Data generation method and device for insurance terms, equipment and medium

Abstract

The invention relates to the technical field of digital finance, in particular to a data generation method and device for insurance terms, equipment and a medium. According to the method, table data and text data filled by an insurance applicant are obtained, key entity information representing the family structure of the insurance applicant is output, a compliance prompt text is generated, and an insurance policy containing insurance terms is output. The method can be applied to the financial investment business field of life insurance and the like, and comprises the following steps: acquiring an insurance applicant table and text data, extracting occupational and health index information from the table data, inputting the text data into an entity recognition model to obtain family structure key entity information, and matching a compliance text according to the occupational and health index information. And generating an initial prompt text, combining the initial prompt text with the compliance text to obtain a compliance prompt text, and finally inputting the compliance prompt text into the pre-adjustment natural language model to output an insurance policy containing insurance terms. Therefore, the information filled by the insurance applicant is comprehensively analyzed, and the insurance policy containing the insurance clauses is automatically generated.

Description

Technical Field

[0001] This application relates to the field of digital financial technology, and in particular to a data generation method, apparatus, device and medium for insurance clauses. Background Technology

[0002] Within the financial investment system, insurance, as a crucial component, involves a complex process in generating insurance clauses during its daily operations. The financial sector emphasizes efficiency, accuracy, and innovation, but traditional methods of generating insurance clauses largely rely on manual processes. This not only leads to inefficiency but also increases the risk of errors and omissions, making it difficult for insurance business to keep pace with the rapid development of the financial industry as a whole and failing to fully meet the demands of the rapid expansion of insurance business under the modern financial investment system.

[0003] Currently, most insurance policy terms are generated by analyzing single types of data. Different premium rates and coverage rules are set based on factors such as the insured's age, occupation, and health condition. Once the insured's information is obtained, the system matches these rules to generate the corresponding insurance terms. However, with changes in the actual insurance market and business development, these rules need continuous adjustment and improvement. But the large number and interrelationship of rules lead to high maintenance costs and are prone to conflicts or omissions. Furthermore, this method primarily focuses on processing structured tabular data, making it difficult to effectively process and analyze text data entered by the insured.

[0004] Therefore, how to comprehensively analyze the information filled in by the policyholder and automatically generate an insurance policy containing insurance terms has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of this application provide a data generation method, apparatus, device, and medium for insurance clauses, to solve the problem of how to comprehensively analyze the information filled in by the policyholder, thereby automatically generating an insurance policy containing insurance clauses.

[0006] In a first aspect, embodiments of this application provide a data generation method for insurance clauses, including: Obtain the form data and text data filled in by the policyholder, and extract the policyholder's occupation information and health indicator information from the form data based on the form structure information; The text data is input into a preset entity recognition model, which outputs key entity information representing the policyholder's family structure. Based on the occupational information and the health indicator information, the corresponding compliance text is matched, and an initial prompt text is generated based on the occupational information, the health indicator information and the key entity information. The initial prompt text is merged with the compliance text to generate a compliance prompt text. The compliance prompt text is then input into a pre-adjusted natural language model to output an insurance policy containing the insurance terms.

[0007] Secondly, according to an embodiment of this application, a data generation device for insurance clauses includes: The structural information extraction module is used to obtain the table data and text data filled in by the policyholder, and extract the policyholder's occupational information and health indicator information from the table data based on the table structural information; The entity information extraction module is used to input the text data into a preset entity recognition model and output key entity information that represents the family structure of the policyholder. The text generation module is used to match the corresponding compliance text based on the occupational information and the health indicator information, and to generate an initial prompt text based on the occupational information, the health indicator information and the key entity information. The policy generation module is used to merge the initial prompt text with the compliance text to generate a compliance prompt text, input the compliance prompt text into a pre-adjusted natural language model, and output a policy containing insurance terms.

[0008] Thirdly, embodiments of this application provide a computer device, the computer device including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the data generation method for insurance clauses as described in the first aspect.

[0009] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the data generation method for insurance clauses as described in the first aspect.

