Automatic intelligent underwriting method, device, equipment and medium

Abstract

The invention relates to the technical field of intelligent decision making, can be applied to business system platforms of financial science and technology, medical health and the like, and discloses an automatic intelligent underwriting method, device, equipment and medium, and the method comprises the steps: obtaining multi-source insuring data of a target insuring object, carrying out the multi-source data fusion of the multi-source insuring data, and carrying out the multi-source data fusion; obtaining target insurance data; carrying out attribute identification on the target insurance data to obtain an object attribute of the target insurance object; obtaining an underwriting parameter model of the target insurance data, and dynamically adjusting the underwriting parameter model according to the object attribute to obtain a target underwriting parameter model matched with the object attribute; and underwriting the target insurance data according to the target underwriting parameter model to obtain an underwriting result of the target insurance data. According to the invention, the underwriting efficiency and the underwriting result accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent decision-making technology, and in particular to an automated intelligent underwriting method, device, equipment and medium. Background Art

[0002] Underwriting is the process by which an insurance company evaluates and analyzes an applicant's health, financial situation, occupational risks, and other information to determine whether to underwrite the policy and determine the terms and premium. Underwriting is an essential part of the insurance industry, protecting the interests of the insurance company while also providing professional assessment and protection for the policyholder. Within the insurance industry, underwriting is a crucial step in determining policy acceptance and premium determination.

[0003] For example, in the healthcare sector, policyholders expect immediate coverage decisions through automated underwriting using electronic medical records. Current rule-based engine systems only support simple dimensional judgments such as "age range filtering" and "underlying disease screening," and are unable to build intelligent decision-making models that integrate multimodal data. When faced with the problem of parsing unstructured medical examination reports, traditional systems lack the ability to correlate medical knowledge graphs, requiring manual intervention to process unusual combinations of medical signs, significantly reducing the efficiency of handling complex cases.

[0004] For example, in the fintech sector, existing underwriting tools rely solely on linear regression models for underwriting assessments. When assessing the correlation between multi-dimensional user profiles and contracting risk, the system struggles to dynamically integrate heterogeneous data sources such as credit scores, geographic information, and device IoT data. This results in a single dimension in underwriting assessments and inaccurate underwriting results.

[0005] Therefore, how to improve underwriting efficiency and the accuracy of underwriting results has become an urgent problem to be solved. Summary of the Invention

[0006] The present invention provides an automated intelligent underwriting method, apparatus, equipment and medium, the main purpose of which is to solve the problems of low underwriting efficiency and inaccurate underwriting results.

[0007] In a first aspect, to achieve the above-mentioned objectives, the present invention provides an automated intelligent underwriting method, comprising:

[0008] Acquire multi-source insurance data of a target insured person, and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data;

[0009] Performing attribute recognition on the target insurance data to obtain object attributes of the target insurance object;

[0010] Obtaining an underwriting parameter model for the target insurance data, and dynamically adjusting the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes;

[0011] The target insurance data is underwritten according to the target underwriting parameter model to obtain an underwriting result of the target insurance data.

[0012] In a second aspect, the present invention further provides an automated intelligent underwriting device, comprising:

[0013] A data fusion module is used to obtain multi-source insurance data of a target insured person, and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data;

[0014] An attribute recognition module, configured to perform attribute recognition on the target insurance data to obtain the object attributes of the target insurance object;

[0015] a dynamic adjustment module, configured to obtain an underwriting parameter model of the target insurance data, and dynamically adjust the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes;

[0016] The real-time underwriting module is used to underwrite the target insurance data according to the target underwriting parameter model to obtain the underwriting result of the target insurance data.

[0017] In a third aspect, the present invention further provides an electronic device, comprising:

[0018] at least one processor; and,

[0019] a memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned automated intelligent underwriting method.

[0021] In a fourth aspect, the present invention also provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is executed by a processor in an electronic device to implement the above-mentioned automated intelligent underwriting method.

[0022] The present invention performs a series of processing such as data fusion and data cleaning on multi-source insurance data, which can identify and correct errors, omissions and inconsistencies in the data, ensure the accuracy and uniqueness of the data, and thus significantly improve the quality of target insurance data; through attribute recognition, it can comprehensively and accurately collect and analyze insurance data to more accurately assess the risks of insurance data, and at the same time quickly extract key information from the target insurance data, which can shorten the underwriting process; through the precise matching of object attributes and underwriting parameter models, it can effectively avoid human errors, ensure the accuracy of underwriting results, reduce the insurance company's claim risk, and flexibly adapt to various complex situations to achieve personalized underwriting; through risk scoring of target insurance data, it can eliminate fluctuations in underwriting standards caused by differences in human experience, ensure the objectivity and standardization of underwriting assessments, thereby improving underwriting efficiency, and also improving the accuracy and comprehensiveness of underwriting. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0024] Figure 1 A schematic diagram of an application environment of an automated intelligent underwriting method according to an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of a process flow of an automated intelligent underwriting method provided by one embodiment of the present invention;

