Insurance product recommendation method and device, computer device, and readable storage medium
By generating and supplementing colloquial terms for professions and industries, and combining this with a deep learning model to identify users' professions, the issues of data security and accuracy in insurance product recommendations have been resolved, enabling precise insurance product recommendations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA PING AN PROPERTY INSURANCE CO LTD
- Filing Date
- 2024-11-22
- Publication Date
- 2026-06-12
AI Technical Summary
Existing technologies struggle to accurately obtain users' occupation and industry information when recommending insurance products, leading to inaccurate recommendations and compromising the security of corporate data.
By acquiring occupational classification standard documents and internal occupational information, the system generates colloquial terms for occupations and industries, establishes mapping relationships, supplements internal occupational information, uses deep learning generative models to identify users' occupations, and recommends insurance products based on their occupations.
It enables the accurate acquisition of users' occupation and industry information while ensuring data security, providing more precise insurance product recommendations and improving data quality and recommendation accuracy.
Smart Images
Figure CN119648326B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of data processing and financial technology, and in particular to a method, apparatus, computer equipment, and computer-readable storage medium for recommending insurance products. Background Technology
[0002] When providing insurance product recommendation services, insurance companies need to obtain the precise occupation and industry type of individual users to assess insurance risks and subsequent insurance pricing. For specific areas like "user occupations used for insurance risk assessment," current open-source models often lack the ability to respond effectively in vertical fields, and their output frequently falls outside the "Occupational Classification Directory."
[0003] Regarding occupational data, different insurance companies will have different and personalized interpretations of policy documents. Due to the confidentiality of corporate data, it cannot be released to external large-scale models for corpus generation. If open-source models are used internally to reverse engineer questions and answers, the data quality will be significantly lower than that of top-tier external large-scale models.
[0004] Therefore, how to accurately obtain users' occupation and industry information while ensuring data security, so as to provide more accurate insurance product recommendation services, has become a technical problem that the insurance industry urgently needs to solve. Summary of the Invention
[0005] The main purpose of this application is to provide an insurance product recommendation method, device, computer equipment, and computer-readable storage medium, which aims to accurately obtain users' occupation and industry information while ensuring data security, thereby providing more accurate insurance product recommendation services.
[0006] To achieve the above objectives, this application provides a method for recommending insurance products, which includes the following steps:
[0007] Obtain occupational classification standard documents and internal occupational information, and generate colloquial terms for occupations and industry terms based on the occupational classification standard documents;
[0008] The internal occupational information is supplemented based on the colloquial terms for the occupations and the colloquial terms for the industries to obtain an internal occupational standard document;
[0009] Acquire dialogue information between the user and the claims adjuster, and identify the user's occupation from the dialogue information based on the internal occupational standards document;
[0010] Recommend insurance products to the user based on their occupation.
[0011] In addition, to achieve the above objectives, this application also provides an insurance product recommendation device, the insurance product recommendation device comprising:
[0012] The colloquial term generation unit is used to acquire occupational classification standard documents and internal occupational information, and to generate colloquial terms for occupations and industry colloquial terms based on the occupational classification standard documents.
[0013] The occupational information supplementation unit is used to supplement the internal occupational information based on the colloquial terms for the occupations and the colloquial terms for the industry, thereby obtaining an internal occupational standard document;
[0014] The user occupation identification unit is used to acquire dialogue information between the user and the claims adjuster, and to identify the user's occupation from the dialogue information according to the internal occupational standard document.
[0015] An insurance product recommendation unit is used to recommend insurance products to the user based on the user's occupation. Furthermore, to achieve the above objectives, this application also provides a computer device, which includes a processor, a memory, and an insurance product recommendation program stored in the memory and executable by the processor, wherein when the insurance product recommendation program is executed by the processor, it implements the steps of the insurance product recommendation method described above.
[0016] In addition, to achieve the above objectives, this application also provides a computer-readable storage medium storing an insurance product recommendation program, wherein when the insurance product recommendation program is executed by a processor, it implements the steps of the insurance product recommendation method as described above.
[0017] This application provides a method, apparatus, computer device, and computer-readable storage medium for recommending insurance products. The method includes acquiring an occupational classification standard document and internal occupational information, and generating colloquial terms for occupations and industry terms based on the occupational classification standard document; supplementing the internal occupational information with the colloquial terms for occupations and industry terms to obtain an internal occupational standard document; then acquiring dialogue information between a user and a claims adjuster, and identifying the user's occupation from the dialogue information based on the internal occupational standard document; and finally recommending insurance products to the user based on their occupation.