[0010] The beneficial effects of the embodiments in this application compared with the prior art are: This application obtains form data and text data filled in by the policyholder. Based on the form structure information, it extracts the policyholder's occupational and health indicator information from the form data. The text data is input into a pre-defined entity recognition model, which outputs key entity information representing the policyholder's family structure. Based on the occupational and health indicator information, corresponding compliance text is matched. An initial prompt text is generated based on the occupational, health indicator, and key entity information. This initial prompt text is merged with the compliance text to generate a compliance prompt text. The compliance prompt text is then input into a pre-adjusted natural language model to output a policy containing insurance terms. This application can be applied to life insurance and other business areas. By obtaining form and text data from the policyholder, extracting occupational and health indicator information from the form data, inputting the text data into an entity recognition model to obtain key entity information about the family structure, matching compliance text based on occupational and health indicator information to generate an initial prompt text, merging the initial prompt text with the compliance text to obtain a compliance prompt text, and finally inputting it into a pre-adjusted natural language model to output a policy containing insurance terms, this application comprehensively analyzes the information filled in by the policyholder to automatically generate a policy containing insurance terms. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an application environment for a data generation method for insurance clauses provided in Embodiment 1 of this application; Figure 2 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 2 of this application; Figure 3 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 3 of this application; Figure 4 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 4 of this application; Figure 5 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 5 of this application; Figure 6 This is a schematic diagram of the structure of a data generation device for insurance clauses provided in Embodiment Six of this application; Figure 7 This is a schematic diagram of the structure of a computer device provided in Embodiment 7 of this application. Detailed Implementation

[0013] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0014] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0015] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0016] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0017] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0018] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0019] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0020] To illustrate the technical solution of this application, specific embodiments are described below.

[0021] The data generation method for insurance clauses provided in Embodiment 1 of this application can be applied to, for example... Figure 1 In this application environment, the client and server communicate with each other. Users can provide conditions, requirements, and operational instructions for generating data used in insurance clauses through the client. The server then generates control instructions for the data generation method based on the content sent by the client. For example, in a financial system, a user executes a business operation by manipulating the client. Upon receiving the business operation, the server generates corresponding operation instructions and processes these instructions according to the data generation method for insurance clauses, resulting in an insurance policy containing the insurance clauses. The client includes, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a dedicated server or a server cluster consisting of multiple servers.

[0022] See Figure 2 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 2 of this application. The aforementioned data generation method for insurance clauses can be applied to... Figure 1 The server-side method for generating data for insurance terms may include the following steps: Step S201: Obtain the form data and text data filled in by the policyholder, and extract the policyholder's occupational information and health indicator information from the form data based on the form structure information.

[0023] Optionally, the step S201, which involves extracting the policyholder's occupational information and health indicator information from the table data based on the table structure information, may include the following steps: Based on the table structure information, the table data is compared one by one with the preset risk mapping table to obtain the risk information corresponding to each data. The risk information corresponding to the data representing occupation is extracted as the policyholder's occupational information, and the risk information corresponding to the data representing health is extracted as the policyholder's health indicator information.

[0024] The pre-defined risk mapping table is a reference table developed by the insurance company based on extensive historical data, industry experience, and risk assessment models. It associates different data items with corresponding risk levels and types. For example, in terms of occupation, the risk mapping table marks professions such as firefighters and miners as high-risk occupations because these professions involve a higher risk of accidental injury and health hazards during work. Regarding health indicators, for indicators such as blood pressure and blood sugar, different risk levels are assigned based on different numerical ranges; for example, hypertension and hyperglycemia correspond to high-risk conditions.

[0025] The one-to-one comparison process uses the table structure information to clearly identify each field in the table data. Then, the specific data in these fields is matched with the corresponding entries in the risk mapping table. For example, if the table shows that the insured's occupation is "construction worker," by comparing it with the risk mapping table, we can know the corresponding risk information for construction workers, which might be "the work environment has risks such as falls from heights and being struck by objects, resulting in a high accident rate." Similarly, if the table records that the insured's blood pressure is 160 / 100 mmHg, after comparing it with the risk mapping table, we can obtain risk information such as "stage 2 hypertension, significantly increased risk of cardiovascular disease."

[0026] The risk information corresponding to the data representing occupation is extracted as the policyholder's occupational information, and the risk information corresponding to the data representing health is extracted as the policyholder's health indicator information. After the comparison is completed, the financial system filters out occupation-related data from a large dataset and uses the corresponding risk information as the policyholder's occupational information. For example, occupation-related fields in the table may include "occupational category" and "job description." The risk information obtained by comparing the data in these fields with the risk mapping table, such as "high-risk occupation, prone to accidental injury at work," constitutes the policyholder's occupational information. Insurance companies can use this information to assess the accidental risks faced by the policyholder during work, and thus determine the premium and coverage of accident insurance or other related insurance.

[0027] Similarly, for data representing health, such as various physical examination indicators (blood routine, urine routine, electrocardiogram, etc.), the risk information obtained by comparing them with a risk mapping table, such as "abnormal liver function, potential risk of liver disease," is used as the insured's health indicator information. This information helps insurance companies assess the insured's health status, determine whether there are potential health problems, and thus decide whether to underwrite and how to set insurance liabilities and rates.