[0026] Figure 3 A schematic diagram of a process for performing multi-source data fusion on the multi-source insurance data provided in one embodiment of the present invention;

[0027] Figure 4 A schematic diagram of a module of an automated intelligent underwriting device provided by one embodiment of the present invention;

[0028] Figure 5 A schematic diagram of the structure of an electronic device for implementing an automated intelligent underwriting method provided by one embodiment of the present invention;

[0029] Figure 6 Another structural schematic diagram of an electronic device for implementing an automated intelligent underwriting method provided by one embodiment of the present invention.

[0030] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the technical solutions of the present disclosure, and to fully understand and implement how the present disclosure applies technical means to solve technical problems and achieve the corresponding technical effects, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. The embodiments of the present disclosure and the various features in the embodiments can be combined with each other without conflict, and the technical solutions formed are all within the scope of protection of the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present disclosure.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, apparatus, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] An embodiment of the present application provides an automated intelligent underwriting method, and the execution subject of the automated intelligent underwriting method includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the device provided by the embodiment of the present application. In other words, the automated intelligent underwriting method can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0034] The present invention provides an automated intelligent underwriting method that can be applied in Figure 1In this application environment, the client communicates with the server via a network. The server can access multi-source insurance data through the client. By performing a series of processing operations such as data fusion and data cleansing on this multi-source insurance data, it can identify and correct errors, omissions, and inconsistencies in the data, ensuring data accuracy and uniqueness, thereby significantly improving the quality of the target insurance data. Attribute recognition enables comprehensive and accurate collection and analysis of insurance data to more accurately assess insurance data risks, while also rapidly extracting key information from the target insurance data, shortening the underwriting process. Precisely matching object attributes with underwriting parameter models effectively avoids human error, ensures the accuracy of underwriting results, and reduces insurance companies' claims risk. It also allows for flexible adaptation to various complex situations and personalized underwriting. By assigning risk scores to the target insurance data, it eliminates fluctuations in underwriting standards caused by human experience, ensuring the objectivity and standardization of underwriting assessments, thereby improving underwriting efficiency and enhancing the accuracy and comprehensiveness of underwriting. Finally, the underwriting results are fed back to the client. The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server side can be implemented by an independent server or a server cluster composed of multiple servers. The present invention will be described in detail below through specific embodiments.

[0035] Reference Figure 2 FIG. 1 is a flow chart of an automated intelligent underwriting method according to an embodiment of the present invention. In this embodiment, the automated intelligent underwriting method includes:

[0036] S1. Obtain multi-source insurance data of a target insured person, perform multi-source data fusion on the multi-source insurance data, and obtain target insurance data.

[0037] In an embodiment of the present invention, the multi-source insurance data includes basic data, health status data, and financial status data of the target insured person, wherein the basic data includes but is not limited to: age, gender, and occupation; the health status data includes but is not limited to: physical examination report and past medical history; the financial status data includes but is not limited to: income certificate and asset status.

[0038] In an embodiment of the present invention, programming languages such as Python and Java can be used to connect to a preset database through the HTTP protocol, and an HTTP request can be sent to obtain multi-source insurance data of the target insured object, thereby obtaining the various insurance data.

[0039] In addition, the uniform resource locator of the database can be opened by a preset web crawler tool to crawl purposefully, thereby obtaining multi-source insurance data of the target insured person.

[0040] For example, in a medical and health scenario, the target insured person may be a patient, medical staff, etc., and the patient's medical records, diagnosis reports, treatment plans, etc. can be obtained through the medical information system. For example, the target insured person's electronic health record (EHR) can be obtained, including past medical history, surgical records, medication status, etc., so as to understand the patient's health status and treatment history.

[0041] Specifically, the patient's health data and hospital records can be combined to assess their health risks, provide a pricing basis for insurance companies, and provide patients with personalized insurance plans.

[0042] In the fintech business scenario, by cooperating with banks and payment institutions, we can obtain the target insured’s account flow, consumption records and other financial data to understand their financial status and payment ability. We can also obtain the target insured’s personal credit report and credit score through the credit reporting agency API interface, evaluate their repayment ability and credit status, design personalized premium plans for insurance companies, and recommend suitable insurance product combinations for them.

