions are higher than or equal to the preset modification permissions, it means that the user currently inputting the result modification command has the authority to modify the risk assessment result. Therefore, the result modification command can be executed, which involves extracting the specified modification type and level from the result modification command. Based on the extracted content, the risk type and risk level in the risk assessment result are then replaced, modifying the risk type in the risk assessment result to the specified modification type and the risk level in the risk assessment result to the specified modification level.

[0102] For example, after the electronic device generates risk assessment results, demand analysis results, and a summary assessment report through backend intelligent analysis rules, if senior business personnel are not satisfied with the risk assessment results, they can choose to manually adjust them. This is done by specifying the modified risk type and risk level through the modification options in the results display interface, thus inputting the result modification command. Upon receiving the result modification command, the electronic device updates the risk level and customer focus for the target risk dimension based on the modification command for that specific risk dimension to meet actual needs. After modifying the risk assessment results, the electronic device can also automatically adjust the demand analysis results and summary assessment report, that is, regenerate the demand analysis results and summary assessment report based on the modified risk assessment results to ensure consistency in the output.

[0103] By applying the technical solutions of the above embodiments, the risk assessment method described in the above embodiments can receive a result modification instruction after generating the risk assessment result. In response to the result modification instruction, it obtains the user permissions corresponding to the result modification instruction. By comparison, if the user permissions are higher than or equal to the preset modification permissions, the specified modification type and specified modification level are extracted from the result modification instruction. This modifies the risk type in the risk assessment result to the specified modification type, and modifies the risk level in the risk assessment result to the specified modification level. This method enables the risk assessment system to support proactive result modification, allowing manual modification when the risk assessment result may not fully meet the requirements, thereby regenerating the requirements analysis results and summary assessment report based on the modified risk assessment result.

[0104] In some embodiments, as a refinement and extension of the specific implementation of the above embodiments, and to fully illustrate the specific implementation process of this embodiment, some embodiments of this application also provide a risk assessment method. This method, based on the risk assessment system's support for proactive modification, dynamically adjusts the assessment rules in the rule base according to the proactively modified content, so that the rule parameters of the second type of rules in the rule base can be dynamically adjusted according to the modification instructions based on the risk assessment results. For example... Figure 6 As shown, the method includes:

[0105] S301. Record the modified risk assessment results;

[0106] S302. Based on the modified risk assessment results, query the rules to be adjusted;

[0107] S303. Count the number of times the risk assessment results corresponding to the rules to be adjusted are modified;

[0108] S304. If the number of modifications exceeds the preset threshold, adjust the rule parameters of the rule to be adjusted according to the specified modification type and the specified modification level.

[0109] The risk assessment system can record the modified risk assessment result each time it actively modifies the risk assessment result, and query the rules to be adjusted based on the modified risk assessment result. The rules to be adjusted are the second type of rules corresponding to the modified risk assessment result.

[0110] For example, the rule base might define income risk as follows: high risk if a user's monthly salary is below a first threshold; medium risk if their monthly salary is between the first and second thresholds; and low risk if their monthly salary is above the second threshold. However, business personnel might discover that users have additional income besides their monthly salary. Therefore, the risk level determined based on monthly salary might differ from the user's actual financial situation. If the risk assessment system classifies a user's income risk as high, but business personnel verify the additional income and determine that the user's income is not low, they can manually change the income risk from high to low.

[0111] At this point, the risk assessment system can record the modified risk assessment result when changing the income risk from high to low risk; that is, the risk type is income risk and the risk level is high risk. Then, based on the recorded modified risk assessment result, it can query the rule to be adjusted, which is the assessment rule for income risk.

[0112] After querying the rules to be adjusted based on the modified risk assessment results, the electronic device can count the number of modifications to the risk assessment results corresponding to the rules to be adjusted, and trigger rule updates based on the count of modifications. If the number of modifications exceeds a preset threshold, it indicates that the modification action corresponding to the assessment rule is an occasional event, so the modification process can continue to be monitored until the rule update conditions are met. If the number of modifications exceeds the preset threshold, it indicates that the modification action corresponding to the assessment rule is not an occasional event, and the assessment rule needs to be updated. Therefore, the rule parameters of the rules to be adjusted can be adjusted according to the specified modification type and the specified modification level. The rule parameters include the assessment interval and the assessment score associated with the assessment interval.