[0028] For example, in the life insurance business, an insured person fills out a form with "pilot" in the occupation field and "high blood lipids (total cholesterol 6.5 mmol / L)" in the health indicator field. The insurance company compares "pilot" and "high blood lipids (total cholesterol 6.5 mmol / L)" against a pre-set risk mapping table based on the form's structure. The risk mapping table shows that pilots are a special profession; although the working environment is relatively safe, long-term high-altitude flight may have some impact on the cardiovascular system, classifying it as a medium-risk profession. For total cholesterol 6.5 mmol / L, the risk mapping table indicates that this value is higher than the normal range, increasing the risk of cardiovascular disease. Therefore, the insured person's occupation information is "pilot (medium-risk profession, long-term high-altitude flight may affect the cardiovascular system)," and the health indicator information is "high blood lipids (total cholesterol 6.5 mmol / L, increased risk of cardiovascular disease)."

[0029] Step S202: Input the text data into a preset entity recognition model and output key entity information representing the policyholder's family structure.

[0030] Optionally, a preset entity recognition model is used to extract entities from the text data to obtain at least one key entity, which is used to represent any person in the policyholder's family structure. Using a preset language model, the contextual semantics corresponding to each key entity are determined from the text data. Based on the contextual semantics corresponding to each key entity, the clause language is converted for each key entity to obtain the key entity information of the policyholder's family structure.

[0031] The pre-defined entity recognition model is a trained artificial intelligence model that is taught to identify specific types of entities in text. In this scenario, the specific type refers to entities related to family structure. By learning from a large amount of text data containing descriptions of family relationships, the model has mastered lexical features and contextual patterns representing family members, such as "father," "mother," "wife," and "son."

[0032] When the text data filled in by the policyholder is input into this model, the model analyzes the text word by word and sentence by sentence. Based on the features and patterns learned previously, it identifies words that represent family members as key entities. For example, if the text mentions "I live with my spouse and son," the entity recognition model will extract the two key entities "spouse" and "son" because they represent members of the policyholder's family structure.

[0033] Using a pre-defined language model, the contextual semantics corresponding to each key entity are determined from the text data. Based on the contextual semantics of each key entity, clause language conversion is performed on each key entity to obtain the key entity information of the policyholder's family structure. The pre-defined language model can understand the semantics and context of text. For each extracted key entity, the language model analyzes its context to determine its precise semantics. For example, in the sentence "My daughter just got married last month and now lives next door with her husband," the language model knows from the context that the "husband" is the "daughter's" spouse. In this way, the language model can construct an accurate semantic environment for each key entity.

[0034] Insurance policy terms typically use standardized and rigorous language. In insurance practice, terms like "partner" are uniformly translated into "spouse," and terms like "son" and "daughter" are categorized as "children." Based on the previously determined contextual semantics, the language model converts each key entity into a term conforming to the insurance policy's specifications. This ensures a consistent and clear definition of family structure information within the insurance contract, avoiding ambiguity caused by non-standard terminology and guaranteeing the clear definition and accurate execution of insurance benefits. For example, by converting "partner" to "spouse" in the previous example and retaining "son," the final key entity information for the policyholder's family structure becomes "spouse, children." For example, in the financial insurance field, the policyholder writes in the text: "I live with my father, my wife, and my daughter." An entity recognition model will extract three key entities from this text: "father," "wife," and "daughter," because these represent members of the policyholder's family structure. The language model analyzes the context and knows that "father" refers to the policyholder's father, "wife" is the policyholder's spouse, and "daughter" is the policyholder's daughter. According to the insurance policy terms, "father" is converted to "father," "wife" to "spouse," and "daughter" to "child." The final key entity information for the policyholder's family structure is "father, spouse, and child."

[0035] Step S203: Match the corresponding compliance text based on the occupational information and the health indicator information, and generate an initial prompt text based on the occupational information, the health indicator information and the key entity information.

[0036] Optionally, the step S203, which involves matching the corresponding compliance text based on the occupational information and the health indicator information, may include the following steps: Based on the health indicator information, an insurance waiting period is generated; based on the occupation information, an accidental injury compensation limit is generated. Based on the insurance waiting period and the accidental compensation limit, and in conjunction with the preset exclusion clauses, a compliance text is generated.

[0037] The initial notification text is a preliminary summary and distillation of the policyholder's relevant information. It includes key elements such as the policyholder's basic information, occupation, health status, and family structure. This text allows insurance company personnel to quickly understand the policyholder's overall risk profile and serves as a clear information carrier during subsequent communications, helping the policyholder better understand the general content of the insurance contract and the potential risks involved.