[0018] As can be seen from the embodiments of this application, the colloquial terms for occupations and industries can be generated by an external large model based on the occupational classification standard file, resulting in higher data quality compared to generating them using an internal open-source model. Then, the internal occupational information is supplemented based on these colloquial terms, resulting in a richer and more complete internal occupational standard file. Therefore, identifying the user's occupation from the dialogue information based on this internal occupational standard file is more accurate, leading to more precise insurance product recommendations. Furthermore, the internal occupational information is not used as a pre-training parameter for the external large model, ensuring the security of the internal data. Attached Figure Description
[0019] Figure 1 A flowchart illustrating an insurance product recommendation method provided in this application;
[0020] Figure 2 A flowchart illustrating another method for recommending insurance products provided in this application;
[0021] Figure 3 A flowchart illustrating yet another insurance product recommendation method provided for this application;
[0022] Figure 4 A functional module diagram of an insurance product recommendation device provided in this application;
[0023] Figure 5 A schematic block diagram of the structure of a computer device provided in this application.
[0024] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0026] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.
[0027] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.
[0028] 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.
[0029] The insurance product recommendation method involved in the embodiments of this application is mainly applied to computer equipment, which can be a PC, a laptop computer, a mobile terminal, or other device with display and processing functions.
[0030] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.
[0031] Reference Figure 1 , Figure 1 This is a flowchart illustrating an insurance product recommendation method provided in this application.
[0032] like Figure 1 As shown in the figure, this application embodiment provides an insurance product recommendation method, which includes steps S101 to S104.
[0033] In this embodiment, the insurance product recommendation method includes the following steps:
[0034] Step S101: Obtain occupational classification standard documents and internal occupational information, and generate colloquial terms for occupations and colloquial terms for industries based on the occupational classification standard documents.
[0035] Occupational classification standard documents refer to standard documents developed by national or international organizations to regulate and unify occupational classifications. These documents are of great significance for human resource management, education planning, vocational training, and statistical analysis. In China, the most commonly used occupational classification standard document is the "Classification of Occupations of the People's Republic of China" (hereinafter referred to as the "Occupational Classification Classification"). Internal occupational information refers to information generated through personalized interpretation of occupational classification standard documents. Colloquial terms refer to words or expressions used in daily communication. Different cultural backgrounds and social environments may have different terminology habits. Occupational colloquial terms are words or expressions used to refer to an occupation in daily communication, while industry colloquial terms are words or expressions used to refer to an industry in daily communication.
[0036] Specifically, the step of generating colloquial terms for occupations and colloquial terms for industries based on the occupational classification standard document includes:
[0037] Extract the occupational descriptions from the occupational classification standard file;
[0038] The occupational descriptions are used as prompts for a deep learning generative model, and colloquial terms for the occupations and industries are generated based on the deep learning generative model.
[0039] For example, job descriptions in occupational classification standard documents could be such as "Landscaper: responsible for the maintenance of greenery within a property" or "Web Editor: editing and processing website articles and other content." These job descriptions can then be extracted and used as prompts for a deep learning generative model.
[0040] For example, the prompts for deep learning generative models could be:
[0041] 1. "Based on the following description of job tasks, list some names of people who perform this type of job in daily life. {Job content from a job classification dictionary}."
[0042] 2. The following is a scope description of the occupation "xxx". List the colloquial sub-occupations, following the format of "Landscape Worker: Responsible for the maintenance of greenery within the property" and "Web Editor: Editing and processing website articles and other content". Please include all sub-occupations.
[0043] For example, Chinese chef, head chef, cook, dim sum chef, and head chef are all colloquial terms for Chinese chef; restaurants, eateries, and snack shops belong to the accommodation and catering industry - catering industry - full-service catering.
[0044] For example, deep learning generative models could be GPT (Generative Pre-trained Transformer), Wenxin Yiyan, Tongyi Qianwen, Claude, NewBing, etc., without limitation.
[0045] It should be noted that the steps "obtain occupational classification standard document" and "obtain internal occupational information" can be obtained synchronously or asynchronously, and the execution order of the two is not limited.
[0046] Step S102: Supplement the internal occupational information according to the colloquial terms for the occupations and the colloquial terms for the industries to obtain an internal occupational standard document.