[0043] like Figure 3 As shown, in the embodiment of the present invention, the multi-source data fusion is performed on the multi-source insurance data to obtain the target insurance data, including:

[0044] Performing data cleaning on the multi-source insurance data to obtain standardized multi-source insurance data;

[0045] Extracting source-specific features from the standardized multi-source insurance data to obtain a structured feature vector;

[0046] Performing cross-modal spatiotemporal alignment on the structured feature vector to obtain target fusion features;

[0047] The target fusion features are subjected to deep semantic extraction according to the self-attention mechanism to obtain target insurance data.

[0048] In detail, the data cleaning includes operations such as missing value processing, outlier detection, data deduplication and format standardization. The missing value processing handles the gaps in the data by deleting, filling or predicting missing values. The data deduplication refers to identifying and deleting duplicate data records to ensure the uniqueness of the data. The format standardization refers to unifying the data format, such as the format of fields such as age and occupation, to facilitate data comparison and analysis.

[0049] Among them, the standardized multi-source insurance data includes a large amount of unstructured data and structured data, such as text, images, audio, etc. For the structured data, feature engineering such as discretization can be used for feature extraction. For the unstructured data, semantic vectors in the text can be extracted based on pre-trained language models such as BERT technology, or visual high-dimensional features in the image can be extracted through computer vision algorithms to generate corresponding structured feature vectors for different modal data.

[0050] In detail, the cross-modal spatiotemporal alignment includes time alignment and space alignment. The time alignment can use dynamic time warping (DTW) or time interpolation to unify multi-source timestamps to map structured feature vectors to a unified vector space; the spatial alignment can eliminate the modality gap through feature mapping networks such as adversarial domain adaptation DAE or graph attention (GAT) mechanism, thereby constructing a joint feature space of multi-source data.

[0051] Specifically, feature weights can be calculated based on a multi-head self-attention mechanism, where the multi-head mechanism captures spatial associations of different features in parallel, thereby mining deep semantic associations of multi-source modal features and generating highly reliable target insurance data.

[0052] For example, in the field of fintech business, insurance companies need to integrate data from different channels, such as hospital records, wearable devices, social media, financial institutions, etc., to fully understand the risk status and needs of the target insured persons. By integrating multi-source insurance data, the risk level of the target insured persons can be more accurately assessed. At the same time, the insurance industry involves a large amount of money transactions and complex contractual relationships. Any data errors may lead to huge economic losses and reputation damage. Through data cleaning, the risks caused by data errors can be effectively reduced.

[0053] Specifically, cleaned data is easier to process and analyze, enabling insurance companies to make decisions more quickly and accurately in formulating strategies, assessing risks, designing new products, etc., providing a basis for premium pricing, product design, etc.

[0054] In an embodiment of the present invention, by performing a series of processes such as data fusion and data cleaning on multi-source insurance data, errors, omissions and inconsistencies in the data can be identified and corrected, ensuring the accuracy and uniqueness of the data, thereby significantly improving the quality of the target insurance data.

[0055] S2. Perform attribute recognition on the target insurance data to obtain object attributes of the target insurance object.

[0056] In an embodiment of the present invention, the object attributes may refer to insurance type attributes such as property insurance, life insurance, liability insurance, etc., and may also refer to age attributes, health status attributes, etc. The insurance type attributes may specifically be critical illness insurance, medical insurance, accident insurance, financial insurance, etc. By obtaining the corresponding object attributes, the authenticity and rationality of the claim application can be more accurately evaluated in the subsequent underwriting process.

[0057] In the embodiment of the present invention, the performing of attribute identification on the target insurance data to obtain the object attributes of the target insurance object includes:

[0058] Dividing the target insurance data into regions to obtain structured regional data of the target insurance data;

[0059] Performing content detection on the structured region data to obtain detection text data;

[0060] The detected text data is matched with the attribute fields in the preset attribute library to obtain the object attributes of the target insured object.

[0061] In detail, computer vision algorithms (such as OpenCV) can be used to analyze the overall layout of the target insurance data to identify different areas in the target insurance data, including policyholder information, insurance subject description, insurance terms, premium calculation, etc.

[0062] Among them, the boundaries of each area are defined according to the results of the analysis, which can be achieved through manual setting or automatic detection. For example, the start and end positions of the area are determined by identifying specific keywords, punctuation marks or format features, and the defined area is extracted from the target insurance data to form structured area data, which includes basic information, health information, financial information and other data.

[0063] Specifically, the structured region data may be subjected to content detection through optical character recognition technology or PDF text stream parsing, and information data such as text and numbers therein may be extracted to form detection text data.

[0064] The present invention pre-constructs an attribute library containing various possible attributes, such as the insured person's name, the insured person's age, the name of the insurance subject, the insurance amount, the insurance period, etc. Each attribute corresponds to one or more attribute fields. The detected text data is matched with the attribute fields in the preset attribute library through string matching and regular expression matching, and the object attributes of the target insured object are extracted according to the matching results.