[0038] The insurance waiting period, also known as the observation period or exclusion period, refers to a period after the insurance contract takes effect during which the insurance company is not liable for certain insurance liabilities. The purpose of setting a waiting period is to prevent policyholders from purchasing insurance with pre-existing conditions, thus protecting the normal operation of the insurance company and the interests of other policyholders.

[0039] Health indicators reflect the insured's current health status. If the insured's health indicators show potential health problems, such as high blood pressure or diabetes, the insurance company may extend the waiting period. For example, for insured individuals with slightly elevated blood pressure or blood sugar levels, the waiting period may be extended from the usual 90 days to 180 days to allow sufficient time to observe whether the insured's health condition deteriorates. Conversely, if the insured's health indicators are good, the insurance company may use a shorter waiting period.

[0040] Accidental death benefit restrictions refer to limitations that insurance companies impose on the amount, scope, or conditions of compensation for accidents under certain circumstances. Different occupations face different levels of accidental risk, therefore, accidental death benefit restrictions need to be set reasonably based on occupational information.

[0041] Some high-risk professions, such as firefighters, construction workers, and miners, face a higher risk of accidental injury during their work. For policyholders in these professions, insurance companies may set lower accidental injury reimbursement ratios or exclude certain types of accidents from coverage. For example, for construction workers, in the event of an accident while working at heights, the insurance company may only reimburse a portion of medical expenses, while occupational diseases caused by the inherent high risk of the job may not be covered. Conversely, for some low-risk professions, such as office workers, the restrictions on accidental injury reimbursement may be relatively fewer.

[0042] Based on the aforementioned waiting period and accidental death benefit limits, and in conjunction with pre-defined exclusion clauses, a compliance text is generated. Exclusion clauses are provisions in the insurance contract under which the insurance company will not assume liability. These clauses are formulated by the insurance company based on laws and regulations and its own risk tolerance. Common exclusion clauses include losses caused by the policyholder's intentional acts, illegal or criminal acts, war, nuclear explosions, etc.

[0043] By combining the previously generated waiting period and accidental death benefit limits with the pre-defined exclusion clauses, a complete compliance document is formed. This document clarifies the circumstances under which the insurance company will not assume liability, as well as the specific provisions within the waiting period and accidental death benefit limits. For example, the compliance document might state: "The waiting period for this insurance contract is 180 days. During the waiting period, the insurance company will not reimburse medical expenses incurred by the insured due to illness. For insured individuals engaged in high-risk occupations, the reimbursement rate for accidents occurring during work hours is 80%. Furthermore, the insurance company will not assume liability for losses caused by the insured's intentional crimes, suicide, or similar acts."

[0044] For example, in the area of ​​policy drafting, the insured's occupation is listed as "coal miner," and their health indicators show "high blood pressure (150 / 95 mmHg)." Key entity information includes "spouse and children." Due to the insured's high blood pressure, the insurance company sets the waiting period at 180 days. Considering that coal miners are in a high-risk profession, the insurance company stipulates that 70% of the compensation will be paid for accidents occurring during coal mining work, and occupational diseases specific to coal mining work are not covered. Combined with pre-defined exclusion clauses, such as those for intentional acts or crimes committed by the insured, the resulting compliance text reads: "The waiting period for this insurance contract is 180 days. During the waiting period, expenses incurred by the insured due to illness will not be reimbursed. As a coal miner, the insured will receive 70% compensation for accidents occurring during coal mining work, but occupational diseases caused by coal mining work are not covered. Furthermore, the insurance company is not liable for losses caused by intentional acts or crimes committed by the insured." Step S204: Merge the initial prompt text with the compliance text to generate a compliance prompt text, input the compliance prompt text into the pre-adjusted natural language model, and output the insurance policy containing the insurance terms.

[0045] Both the initial notification text and the compliance text revolve around insurance-related information, but with different focuses. The initial notification text emphasizes the policyholder's individual information and preliminary risk warnings, while the compliance text focuses on the compliance requirements and liability limitations of the insurance contract. Combining them makes the information more complete and comprehensive, allowing policyholders to understand the connection between their situation and the insurance, as well as the various provisions of the insurance contract, all from a single text, avoiding comprehension difficulties and potential omissions caused by fragmented information.