[0047] Specifically, the process of supplementing the internal occupational information based on the colloquial terms for the occupations and the colloquial terms for the industries to obtain an internal occupational standard document includes:
[0048] Obtain the job title from the internal job information;
[0049] Establish a mapping relationship between the occupational name and the colloquial terms for the occupation and the colloquial terms for the industry;
[0050] The internal occupational information is supplemented based on the mapping relationship to obtain an internal occupational standard document.
[0051] For example, the occupational classification standard document stores relevant content in tabular form. For instance, the table includes the following four columns: occupational code, occupational name, occupational category, and occupational description. By extracting information such as occupational description or occupational name as prompt words for the external large model, the colloquial terms for the occupation and the colloquial terms for the industry can be obtained. Then, the above table is expanded to six columns by adding two more columns: occupational code, occupational name, occupational category, occupational description, colloquial terms for the occupation, and colloquial terms for the industry.
[0052] For example, the internal occupational information is also stored in the form of a table, which includes the following 5 columns: occupation code, occupation name, occupation category, occupation description, and risk level.
[0053] The next step is to identify the job titles in the table and establish a mapping relationship between job titles and their colloquial terms and industry terms. This mapping relationship can then be used to supplement internal job information, resulting in an internal job standard document. After merging the two tables, the internal job standard document includes seven columns: job code, job title, job category, job description, colloquial term for the job, colloquial term for the industry, and risk level.
[0054] For example, industry types can be stored in a multi-level classification manner, such as: Agriculture, Forestry, Animal Husbandry, and Fishery -> Agriculture -> Grain and Other Crop Cultivation. Companies that might work under the Grain and Other Crop Cultivation type could include: agricultural companies, farms and plantations, greenhouse vegetable suppliers, fruit seedling cultivation companies, grain planting companies, grain farms, fruit and vegetable planting companies, seed breeding companies, agricultural product research and development companies, crop seed companies, agricultural cooperatives, agricultural technology companies, agricultural biotechnology companies, agricultural supply chain management companies, etc. Then, using the above content as prompts, colloquial terms for the industry can be generated.
[0055] Step S103: Obtain the dialogue information between the user and the claims adjuster, and identify the user's occupation from the dialogue information according to the internal occupational standard document.
[0056] In one specific embodiment of this application, identifying the user's occupation from the dialogue information based on the internal occupational standard document includes:
[0057] Convert the dialogue information into dialogue text;
[0058] The dialogue information text is segmented to obtain at least one text keyword;
[0059] Extract multiple candidate occupations from the internal occupational standard document based on at least one of the aforementioned text keywords;
[0060] Calculate the similarity between each of the candidate occupations and at least one of the text keywords;
[0061] Scores for multiple candidate occupations are calculated based on the similarity.
[0062] The user's occupation is determined based on the scores of multiple candidate occupations.
[0063] For example, the dialogue information may include the conversation between the user and the insurance salesperson, as well as emotional information during the conversation and environmental information during the conversation.
[0064] For example, if the conversation is text, the conversation text can be obtained directly; if the conversation is voice, the voice can be converted into conversation text using speech-to-text technology.
[0065] If the dialogue information includes emotional information from the conversation, this emotional information can be used to further recommend insurance products to the user. For example, if the emotional information is negative, the risk of future default is judged to be high, so the user's risk level can be increased, and the second method described below can be used to recommend insurance products. If the emotional information is positive, the risk of future default is judged to be low, so the user's risk level can be decreased, and the first method described below can be used to recommend insurance products.
[0066] For example, if the dialogue information includes environmental information about the conversation, this information can aid in determining the user's occupation and industry. For instance, if the keywords in the current conversation are related to fruit and vegetable cultivation, the user's occupation could be either a farmer or an agricultural researcher. The environmental information from the conversation can further assist in this determination. If the environmental information confirms high noise levels, the user is likely a farmer working in the fields; if the environmental information confirms low noise levels, the user is likely an agricultural researcher cultivating crops in a laboratory. The two occupations are significantly different, resulting in vastly different risk levels and therefore significantly different recommended insurance products. Furthermore, the noise level can be determined based on a set decibel threshold.
[0067] Word segmentation is a crucial step in natural language processing (NLP). It divides a continuous text into individual lexical units, often referred to as "words" or "lexical units." For example, word segmentation can be rule-based, statistical, or deep learning-based. Rule-based segmentation uses predefined dictionaries and rules to segment text; statistical segmentation trains models with large corpora and uses statistical probabilities to determine word boundaries; and deep learning-based segmentation utilizes neural network models, such as recurrent neural networks (RNNs), long short-term memory networks (LSTMs), and Transformers, for word segmentation.