[0065] For example, in healthcare scenarios, target insurance data attribute recognition can be applied to the intelligent claims process for health insurance. Computer vision algorithms analyze the overall layout of medical documents (such as outpatient invoices, hospitalization invoices, expense lists, and discharge summaries) and identify different areas, such as patient information, medical information, and expense information. For example, when identifying a hospitalization invoice, the computer vision algorithm will first locate key areas such as the invoice title, patient name, hospitalization date, discharge date, and total cost.

[0066] Specifically, optical character recognition (OCR) is used to detect the content of structured region data and extract information such as text and numbers. For example, in a discharge summary, OCR can extract key information such as the patient's diagnosis, treatment plan, and discharge instructions.

[0067] Among them, based on the pre-construction of an attribute library containing the attributes required for health insurance claims, such as patient name, ID number, date of consultation, diagnosis results, treatment costs, etc., for example, the extracted patient name and ID number are matched with the "insured person's name" and "insured person's ID number" fields in the attribute library to obtain the corresponding object attributes.

[0068] For example, in FinTech business scenarios, the identification of target insurance data attributes can be applied to the provision of personalized financial services. Computer vision algorithms analyze the overall layout of financial documents (such as insurance contracts, bank statements, credit reports, etc.) and identify different areas, such as customer information, transaction records, and contract terms. For example, when identifying an insurance contract, the algorithm will first locate key areas such as the contract title, policyholder information, insurance subject, insurance amount, and insurance period.

[0069] Specifically, optical character recognition or natural language processing (NLP) technology is used to detect the content of structured data and extract information such as text and numbers. For example, in bank statements, OCR technology can extract key information such as transaction date, transaction amount, and counterparty.

[0070] A pre-built attribute library contains attributes required for financial services, such as customer name, ID number, bank account, credit score, spending habits, risk appetite, etc. For example, the extracted customer name, ID number, and credit score are matched with the "customer name," "customer ID number," and "customer credit score" fields in the attribute library to obtain the corresponding object attributes.

[0071] In the embodiment of the present invention, attribute recognition can be used to comprehensively and accurately collect and analyze insurance data to more accurately assess the risks of insurance data, while quickly extracting key information from the target insurance data, which can shorten the underwriting process and greatly improve underwriting efficiency.

[0072] S3. Obtain an underwriting parameter model for the target insurance data, and dynamically adjust the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes.

[0073] In an embodiment of the present invention, the underwriting parameter model can be a criterion model used by insurance companies to evaluate the risk level of insurance applications, decide whether to accept insurance, and determine premium rates. It usually covers factors such as the insured's personal information, health status, value of the insured object, and risk assessment, aiming to control risks and ensure that insurance companies can provide sustainable insurance services.

[0074] Among them, the underwriting parameter model involves multiple aspects, such as the underwriting standards and review requirements for different types of insurance, as well as special review requirements for applicants in high-risk occupations, with a history of chronic diseases or genetic disease risks.

[0075] Specifically, underwriting parameter models can be obtained through internal insurers, industry regulation, or through big data analysis. For example, underwriting parameter models can be developed by an insurance company's actuaries and underwriters based on factors such as the company's business strategy, market demand, and risk assessment results. For example, insurance industry regulators will develop some common underwriting parameter models and standards to ensure the compliance of insurance companies. For example, the "Insurance Law of the People's Republic of China" and related industry regulations stipulate underwriting and insurance, including the establishment of insurance contracts, the insurance company's right of verification, and the policyholder's obligation to disclose truthfully.

[0076] In an embodiment of the present invention, dynamically adjusting the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes includes:

[0077] Extracting parameter keywords from the underwriting parameter model to obtain parameter identification keywords;

[0078] Performing real-time matching calculation on the object attributes and the parameter identification keywords according to the real-time changing dimension of the object attributes to obtain a dynamic matching degree;

[0079] The dynamic matching degree is compared with a preset matching threshold to determine a target identification keyword corresponding to the object attribute, and an underwriting parameter model corresponding to the target identification keyword is selected as a target underwriting parameter model.

[0080] In detail, keyword extraction algorithms such as TF-IDF, TextRank, LDA and other algorithms can be used to extract keywords from the underwriting parameters in the underwriting parameter model. The parameter identification keywords include disease names, wealth levels, etc., and then the matching degree is calculated through string comparison method or Jaccard similarity coefficient. The string comparison method refers to directly comparing whether the object attributes and parameter identification keywords are completely or partially matched. The Jaccard similarity coefficient calculates the similarity between the object attributes and parameter identification keywords through set intersection and union.

[0081] In the present invention, the cosine similarity algorithm, Jaccard similarity algorithm, etc. can be used to calculate the similarity between object attributes and parameter identification keywords in various matching dimensions. For example, for age attributes, matching can be performed according to age groups, and for health status attributes, matching can be performed according to the type and severity of the disease.