[0046] During the merger process, it is necessary to organize the structure of the two parts reasonably to ensure logical clarity. The information in the initial notification text can be used as an introduction, first introducing the policyholder's basic information and preliminary risk warnings, and then listing the insurance waiting period, accidental injury compensation limitations, exclusions, etc., from the compliance text in sequence, forming a coherent and organized document. For example, if the initial notification text states, "The policyholder's occupation is a coal miner, health indicators show high blood pressure, and family structure includes spouse and children. The insurance waiting period is 180 days, and the accidental injury compensation ratio during work is 70%," and the compliance text stipulates, "The insurance company is not liable for losses caused by the insured's intentional acts, illegal activities, or crimes," the merged compliance notification text can be stated as: "The policyholder's occupation is a coal miner, health indicators show high blood pressure, and family structure includes spouse and children. The insurance waiting period for this insurance contract is 180 days. As a coal miner, the insured's accidental injury during work is compensated at a ratio of 70%. However, the insurance company is not liable for losses caused by the insured's intentional acts, illegal activities, or crimes." The compliance prompt text is input into a pre-tuned natural language model, which outputs an insurance policy containing the insurance terms. A natural language model is an artificial intelligence model capable of processing and generating natural language text. In insurance business, the model needs to be pre-tuned to adapt to the professional terminology, grammatical rules, and business logic of the insurance field. The pre-tuning process typically involves training the model using a large amount of insurance terms data, allowing the model to learn the standardized expression, clause structure, and legal requirements of insurance contracts. Through pre-tuning, the model can better understand the information in the compliance prompt text and transform it into formal terms that conform to insurance industry standards.

[0047] Once the compliance notification text is input into the pre-adjusted natural language model, the model analyzes and processes the text, reorganizing and expanding the information according to the standard format and clause structure of an insurance contract. It uses standardized insurance terminology to elaborate on insurance liabilities, compensation conditions, exclusions, and the rights and obligations of both parties, ultimately generating a complete policy containing the insurance terms.

[0048] For example, in the business scenario of formulating life insurance policy regulations, the natural language model will expand the "insurance waiting period is 180 days" in the compliance prompt text to "180 days from the effective date of this insurance contract is the insurance waiting period. During the waiting period, the company will not be liable for paying insurance benefits in the event of medical expenses incurred by the insured due to illness, death or total disability, except for insured events caused by accidental injury." See Figure 3 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 3 of this application. Figure 3 As shown, the step S203 above, which involves generating initial prompt text based on the occupational information, the health indicator information, and the key entity information, may include the following steps: Step S301: Based on the occupational information, generate corresponding claims case generation scripts; Step S302: Generate a script by combining the occupational information, health indicator information, key entity information, and claims case to obtain the combined text as the initial prompt text.

[0049] Different professions face different risks, and insurance claims will vary accordingly. By generating claim case scripts that correspond to occupational information, policyholders can more intuitively understand the potential claim situations they may encounter in insurance scenarios related to their profession, enhancing their awareness of the actual role of insurance.

[0050] Insurance companies compile typical claim cases based on extensive past claims data, categorized by profession. Once they obtain the policyholder's occupational information, they can select relevant cases from this database and translate them into appropriate sales pitches.

[0051] By piecing together information from multiple sources, this system provides policyholders with a comprehensive overview of insurance-related information. Occupational information reflects the risks associated with the policyholder's work; health indicators demonstrate the impact of their physical condition on insurance; key entity information relates to the connection between the policyholder's family structure and insurance benefits; and claims case scripts use real-life examples to give policyholders a more concrete understanding of the insurance claims process. Integrating this information allows policyholders to grasp the multifaceted connection between their situation and insurance within a single text document.

[0052] When assembling the information, it's crucial to maintain a logical organization to ensure the text is clear and easy to understand. You can begin by introducing the policyholder's basic information, such as occupation and health indicators, then mention key details related to their family structure, and finally introduce a claims case to generate the appropriate message. For example, the initial prompt could be: "You work as a courier, and your current health indicators show slightly high blood pressure. Your family structure includes your spouse and children. In past insurance claims, some couriers have been injured in traffic accidents while delivering goods and ultimately received corresponding medical expense compensation and lost wages compensation." In this embodiment of the application, through the initial prompt text, the policyholder can comprehensively understand the relationship between their occupation, health, family structure and insurance claims, thereby better assessing their insurance needs.

[0053] See Figure 4 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 4 of this application. Figure 4 As shown, in step S204, after inputting the compliance prompt text into the pre-adjusted natural language model and outputting an insurance policy containing the insurance terms, the data generation method for the insurance terms may further include the following steps: Step S401: Extract each insurance clause from the policy, and for any insurance clause, split the insurance clause into at least one sentence; Step S402: Perform similarity matching between each sentence and each clause template in the preset knowledge base to obtain all similarity matching results for the corresponding sentence; Step S403: For any sentence, if all similarity matching results corresponding to the sentence do not meet the preset conditions, then the missing content of the sentence is determined.