[0068] Specifically, multiple candidate occupations are extracted from the internal occupational standard document based on at least one of the text keywords. For example, the at least one text keyword can be retrieved from the internal occupational standard document, and then the occupations associated with the retrieved keyword can be used as candidate occupations.
[0069] Similarity is used to measure the degree of similarity between two texts. For example, the similarity in this application includes at least one of the following: cosine similarity, Jaccard similarity, edit distance (Levenshtein distance), and word embedding similarity. Cosine similarity measures the similarity of two vectors by calculating the cosine of the angle between them. Jaccard similarity measures the similarity of two sets by calculating the ratio of their intersection to their union. Edit distance measures the similarity by calculating the minimum number of editing operations (insertion, deletion, replacement) required to convert one string into another. Word embeddings (such as Word2Vec, GloVe, BERT, etc.) map words to a high-dimensional vector space and measure the similarity of words or sentences by calculating the distance between vectors (such as Euclidean distance or cosine similarity).
[0070] For example, the score can be represented by similarity. For instance, if the similarity is 98%, the score is 98; if the similarity is 80%, the score is 80; and if the similarity is 34%, the score is 34.
[0071] Specifically, the internal occupational standard document includes the spoken occupational name for each occupation, the standard occupational name corresponding to the spoken occupational name, and the industry to which it belongs; the calculation of the similarity between each candidate occupation and the keyword includes:
[0072] Each candidate occupation is segmented into words to obtain the first spoken occupational name keyword and the first industry keyword.
[0073] Determine the second spoken occupational name keyword and the second industry keyword from at least one text keyword;
[0074] Calculate the name similarity between the first spoken occupational name keyword and the second spoken occupational name keyword, and calculate the industry similarity between the first industry keyword and the second industry keyword;
[0075] The similarity between each candidate occupation and the keyword is determined based on the name similarity and the industry similarity.
[0076] For example, determining the similarity between each candidate occupation and the keyword based on the name similarity and the industry similarity can be achieved by assigning corresponding weights to both name similarity and industry similarity, then calculating a first product of name similarity and its weight, a second product of industry similarity and its weight, and finally using the sum of the first and second products as the similarity between each candidate occupation and the keyword. Further, since the similarity between each candidate occupation and the keyword is determined, the weight of name similarity can be set to be greater than the weight of industry similarity, while ensuring that their sum is 1.
[0077] Step S104: Recommend insurance products to the user based on the user's occupation.
[0078] In one specific embodiment of this application, recommending insurance products to the user based on the user's occupation includes:
[0079] Determine the target risk level corresponding to the user's occupation, and recommend insurance products with a risk level higher than or equal to the target risk level for the user.
[0080] For example, after merging the two tables above, the internal occupational standard document includes seven columns: occupational code, occupational name, occupational category, occupational description, colloquial terminology for the occupation, colloquial terminology for the industry, and risk level. After identifying the user's occupation, the occupation column in this internal occupational standard document is searched to find the corresponding risk level as the target risk level. Then, insurance products with a risk level higher than or equal to the target risk level are recommended to the user. For example, if candidate occupations are A, B, and C, with risk levels of 1, 2, and 3 (risk level 1 < 2 < 3), and the identified occupation is B, then the recommended product would be 2 or 3.
[0081] As can be seen, the embodiments of this application can recommend a wider range of insurance products for users to choose from while reducing the risk to insurance companies. However, it first needs to determine the occupation from multiple candidate occupations, thus requiring the calculation of multiple similarities, making the recommendation speed relatively slow.
[0082] Furthermore, the internal occupational standard document includes a risk level for each occupation. Before determining the target risk level corresponding to the user's occupation, the process of determining the user's occupation based on the scores of multiple candidate occupations includes:
[0083] The candidate occupation with the highest score is determined as the user's occupation;
[0084] In another specific embodiment of this application, the insurance product recommendation method further includes:
[0085] Determine the target risk level for each of the candidate occupations to obtain multiple target risk levels;
[0086] Recommend insurance products with a risk level higher than or equal to the highest target risk level to the user.