[0113] For example, within a preset risk assessment period, if the risk assessment system detects that a user frequently modifies the assessment rules for income risk, then the income risk assessment rules can be identified as rules requiring adjustment. Furthermore, by statistically analyzing the number of modifications to the income risk assessment results, when the number of modifications reaches a preset threshold of 30, dynamic adjustments to the corresponding income risk assessment rules can be triggered. During the adjustment process, the electronic device can run an AI model to re-expand the rules based on the basic assessment rules, or perform pattern analysis based on the specific modifications made in the past 30 results to redetermine the first, second, and third values ​​in the income risk assessment rules. This ensures that the results of subsequent automatic risk assessments better meet user needs and improves the accuracy of risk assessments.

[0114] like Figure 7 As shown, in some embodiments, to facilitate dynamic adjustment of assessment rules, the risk assessment system can also incorporate a highly scalable risk dimension generation engine that supports flexible modification. The risk dimension generation engine is responsible for parsing customer data, automatically managing various risk dimension generation rules, and applying these rules to calculate customer risk dimensions. If additional risk dimensions are needed later, simply add a new risk dimension generation rule and register it with the risk dimension generation engine.

[0115] By applying the technical solutions of the above embodiments, the risk assessment method described in the above embodiments can record the modified risk assessment result when the user modifies the risk assessment result, and query the rules to be adjusted based on the modified risk assessment result. Then, it counts the number of modifications to the risk assessment result corresponding to the rule to be adjusted. If the number of modifications exceeds a preset threshold, the rule parameters of the rule to be adjusted are adjusted according to a specified modification type and a specified modification level. This method, based on the risk assessment system's support for proactive modification of assessment results, can trigger rule updates based on proactive modification actions, thereby continuously improving the risk assessment process of the system and enhancing the accuracy of risk assessment.

[0116] In some embodiments, as a specific implementation of the risk assessment method described in the above embodiments, some embodiments of this application also provide a risk assessment device, such as... Figure 8 As shown, the device includes:

[0117] The information acquisition module is used to acquire user information, which includes at least one of first information and second information; the first information is information input through an information input interface; the second information is an uploaded information file.

[0118] A field recognition module is used to identify key fields from the user information to generate a key field set, the key field set including at least one of the key fields; the key fields are used to characterize text information that is related to the risk level;

[0119] The rule query module is used to query target assessment rules in the rule base based on the set of key fields. The rule base includes a first type of rules and a second type of rules. The first type of rules are basic assessment rules based on risk level settings. The second type of rules are extended assessment rules generated by machine learning algorithms based on the first type of rules. The rule parameters in the second type of rules can be dynamically adjusted according to the modification instructions based on the risk assessment results.

[0120] The risk assessment module is used to generate risk assessment results based on the target assessment rules and the set of key fields, and to output the risk assessment results, which include risk type and risk level.

[0121] By applying the technical solutions of the above embodiments, the risk assessment device provided in the above embodiments can acquire user information through the information acquisition module and identify key fields from the user information through the field recognition module to generate a key field set. Then, the rule query module queries the target assessment rule in the rule base based on the key field set, and the risk assessment module generates and outputs the risk assessment result based on the target assessment rule and the key field set. The rule base includes basic assessment rules based on risk level settings and extended assessment rules generated by machine learning algorithms based on the basic assessment rules. Furthermore, the rule parameters in the assessment rules can be dynamically adjusted according to the modification instructions based on the risk assessment result. The device can automatically perform risk assessment based on user information and assessment rules, accurately analyze user profiles, and assess risk types and risk levels, improving the efficiency and accuracy of risk assessment.

[0122] It should be noted that other corresponding descriptions of the functional units involved in the risk assessment device provided in the embodiments of this application can be found in the corresponding descriptions in the risk assessment method provided in the above embodiments, and will not be repeated here.

[0123] This application also provides a computer device, specifically a personal computer, server, network device, etc. The computer device includes a bus, processor, memory, and communication interface, and may also include input / output interfaces and a display device. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores location information. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0124] Those skilled in the art will understand that the structure of the computer device described above is only a partial structure related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. A specific computer device may include more or fewer components, or combine certain components, or have different component arrangements.

[0125] In one embodiment, a computer-readable storage medium is also provided, which may be non-volatile or volatile, and a computer program is stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0126] In one embodiment, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0127] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0128] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0129] Any references to memory, database, or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc.

[0130] Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can take many forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0131] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchain. The processors involved in the embodiments provided in this application may be, but are not limited to, general-purpose processors, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc.

[0132] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0133] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A risk assessment method, characterized in that, The method includes: Obtain user information, wherein the user information includes at least one of first information and second information; the first information is information entered through an information input interface; and the second information is an uploaded information file. Key fields are identified from the user information to generate a set of key fields, the set of key fields including at least one of the key fields; the key fields are used to characterize text information that is related to the risk level; Based on the set of key fields, the target assessment rules are queried in the rule base, which includes a first type of rules and a second type of rules. The first type of rules are basic assessment rules based on risk level settings. The second type of rules are extended assessment rules generated by machine learning algorithms based on the first type of rules. The rule parameters in the second type of rules can be dynamically adjusted according to the modification instructions based on the risk assessment results. The risk assessment results are generated and output based on the target assessment rules and the set of key fields. The risk assessment results include risk type and risk level.