[0054] An insurance policy typically contains multiple different insurance clauses, each specifying a particular aspect of the insurance contract, such as coverage, exclusions, and claims procedures. This step requires extracting each clause from the policy individually so that each clause can be processed separately later.

[0055] Each extracted insurance clause is further broken down into individual sentences. This is done to analyze the clause content more meticulously, as sentences are the basic units for expressing complete meaning; analyzing sentences allows for a more accurate identification of potential problems within the clauses.

[0056] The preset knowledge base is a database storing a large number of standard insurance clause templates. These templates are summarized based on insurance industry norms, laws and regulations, and the experience of insurance companies, representing the standard expression and content structure of insurance clauses. Each clause template contains the key information and sentence structure that the clause should have.

[0057] Each segmented sentence is matched against each clause template in a pre-defined knowledge base based on similarity. Various similarity matching methods can be used, such as semantic-based matching and keyword-based matching. The matching result for each sentence and each template is obtained by calculating the similarity score between the sentence and the template.

[0058] The preset condition is a pre-defined similarity threshold used to determine whether the match between a sentence and a clause template is high enough. If the similarity score of a sentence with a clause template reaches or exceeds this threshold, the sentence is considered to match the template well and the content is basically complete. Conversely, if the similarity score of a sentence with all clause templates is below this threshold, it indicates that the sentence may have missing content.

[0059] In this embodiment of the application, the review process can promptly identify any missing content in the insurance terms, allowing for further improvement and correction of the policy, thereby enhancing the quality and standardization of the insurance terms.

[0060] See Figure 5 This is a flowchart illustrating a data generation method for insurance clauses provided in Embodiment 5 of this application. Figure 5 As shown, after inputting the compliance prompt text into the pre-adjusted natural language model and outputting an insurance policy containing the insurance terms, the data generation method for the insurance terms may further include the following steps: Step S501: Extract the premium information from the policy, and extract the policyholder's occupational and gender characteristics from the policy; Step S502: Based on the occupational characteristics and the gender characteristics, determine the occupation and gender that are significantly different from the occupational characteristics and the gender characteristics; Step S503: Obtain the target premium for the occupation and gender with significant differences, compare the premium information with the difference of the target premium, and obtain the comparison result; Step S504: Adjust the premium information according to the comparison results to obtain the adjusted policy.

[0061] Among them, the premium is the fee that the policyholder needs to pay to the insurance company in the insurance contract, and it is an important part of the insurance transaction. Accurately extracting premium information from the policy, including the specific premium amount, payment method (such as annual payment or monthly payment), payment period, etc., provides basic data for subsequent premium assessment and adjustment.

[0062] An insured's occupational and gender characteristics reflect, to some extent, the level of risk they face. Different occupations present different work risks; for example, construction workers may face a higher risk of accidental injury. Differences in health status and life expectancy between genders also affect insurance risk. Extracting these characteristics from insurance policies allows for subsequent analysis of premium differences among different characteristic groups.

[0063] Insurance companies analyze the differences in insurance risk across different occupations and genders based on extensive historical claims data and risk assessment models. Based on these criteria, they identify occupational and gender groups that exhibit significant differences from the insured's occupational and gender characteristics.

[0064] Insurance companies set corresponding premium standards for different occupational and gender groups, which are determined based on the average risk level of that group. They also use internal data systems or actuarial models to obtain target premiums for groups whose characteristics differ significantly from those of the insured.

[0065] The policyholder's premium information is compared with the target premium for a significantly different group, and the difference between the two is calculated. The comparison indicators may include the difference in premium amount, the difference in premium rate, etc.

[0066] Based on the comparison results, if there is a significant difference between the policyholder's premium and the target premium for groups with large differences, the premium needs to be adjusted. The principle of adjustment is to make the premium more fair and reasonable, reflecting the policyholder's actual risk level. If the policyholder's premium is too high, it may be necessary to appropriately reduce the premium; if the premium is too low, it may be necessary to appropriately increase the premium.

[0067] After adjusting the premium, the new premium information is updated in the policy, generating the adjusted policy. This new policy has a premium setting that better suits the risk profiles of different groups, improving the fairness and market competitiveness of insurance products.

[0068] In this embodiment, an adjusted insurance policy is generated after comprehensive evaluation. This reflects the risk differences between different occupations and makes the premium setting more fair and reasonable.

[0069] Corresponding to the data generation method for insurance clauses in the above embodiments, Figure 6 This paper shows a structural block diagram of a data generation apparatus for insurance clauses provided in Embodiment Six of this application. The aforementioned data generation apparatus for insurance clauses can be applied to… Figure 1 The server-side component is shown. For ease of explanation, only the parts relevant to the embodiments of this application are shown.