[0082] Specifically, the similarities are weighted and summed according to the importance weights of each dimension to obtain the final dynamic matching degree. For example, assuming that the weight of the age dimension is 0.3, the weight of the health status dimension is 0.7, the similarity of the object attributes in the age dimension is 0.8, and the similarity in the health status dimension is 0.6, then the dynamic matching degree is 0.3*0.8+0.7*0.6=0.66.

[0083] Specifically, a comparison is made between the predefined matching degree threshold and the dynamic matching degree. If the matching degree exceeds the matching degree threshold, it is considered that the parameter identification keyword matches the object attribute, and the parameter identification keyword with a matching degree exceeding the threshold is screened out as the target identification keyword, and the underwriting parameter model corresponding to the target identification keyword is determined as the target underwriting parameter model.

[0084] In the embodiment of the present invention, extracting parameter keywords from the underwriting parameter model to obtain parameter identification keywords includes:

[0085] selecting one of the multiple underwriting parameter models as the underwriting parameter model to be processed, and performing word segmentation on the underwriting parameters in the underwriting parameter model to be processed to obtain word segmentation of the underwriting parameter model;

[0086] Select one of the underwriting parameter model segmentations one by one as the segmentation to be processed;

[0087] Counting the number of local occurrences of the to-be-processed segmentation word in the to-be-processed underwriting parameter model, and counting the number of global occurrences of the to-be-processed segmentation word in all the underwriting parameter models;

[0088] Dividing the local occurrence count by the global occurrence count to obtain the criticality of the word to be processed;

[0089] The to-be-processed word set with a criticality greater than a preset threshold is used as a parameter identification keyword of the underwriting parameter model.

[0090] For example, in the healthcare field, underwriting parameter models typically involve parameters such as disease names and health conditions, which are used to assess the health risks of target policyholders. TF-IDF and other algorithms are used to extract parameter keywords from the underwriting parameter model. For example, parameter keywords such as "diabetes," "blood sugar levels," and "complications" are extracted from the underwriting parameter model for diabetic patients. The matching degree between the object attributes and the parameter identification keywords is calculated using string comparison or the Jaccard similarity coefficient. For example, if the object attribute is "diabetic patient, high blood sugar level, and mild complications," then the matching degree with the extracted keywords "diabetes," "blood sugar level," and "complications" is high.

[0091] The dynamic matching degree is compared with a preset matching threshold. If the matching degree exceeds the threshold, the parameter identification keyword is considered to match the object attribute. For example, if the preset matching threshold is 80% and the matching degrees exceed the threshold, the underwriting parameter model is determined to be the target underwriting parameter model.

[0092] In the embodiment of the present invention, by accurately matching object attributes with the underwriting parameter model, human errors are effectively avoided, the accuracy of the underwriting results is ensured, and the insurance company's claim risk is reduced. At the same time, it can flexibly adapt to various complex situations and realize personalized underwriting.

[0093] S4. Underwrite the target insurance data according to the target underwriting parameter model to obtain an underwriting result of the target insurance data.

[0094] In an embodiment of the present invention, a risk score is performed on the target insurance data according to a target underwriting parameter model, and an underwriting result of the target insurance data is determined based on the scored underwriting risk data.

[0095] In an embodiment of the present invention, the step of underwriting the target insurance data according to the target underwriting parameter model to obtain an underwriting result of the target insurance data includes:

[0096] Performing word screening on the target insurance data to obtain insurance risk words;

[0097] Performing risk scoring on the insurance risk terms according to the target underwriting parameter model to obtain underwriting risk score data;

[0098] The underwriting result of the target insurance data is determined based on the underwriting risk score data.

[0099] In detail, the risk-related vocabulary library and natural language processing algorithm can be used to screen words. The risk-related words include disease names, dangerous occupations, high-risk activities, etc. The target insurance data is matched with the risk-related vocabulary library, and the successfully matched words are used as insurance risk words.

[0100] Specifically, the screened risk words are evaluated according to the target underwriting parameter model. Different risk words may have different weights or scores. Each insured risk word is scored using the scoring criteria in the underwriting parameter library in the underwriting parameter model. For example, if a risk word is marked as high risk in the underwriting parameter library, it is given a higher score; if it is marked as low risk, it is given a lower score.

[0101] Among them, the insurance data is comprehensively evaluated based on the underwriting risk score data. The underwriting result usually includes the presence or absence of underwriting risk, such as approval or rejection of the insurance. The target underwriting parameter model can be used for underwriting.

[0102] In an embodiment of the present invention, the underwriting results can be displayed on a visual interface. The visual interface display is an interface display format for communication between people and computers, allowing users to use input devices such as a mouse to manipulate icons or menu options on the screen to select commands, call files, start programs or perform other daily tasks, so as to improve the efficiency of obtaining underwriting results.