[0087] For example, candidate occupations are A, B, and C, with risk levels of 1, 2, and 3; the highest level is 3. To reduce the risk for the insurance company, product 3 is recommended. It can be seen that in this embodiment, the highest risk level is determined directly based on multiple candidate occupations, and then products with a risk level higher than or equal to the highest risk level are recommended. This method may recommend fewer insurance products to the user, but it does not require calculating multiple similarities, thus the calculation speed is relatively fast.
[0088] As can be seen from the embodiments of this application, the colloquial terms for occupations and industries can be generated by an external large model based on the occupational classification standard file, resulting in higher data quality compared to generating them using an internal open-source model. Then, the internal occupational information is supplemented based on these colloquial terms, resulting in a richer and more complete internal occupational standard file. Therefore, identifying the user's occupation from the dialogue information based on this internal occupational standard file is more accurate, leading to more precise insurance product recommendations. Furthermore, the internal occupational information is not used as a pre-training parameter for the external large model, ensuring the security of the internal data.
[0089] Reference Figure 2 , Figure 2 A flowchart illustrating another method for recommending insurance products provided in this application.
[0090] like Figure 2 As shown in the figure, this application embodiment provides an insurance product recommendation method, which includes steps S201 to S217.
[0091] In this embodiment, the insurance product recommendation method includes the following steps:
[0092] Step S201: Obtain occupational classification standard documents and internal occupational information.
[0093] Step S202: Extract the occupational descriptions from the occupational classification standard file.
[0094] Step S203: Use the occupational description as prompt words for the deep learning generative model, and generate colloquial terms for the occupation and industry based on the deep learning generative model.
[0095] Step S204: Obtain the occupation name from the internal occupation information.
[0096] Step S205: Establish a mapping relationship between the occupational name and the colloquial terms for the occupation and the colloquial terms for the industry.
[0097] Step S206: Supplement the internal occupational information according to the mapping relationship to obtain an internal occupational standard document. The internal occupational standard document includes the spoken occupational name of each occupation, the standard occupational name corresponding to the spoken occupational name, the industry to which it belongs, and the risk level.
[0098] Step S207: Obtain the dialogue information between the user and the claims adjuster.
[0099] Step S208: Convert the dialogue information into dialogue information text.
[0100] Step S209: Perform word segmentation on the dialogue information text to obtain at least one text keyword.
[0101] Step S210: Extract multiple candidate occupations from the internal occupational standard document based on at least one of the text keywords.
[0102] Step S211: Perform word segmentation on each candidate occupation to obtain the first spoken occupation name keyword and the first industry keyword.
[0103] Step S212: Determine the second spoken occupational name keyword and the second industry keyword from at least one text keyword.
[0104] Step S213: Calculate the name similarity between the first spoken occupational name keyword and the second spoken occupational name keyword, and calculate the industry similarity between the first industry keyword and the second industry keyword.
[0105] Step S214: Determine the similarity between each candidate occupation and the keyword based on the name similarity and the industry similarity.
[0106] Step S215: Calculate the scores of multiple candidate occupations based on the similarity.
[0107] Step S216: Determine the candidate occupation with the highest score as the user's occupation.
[0108] Step S217: Determine the target risk level corresponding to the user's occupation, and recommend insurance products with a risk level higher than or equal to the target risk level for the user.
[0109] The specific implementation methods of each step in this embodiment are the same as those in the above embodiments, and will not be repeated here.
[0110] Reference Figure 3 , Figure 3 A flowchart illustrating another insurance product recommendation method provided in this application.
[0111] like Figure 3 As shown in the figure, this application embodiment provides an insurance product recommendation method, which includes steps S301 to S312.
[0112] In this embodiment, the insurance product recommendation method includes the following steps:
[0113] Step S301: Obtain occupational classification standard documents and internal occupational information.
[0114] Step S302: Extract the occupational descriptions from the occupational classification standard file.
[0115] Step S303: Use the occupational description as prompt words for the deep learning generative model, and generate colloquial terms for the occupation and industry based on the deep learning generative model.
[0116] Step S304: Obtain the occupation name from the internal occupation information.
[0117] Step S305: Establish a mapping relationship between the occupational name and the colloquial terms for the occupation and the colloquial terms for the industry.
[0118] Step S306: Supplement the internal occupational information according to the mapping relationship to obtain an internal occupational standard document. The internal occupational standard document includes the spoken occupational name of each occupation, the standard occupational name corresponding to the spoken occupational name, the industry to which it belongs, and the risk level.