[0070] See Figure 6 The data generation device for insurance terms includes: The structural information extraction module 61 is used to obtain the table data and text data filled in by the policyholder, and extract the policyholder's occupational information and health indicator information from the table data based on the table structural information; The entity information extraction module 62 is used to input the text data into a preset entity recognition model and output key entity information representing the policyholder's family structure; The text generation module 63 is used to match the corresponding compliance text based on the occupational information and the health indicator information, and to generate an initial prompt text based on the occupational information, the health indicator information and the key entity information. The policy generation module 64 is used to merge the initial prompt text with the compliance text to generate a compliance prompt text, input the compliance prompt text into a pre-adjusted natural language model, and output a policy containing insurance terms.

[0071] Optionally, the structural information extraction module 61 includes: The risk information acquisition unit is used to compare the table data with a preset risk mapping table one by one based on the table structure information to obtain the risk information corresponding to each data. The information extraction unit is used to extract the risk information corresponding to the data representing occupation as the policyholder's occupational information, and to extract the risk information corresponding to the data representing health as the policyholder's health indicator information.

[0072] Optionally, the text generation module 62 includes: The claims generation unit is used to generate an insurance waiting period based on the health indicator information and to generate accidental compensation restrictions based on the occupation information. The compliance text generation unit is used to generate compliance text based on the insurance waiting period and the accidental compensation limit, combined with preset disclaimers.

[0073] Optionally, the entity information extraction module 62 includes: An entity extraction unit is used to extract entities from the text data using a preset entity recognition model to obtain at least one key entity, which is used to represent any person in the policyholder's family structure. The conversion unit is used to determine the contextual semantics corresponding to each key entity from the text data using a preset language model, and to perform clause language conversion on each key entity according to the contextual semantics corresponding to each key entity, so as to obtain the key entity information of the policyholder's family structure after conversion.

[0074] Optionally, the text generation module 63 includes: Based on the occupational information, generate corresponding claims case generation scripts; The occupational information, health indicator information, key entity information, and claims case are used to generate a script, which is then spliced ​​together to obtain the spliced ​​text as the initial prompt text.

[0075] Optionally, the data generation device for insurance terms also includes: The splitting module is used to extract each insurance clause in the policy after the compliance prompt text is input into the pre-adjusted natural language model and the policy containing the insurance clauses is output, and to split each insurance clause into at least one sentence for any insurance clause. The similarity matching module is used to perform similarity matching between each sentence and each clause template in the preset knowledge base to obtain all similarity matching results for the corresponding sentence; The sentence content determination module is used to determine the missing content of a sentence if all similarity matching results corresponding to the sentence do not meet the preset conditions.

[0076] Optionally, the data generation device for insurance terms also includes: The feature extraction module is used to extract the premium information from the policy after the compliance prompt text is input into the pre-adjusted natural language model and the policy containing the insurance terms is output, and to extract the occupational and gender characteristics of the policyholder from the policy. The feature determination module is used to determine the occupation and gender that are significantly different from the occupation and gender features, based on the occupation features and the gender features. The comparison module is used to obtain the target premiums for the occupations and genders with significant differences, compare the premium information with the differences in the target premiums, and obtain the comparison results. The adjustment module is used to adjust the premium information according to the comparison results to obtain the adjusted policy.

[0077] It should be noted that the information interaction and execution process between the above modules, units, and sub-units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0078] Figure 7 This is a schematic diagram of the structure of a computer device provided in Embodiment Seven of this application. Figure 7 As shown, the computer device of this embodiment includes: at least one processor ( Figure 7Only one is shown in the diagram), a memory, and a computer program stored in the memory and executable on at least one processor. When the processor executes the computer program, it implements the steps of any of the above-described data generation methods for insurance clauses or embodiments of data generation methods for insurance clauses.

[0079] This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 7 The examples of computer devices are merely examples and do not constitute a limitation on computer devices. Computer devices may include more or fewer components than shown in the illustration, or combinations of certain components, or different components, such as network interfaces, displays, and input devices.

[0080] The processor referred to can be a CPU, but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0081] Memory includes readable storage media, internal memory, etc., wherein internal memory can be the RAM of a computer device, providing an environment for the operation of the operating system and computer-readable instructions stored in the readable storage media. The readable storage media can be the hard drive of a computer device, or in other embodiments, it can be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, memory can include both internal storage units and external storage devices of the computer device. Memory is used to store the operating system, applications, bootloader, data, and other programs, such as program code for computer programs. Memory can also be used to temporarily store data that has been output or will be output.