[0103] In detail, in the visualization interface, there are usually multiple chart types preset, such as bar charts, line charts, pie charts, etc. and corresponding chart configuration items, such as axis labels, legends, titles, data formats, etc. The underwriting results are bound to the chart configuration items, that is, the binding is achieved by modifying the properties in the configuration item object, and then the underwriting results are mapped to the coordinate axis of the chart, the chart title, coordinate axis labels are set, the color, font size, etc., and the underwriting results are displayed in the visualization chart in an intuitive and easy-to-understand manner.

[0104] In an embodiment of the present invention, the transparency and explainability of the underwriting results can be ensured through a visual interface display, avoiding the black box problem. At the same time, the present invention can also use AI technology to train historical multi-source insurance data to identify abnormal or complex cases, so that relevant underwriting results can be automatically obtained during the underwriting process, avoiding misjudgment or failure to underwrite, thereby improving underwriting efficiency.

[0105] For example, in medical and health scenarios, it can be applied to automated underwriting of health insurance applications. The BERT model can be used to identify the semantic correlation between medical records and risk words in the dictionary to obtain risk words, and then determine the underwriting risk score data based on the weight score of the risk words.

[0106] In an embodiment of the present invention, determining the underwriting result of the target insurance data according to the underwriting risk score data includes:

[0107] Comparing the underwriting risk score data with a preset risk threshold to obtain a comparison result;

[0108] If the comparison result is that the underwriting risk score data is greater than the risk threshold, the underwriting result is determined as the target insurance data having an underwriting risk;

[0109] If the comparison result is that the underwriting risk score data is less than or equal to the risk threshold, the underwriting result is determined as there is no underwriting risk for the target insurance data.

[0110] In an embodiment of the present invention, by performing risk screening on the target insurance data, the risk data in the insurance data can be quickly processed, and at the same time, fluctuations in underwriting standards caused by differences in human experience can be eliminated, thereby ensuring the objectivity and standardization of underwriting assessments, thereby improving underwriting efficiency, and also improving the accuracy and comprehensiveness of underwriting.

[0111] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean 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 the present invention.

[0112] like Figure 4 , which is a functional module diagram of an automated intelligent underwriting device provided by one embodiment of the present invention.

[0113] In an embodiment of the present disclosure, an automated intelligent underwriting device is provided, and the automated intelligent underwriting device corresponds one-to-one with the automated intelligent underwriting method in the above embodiment. Figure 5 As shown, the automated intelligent underwriting device 100 can be installed in an electronic device. According to the functions to be implemented, the automated intelligent underwriting device 100 includes a data fusion module 101, an attribute recognition module 102, a dynamic adjustment module 103, and a real-time underwriting module 104. The functional modules are described in detail as follows:

[0114] The data fusion module 101 is used to obtain multi-source insurance data of a target insured person and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data;

[0115] The attribute recognition module 102 is used to perform attribute recognition on the target insurance data to obtain the object attributes of the target insurance object;

[0116] A dynamic adjustment module 103 is configured to obtain an underwriting parameter model of the target insurance data, and dynamically adjust the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes;

[0117] The real-time underwriting module 104 is configured to underwrite the target insurance data according to the target underwriting parameter model to obtain an underwriting result of the target insurance data.

[0118] In one embodiment, when the data fusion module 101 performs multi-source data fusion on the multi-source insurance data to obtain target insurance data, it is configured to:

[0119] Performing data cleaning on the multi-source insurance data to obtain standardized multi-source insurance data;

[0120] Extracting source-specific features from the standardized multi-source insurance data to obtain a structured feature vector;

[0121] Performing cross-modal spatiotemporal alignment on the structured feature vector to obtain target fusion features;

[0122] The target fusion features are subjected to deep semantic extraction according to the self-attention mechanism to obtain target insurance data.

[0123] In one embodiment, when performing attribute recognition on the target insurance data to obtain the object attributes of the target insurance object, the attribute recognition module 102 is configured to:

[0124] Dividing the target insurance data into regions to obtain structured regional data of the target insurance data;

[0125] Performing content detection on the structured region data to obtain detection text data;

[0126] The detected text data is matched with the attribute fields in the preset attribute library to obtain the object attributes of the target insured object.

[0127] In one embodiment, when the dynamic adjustment module 103 dynamically adjusts the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes, it is configured to:

[0128] Extracting parameter keywords from the underwriting parameter model to obtain parameter identification keywords;

[0129] Performing real-time matching calculation on the object attributes and the parameter identification keywords according to the real-time changing dimension of the object attributes to obtain a dynamic matching degree;

[0130] The dynamic matching degree is compared with a preset matching threshold to determine a target identification keyword corresponding to the object attribute, and an underwriting parameter model corresponding to the target identification keyword is selected as a target underwriting parameter model.