[0119] Step S307: Obtain the dialogue information between the user and the claims adjuster.
[0120] Step S308: Convert the dialogue information into dialogue information text.
[0121] Step S309: Perform word segmentation on the dialogue information text to obtain at least one text keyword.
[0122] Step S310: Extract multiple candidate occupations from the internal occupational standard document based on at least one of the text keywords.
[0123] Step S311: Determine the target risk level corresponding to each of the candidate occupations to obtain multiple target risk levels.
[0124] Step S312: Recommend insurance products with a risk level higher than or equal to the highest target risk level to the user.
[0125] The specific implementation methods of each step in this embodiment are the same as those in the above embodiments, and will not be repeated here.
[0126] Please see Figure 4 , Figure 4 A schematic diagram of the functional modules of an insurance product recommendation device provided in this application.
[0127] like Figure 4 As shown, the insurance product recommendation device includes:
[0128] The colloquial term generation unit 401 is used to acquire occupational classification standard documents and internal occupational information, and to generate colloquial terms for occupations and colloquial terms for industries based on the occupational classification standard documents.
[0129] Occupational information supplementation unit 402 is used to supplement the internal occupational information according to the colloquial terms of the occupation and the colloquial terms of the industry to obtain an internal occupational standard document;
[0130] User occupation identification unit 403 is used to acquire dialogue information between the user and the claims adjuster, and to identify the user's occupation from the dialogue information according to the internal occupational standard document;
[0131] The insurance product recommendation unit 404 is used to recommend insurance products to the user based on the user's occupation.
[0132] Specifically, in generating colloquial terms for occupations and industries based on the occupational classification standard document, the colloquial term generation unit 401 is specifically used for:
[0133] Extract the occupational descriptions from the occupational classification standard file;
[0134] The occupational descriptions are used as prompts for a deep learning generative model, and colloquial terms for the occupations and industries are generated based on the deep learning generative model.
[0135] Specifically, in supplementing the internal occupational information based on the colloquial terms for the occupations and the colloquial terms for the industries to obtain internal occupational standard documents, the occupational information supplementation unit 402 is specifically used for:
[0136] Obtain the job title from the internal job information;
[0137] Establish a mapping relationship between the occupational name and the colloquial terms for the occupation and the colloquial terms for the industry;
[0138] The internal occupational information is supplemented based on the mapping relationship to obtain an internal occupational standard document.
[0139] Specifically, in the aspect of identifying the user's occupation from the dialogue information based on the internal occupational standard document, the user occupational identification unit 403 is specifically used for:
[0140] Convert the dialogue information into dialogue text;
[0141] The dialogue information text is segmented to obtain at least one text keyword;
[0142] Extract multiple candidate occupations from the internal occupational standard document based on at least one of the aforementioned text keywords;
[0143] Calculate the similarity between each of the candidate occupations and at least one of the text keywords;
[0144] Scores for multiple candidate occupations are calculated based on the similarity.
[0145] The user's occupation is determined based on the scores of multiple candidate occupations.
[0146] Furthermore, the internal occupational standard document includes the spoken occupational name for each occupation, the standard occupational name corresponding to the spoken occupational name, and the industry to which it belongs; in calculating the similarity between each candidate occupation and the keyword, the user occupational identification unit 403 is specifically used for:
[0147] Each candidate occupation is segmented into words to obtain the first spoken occupational name keyword and the first industry keyword.
[0148] Determine the second spoken occupational name keyword and the second industry keyword from at least one text keyword;
[0149] Calculate the name similarity between the first spoken occupational name keyword and the second spoken occupational name keyword, and calculate the industry similarity between the first industry keyword and the second industry keyword;
[0150] The similarity between each candidate occupation and the keyword is determined based on the name similarity and the industry similarity.
[0151] Furthermore, the internal occupational standard document includes a risk level for each occupation. In determining a user's occupation based on scores from multiple candidate occupations, the user occupational identification unit 403 is specifically used for:
[0152] The candidate occupation with the highest score is determined as the user's occupation;
[0153] In recommending insurance products to the user based on the user's occupation, the insurance product recommendation unit 404 is specifically used for:
[0154] Determine the target risk level corresponding to the user's occupation, and recommend insurance products with a risk level higher than or equal to the target risk level for the user.
[0155] Furthermore, the insurance product recommendation unit 404 is specifically used for:
[0156] Determine the target risk level for each of the candidate occupations to obtain multiple target risk levels;
[0157] Recommend insurance products with a risk level higher than or equal to the highest target risk level to the user.