[0082] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium can include at least: any entity or device capable of carrying computer program code, a recording medium, a computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0083] The implementation of all or part of the processes in the methods of the above embodiments can also be accomplished by a computer program product. When the computer program product is run on a computer device, it enables the computer device to execute the steps in the above method embodiments.

[0084] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0085] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0086] In the embodiments provided in this application, it should be understood that the disclosed apparatus / computer devices and methods can be implemented in other ways. For example, the apparatus / computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0087] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0088] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A data generation method for insurance clauses, characterized in that, include: Obtain the form data and text data filled in by the policyholder, and extract the policyholder's occupation information and health indicator information from the form data based on the form structure information; The text data is input into a preset entity recognition model, which outputs key entity information representing the policyholder's family structure. Based on the occupational information and the health indicator information, the corresponding compliance text is matched, and an initial prompt text is generated based on the occupational information, the health indicator information and the key entity information. The initial prompt text is merged with the compliance text to generate a compliance prompt text. The compliance prompt text is then input into a pre-adjusted natural language model to output an insurance policy containing the insurance terms.

2. The data generation method for insurance clauses according to claim 1, characterized in that, The step of extracting the policyholder's occupational information and health indicator information from the table data based on the table structure information includes: Based on the table structure information, the table data is compared one by one with the preset risk mapping table to obtain the risk information corresponding to each data. The risk information corresponding to the data representing occupation is extracted as the policyholder's occupational information, and the risk information corresponding to the data representing health is extracted as the policyholder's health indicator information.

3. The data generation method for insurance clauses according to claim 2, characterized in that, The step of matching the corresponding compliance text based on the occupational information and the health indicator information includes: Based on the health indicator information, an insurance waiting period is generated; based on the occupation information, an accidental injury compensation limit is generated. Based on the insurance waiting period and the accidental compensation limit, and in conjunction with the preset exclusion clauses, a compliance text is generated.

4. The data generation method for insurance clauses according to claim 1, characterized in that, The step of inputting the text data into a preset entity recognition model and outputting key entity information representing the policyholder's family structure includes: Using a preset entity recognition model, entities are extracted from the text data to obtain at least one key entity, which is used to represent any person in the policyholder's family structure. Using a preset language model, the contextual semantics corresponding to each key entity are determined from the text data. Based on the contextual semantics corresponding to each key entity, the clause language is converted for each key entity to obtain the key entity information of the policyholder's family structure.

5. The data generation method for insurance clauses according to claim 1, characterized in that, The step of generating initial prompt text based on the occupational information, the health indicator information, and the key entity information includes: Based on the occupational information, generate corresponding claims case generation scripts; The occupational information, health indicator information, key entity information, and claims case are used to generate a script, which is then spliced ​​together to obtain the spliced ​​text as the initial prompt text.

6. The data generation method for insurance clauses according to claim 1, characterized in that, After inputting the compliance prompt text into the pre-tuned natural language model and outputting an insurance policy containing the insurance terms, the process further includes: Extract each insurance clause from the policy, and for any given insurance clause, break it down into at least one sentence; Each sentence is matched against each clause template in the preset knowledge base to obtain all similarity matching results for the corresponding sentence; If, for any given sentence, all similarity matching results for that sentence do not meet the preset conditions, then the sentence is determined to be missing content.

7. The data generation method for insurance clauses according to any one of claims 1 to 6, characterized in that, After inputting the compliance prompt text into the pre-tuned natural language model and outputting an insurance policy containing the insurance terms, the process further includes: Extract the premium information from the policy, and extract the policyholder's occupational and gender characteristics from the policy; Based on the occupational characteristics and the gender characteristics, determine the occupations and genders that differ significantly from the occupational characteristics and the gender characteristics; Obtain the target premiums for the occupations and genders with significant differences, compare the premium information with the differences in the target premiums, and obtain the comparison results; Based on the comparison results, the premium information is adjusted to obtain the adjusted policy.

8. A data generation device for insurance clauses, characterized in that, include: The structural information extraction module is used to obtain the table data and text data filled in by the policyholder, and extract the policyholder's occupational information and health indicator information from the table data based on the table structural information; The entity information extraction module is used to input the text data into a preset entity recognition model and output key entity information that represents the family structure of the policyholder. The text generation module is used to match the corresponding compliance text based on the occupational information and the health indicator information, and to generate an initial prompt text based on the occupational information, the health indicator information and the key entity information. The policy generation module is used to merge the initial prompt text with the compliance text to generate a compliance prompt text, input the compliance prompt text into a pre-adjusted natural language model, and output a policy containing insurance terms.

9. A computer device, characterized in that, The computer device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the data generation method for insurance clauses as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the data generation method for insurance clauses as described in any one of claims 1 to 7.

Images