[0131] In one embodiment, when the dynamic adjustment module 103 extracts parameter keywords from the underwriting parameter model to obtain parameter identification keywords, it is configured to:

[0132] selecting one of the multiple underwriting parameter models as the underwriting parameter model to be processed, and performing word segmentation on the underwriting parameters in the underwriting parameter model to be processed to obtain word segmentation of the underwriting parameter model;

[0133] Select one of the underwriting parameter model segmentations one by one as the segmentation to be processed;

[0134] Counting the number of local occurrences of the to-be-processed segmentation word in the to-be-processed underwriting parameter model, and counting the number of global occurrences of the to-be-processed segmentation word in all the underwriting parameter models;

[0135] Dividing the local occurrence count by the global occurrence count to obtain the criticality of the word to be processed;

[0136] The to-be-processed word set with a criticality greater than a preset threshold is used as a parameter identification keyword of the underwriting parameter model.

[0137] In one embodiment, when the real-time underwriting module 104 performs underwriting on the target insurance data according to the target underwriting parameter model and obtains the underwriting result of the target insurance data, it is configured to:

[0138] Performing word screening on the target insurance data to obtain insurance risk words;

[0139] Performing risk scoring on the insurance risk terms according to the target underwriting parameter model to obtain underwriting risk score data;

[0140] The underwriting result of the target insurance data is determined based on the underwriting risk score data.

[0141] In one embodiment, when determining the underwriting result of the target insurance data based on the underwriting risk score data, the real-time underwriting module 104 is configured to:

[0142] Comparing the underwriting risk score data with a preset risk threshold to obtain a comparison result;

[0143] If the comparison result is that the underwriting risk score data is greater than the risk threshold, the underwriting result is determined as the target insurance data having an underwriting risk;

[0144] If the comparison result is that the underwriting risk score data is less than or equal to the risk threshold, the underwriting result is determined as there is no underwriting risk for the target insurance data.

[0145] In the present invention, the specific definition of an automated intelligent underwriting device can be found in the definition of an automated intelligent underwriting method described above and will not be repeated here. Each module in the aforementioned automated intelligent underwriting device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of the aforementioned modules can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a memory in the computer device in software form, so that the processor can call and execute the corresponding operations of each of the aforementioned modules.

[0146] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the service side of an automated intelligent underwriting method.

[0147] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, memory, a network interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When executed by the processor, the computer program implements the functions or steps on the client side of an automated intelligent underwriting method.

[0148] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed:

[0149] Acquire multi-source insurance data of a target insured person, and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data;

[0150] Performing attribute recognition on the target insurance data to obtain object attributes of the target insurance object;

[0151] Obtaining an underwriting parameter model for the target insurance data, and dynamically adjusting the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes;

[0152] The target insurance data is underwritten according to the target underwriting parameter model to obtain an underwriting result of the target insurance data.

[0153] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and apparatuses can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the module division is merely a logical function division, and actual implementation may employ other division methods.

[0154] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0155] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0156] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0157] In some implementations of this embodiment, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the method described in the above embodiment are implemented.

[0158] The readable storage medium of the present invention stores a computer program, which, when executed by a processor of an electronic device, can implement:

[0159] Acquire multi-source insurance data of a target insured person, and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data;

[0160] Performing attribute recognition on the target insurance data to obtain object attributes of the target insurance object;

[0161] Obtaining an underwriting parameter model for the target insurance data, and dynamically adjusting the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes;

[0162] The target insurance data is underwritten according to the target underwriting parameter model to obtain an underwriting result of the target insurance data.

[0163] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can be found in the relevant descriptions of the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0164] The computer-readable storage medium may also store at least one computer-executable program / instruction, such as a computer-readable instruction. Computer-readable storage media include, but are not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Computer-readable storage media may include, for example, read-only memory (ROM), a hard disk, a flash memory, etc. For example, a non-transitory computer-readable storage medium may be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions stored on the computer-readable storage medium, the various methods described above may be performed.

[0165] In addition, the computer device may also include (but is not limited to) a data bus, an input / output (I / O) bus, a display, and input / output devices (eg, keyboard, mouse, speaker, etc.).

[0166] The processor can communicate with external devices via an I / O bus via a wired or wireless network.

[0167] In one embodiment, the at least one computer executable instruction may also be compiled into or constitute a software product / computer program product, wherein one or more computer executable instructions are executed by a processor to perform the various functions and / or method steps in the embodiments described in the present technology.

[0168] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. 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-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0169] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by 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.