2. The method according to claim 1, characterized in that, Identify key fields from the user information to generate a set of key fields, including: Extract non-text data from the user information; Based on the data type of the non-text data, a text conversion tool is invoked; The text conversion tool is used to convert the non-text data into text data, and to generate text to be recognized based on the text data; the text to be recognized includes the text data and the text content in the user information. The key fields are extracted from the text to be identified by referring to a preset dictionary. The key fields include field names and field values. The key fields are recorded based on a preset input template to generate the set of key fields. The preset input template is a template pre-built based on the layout of the input controls in the information input interface.

3. The method according to claim 1, characterized in that, Identify key fields from the user information to generate a set of key fields, including: The semantic recognition model is invoked; the semantic recognition model is a neural network model trained based on sample business data. The user information is input into the semantic recognition model so that the semantic information of the user information can be recognized by the semantic recognition model; The user information is segmented based on the semantic information to extract the key fields from the user information; The multiple key fields are combined into the key field set.

4. The method according to claim 1, characterized in that, Based on the aforementioned set of key fields, query the target evaluation rules in the rule base, including: Obtain the risk assessment intent, which is specified by the user or obtained by performing intent recognition on the user information; The target rule base is invoked based on the stated risk assessment intent; The risk scanning dimensions are calculated based on the set of key fields, and the target assessment rules associated with the risk scanning dimensions are extracted from the target rule base.

5. The method according to claim 1, characterized in that, Risk assessment results are generated based on the target assessment rules and the set of key fields, including: Read the field name and field value of the key field from the set of key fields; Based on the field name, sub-rules are extracted from the target assessment rule. The sub-rules include multiple assessment intervals, and the assessment intervals are set with associated risk types and assessment scores. Query the target interval to which the field value belongs among multiple evaluation intervals, and obtain the evaluation score associated with the target interval; Calculate the total evaluation score based on the evaluation scores corresponding to multiple key fields in the set of key fields; The risk level is calculated based on the total assessment score.

6. The method according to claim 1, characterized in that, The method further includes: Receive a result modification instruction, the result modification instruction being used to modify the risk assessment result; the result modification instruction includes specifying the modification type and specifying the modification level; In response to the result modification instruction, the user permission corresponding to the result modification instruction is obtained, and the user permission is used to represent the modification permission of the user who sent the result modification instruction; If the user's permissions are higher than or equal to the preset modification permissions, extract the specified modification type and the specified modification level from the result modification instruction; Modify the risk type in the risk assessment result to the specified modification type, and modify the risk level in the risk assessment result to the specified modification level.

7. The method according to claim 6, characterized in that, The method further includes: Record the modified risk assessment results; Based on the modified risk assessment results, query the rules to be adjusted, where the rules to be adjusted are the second type of rules corresponding to the modified risk assessment results; The number of times the risk assessment results corresponding to the rules to be adjusted were modified was recorded; If the number of modifications exceeds a preset threshold, the rule parameters of the rule to be adjusted are adjusted according to the specified modification type and the specified modification level. The rule parameters include the evaluation interval and the evaluation score associated with the evaluation interval.

8. The method according to claim 1, characterized in that, Output the risk assessment results, including: The scope of applications for obtaining the aforementioned user information; A demand analysis result is generated based on the application scope and the risk assessment result. The demand analysis result includes a set of recommended information that meets the risk type and risk level within the application scope. A summary assessment report is generated based on the requirements analysis results and the risk assessment results. Output the summary and evaluation report.

9. A risk assessment device, characterized in that, The device includes: The information acquisition module is used to acquire user information, which includes at least one of first information and second information; the first information is information input through an information input interface; the second information is an uploaded information file. A field recognition module is used to identify key fields from the user information to generate a key field set, the key field set including at least one of the key fields; the key fields are used to characterize text information that is related to the risk level; The rule query module is used to query target assessment rules in the rule base based on the set of key fields. The rule base includes a first type of rules and a second type of rules. The first type of rules are basic assessment rules based on risk level settings. The second type of rules are extended assessment rules generated by machine learning algorithms based on the first type of rules. The rule parameters in the second type of rules can be dynamically adjusted according to the modification instructions based on the risk assessment results. The risk assessment module is used to generate risk assessment results based on the target assessment rules and the set of key fields, and to output the risk assessment results, which include risk type and risk level.

10. An electronic device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 8.

Images