[0170] In the embodiments provided in the present disclosure, it should be understood that the disclosed devices and methods may also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, may be implemented with a dedicated hardware-based system that performs the specified function or action, or may be implemented with a combination of dedicated hardware and computer instructions.

[0171] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

[0172] It should be noted that if software tools or components other than those of our company appear in the embodiments of this application, they are only used for illustration and do not represent actual use.

Claims

1. An automated intelligent underwriting method, characterized in that: The method comprises: Acquire multi-source insurance data of a target insured person, and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data; Performing attribute recognition on the target insurance data to obtain object attributes of the target insurance object; Obtaining an underwriting parameter model for the target insurance data, and dynamically adjusting the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes; The target insurance data is underwritten according to the target underwriting parameter model to obtain an underwriting result of the target insurance data.

2. The automated intelligent underwriting method according to claim 1, wherein: The multi-source data fusion is performed on the multi-source insurance data to obtain target insurance data, including: Performing data cleaning on the multi-source insurance data to remove duplicate data, fill missing values, and correct outliers to obtain standardized multi-source insurance data; Extracting source-specific features from the standardized multi-source insurance data to obtain a structured feature vector; Performing cross-modal spatiotemporal alignment on the structured feature vector to obtain target fusion features; The target fusion features are subjected to deep semantic extraction according to the self-attention mechanism to obtain target insurance data.

3. The automated intelligent underwriting method according to claim 1, wherein: The performing attribute identification on the target insurance data to obtain the object attributes of the target insurance object includes: Dividing the target insurance data into regions to obtain structured regional data of the target insurance data; Performing content detection on the structured region data to obtain detection text data; The detected text data is matched with the attribute fields in the preset attribute library to obtain the object attributes of the target insured object.

4. The automated intelligent underwriting method according to claim 1, wherein: The dynamically adjusting the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes includes: Extracting parameter keywords from the underwriting parameter model to obtain parameter identification keywords; Performing real-time matching calculation on the object attributes and the parameter identification keywords according to the real-time changing dimension of the object attributes to obtain a dynamic matching degree; The dynamic matching degree is compared with a preset matching threshold to determine a target identification keyword corresponding to the object attribute, and an underwriting parameter model corresponding to the target identification keyword is selected as a target underwriting parameter model.

5. The automated intelligent underwriting method according to claim 4, wherein: The step of extracting parameter keywords from the underwriting parameter model to obtain parameter identification keywords includes: selecting one of the multiple underwriting parameter models as the underwriting parameter model to be processed, and performing word segmentation on the underwriting parameters in the underwriting parameter model to be processed to obtain word segmentation of the underwriting parameter model; Select one of the underwriting parameter model segmentations one by one as the segmentation to be processed; Counting the number of local occurrences of the to-be-processed segmentation word in the to-be-processed underwriting parameter model, and counting the number of global occurrences of the to-be-processed segmentation word in all the underwriting parameter models; Dividing the local occurrence count by the global occurrence count to obtain the criticality of the word to be processed; The to-be-processed word set with a criticality greater than a preset threshold is used as a parameter identification keyword of the underwriting parameter model.

6. The automated intelligent underwriting method according to claim 1, wherein: The step of underwriting the target insurance data according to the target underwriting parameter model to obtain an underwriting result of the target insurance data includes: Performing word screening on the target insurance data to obtain insurance risk words; Performing risk scoring on the insurance risk terms according to the target underwriting parameter model to obtain underwriting risk score data; The underwriting result of the target insurance data is determined based on the underwriting risk score data.

7. The automated intelligent underwriting method according to claim 6, wherein: Determining the underwriting result of the target insurance data according to the underwriting risk score data includes: Comparing the underwriting risk score data with a preset risk threshold to obtain a comparison result; If the comparison result is that the underwriting risk score data is greater than the risk threshold, the underwriting result is determined as the target insurance data having an underwriting risk; If the comparison result is that the underwriting risk score data is less than or equal to the risk threshold, the underwriting result is determined as there is no underwriting risk for the target insurance data.

8. An automated intelligent underwriting device, characterized in that: The device comprises: A data fusion module is used to obtain multi-source insurance data of a target insured person, and perform multi-source data fusion on the multi-source insurance data to obtain target insurance data; An attribute recognition module, configured to perform attribute recognition on the target insurance data to obtain the object attributes of the target insurance object; a dynamic adjustment module, configured to obtain an underwriting parameter model of the target insurance data, and dynamically adjust the underwriting parameter model according to the object attributes to obtain a target underwriting parameter model that matches the object attributes; The real-time underwriting module is used to underwrite the target insurance data according to the target underwriting parameter model to obtain the underwriting result of the target insurance data.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute an automated intelligent underwriting method 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, an automated intelligent underwriting method as described in any one of claims 1 to 7 is implemented.

Images