[0158] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the above-described apparatus and modules can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0159] The aforementioned device can be implemented as a computer program, which can be used in, for example... Figure 4 It runs on the computer device shown.
[0160] Please see Figure 5 , Figure 5 This application provides a schematic block diagram of the structure of a computer device. The computer device may be a server.
[0161] See Figure 5 The computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.
[0162] Non-volatile storage media can store operating systems and computer programs. These computer programs include program instructions that, when executed, cause the processor to perform any insurance product recommendation method.
[0163] The processor provides computing and control capabilities, supporting the operation of the entire computer device.
[0164] Internal memory provides an environment for the execution of computer programs stored in non-volatile storage media. When executed by a processor, the computer program can cause the processor to perform any insurance product recommendation method.
[0165] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0166] It should be understood that the processor can be a Central Processing Unit (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. Among these, a general-purpose processor can be a microprocessor or any conventional processor.
[0167] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps:
[0168] Obtain occupational classification standard documents and internal occupational information, and generate colloquial terms for occupations and industry terms based on the occupational classification standard documents;
[0169] The internal occupational information is supplemented based on the colloquial terms for the occupations and the colloquial terms for the industries to obtain an internal occupational standard document;
[0170] Acquire dialogue information between the user and the claims adjuster, and identify the user's occupation from the dialogue information based on the internal occupational standards document;
[0171] Recommend insurance products to the user based on their occupation.
[0172] In one embodiment, in generating colloquial terms for occupations and industry terms based on the occupational classification standard document, the processor is configured to run a computer program stored in memory to perform the following steps:
[0173] Extract the occupational descriptions from the occupational classification standard file;
[0174] The occupational descriptions are used as prompts for a deep learning generative model, and colloquial terms for the occupations and industries are generated based on the deep learning generative model.
[0175] In one embodiment, in supplementing the internal occupational information based on colloquial terms for the occupations and industry colloquial terms to obtain an internal occupational standard document, the processor is configured to run a computer program stored in memory to perform the following steps:
[0176] Obtain the job title from the internal job information;
[0177] Establish a mapping relationship between the occupational name and the colloquial terms for the occupation and the colloquial terms for the industry;
[0178] The internal occupational information is supplemented based on the mapping relationship to obtain an internal occupational standard document.
[0179] In one embodiment, while identifying a user's occupational aspect from the dialogue information based on the internal occupational standards document, the processor is configured to run a computer program stored in memory to perform the following steps:
[0180] Convert the dialogue information into dialogue text;
[0181] The dialogue information text is segmented to obtain at least one text keyword;
[0182] Extract multiple candidate occupations from the internal occupational standard document based on at least one of the aforementioned text keywords;
[0183] Calculate the similarity between each of the candidate occupations and at least one of the text keywords;
[0184] Scores for multiple candidate occupations are calculated based on the similarity.
[0185] The user's occupation is determined based on the scores of multiple candidate occupations.
[0186] In one embodiment, the internal occupational standard document includes the spoken occupational name for each occupation, the standard occupational name corresponding to the spoken occupational name, and the industry to which it belongs; in calculating the similarity between each candidate occupation and the keyword, the processor is configured to run a computer program stored in memory to perform the following steps:
[0187] Each candidate occupation is segmented into words to obtain the first spoken occupational name keyword and the first industry keyword.
[0188] Determine the second spoken occupational name keyword and the second industry keyword from at least one text keyword;
[0189] Calculate the name similarity between the first spoken occupational name keyword and the second spoken occupational name keyword, and calculate the industry similarity between the first industry keyword and the second industry keyword;
[0190] The similarity between each candidate occupation and the keyword is determined based on the name similarity and the industry similarity.
[0191] In one embodiment, the internal occupational standards document includes a risk level for each occupation, and in determining a user's occupation based on scores from a plurality of candidate occupations, the processor is configured to run a computer program stored in memory to perform the following steps:
[0192] The candidate occupation with the highest score is determined as the user's occupation;
[0193] The step of recommending insurance products to the user based on the user's occupation includes:
[0194] Determine the target risk level corresponding to the user's occupation, and recommend insurance products with a risk level higher than or equal to the target risk level for the user.
[0195] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps:
[0196] Determine the target risk level for each of the candidate occupations to obtain multiple target risk levels;
[0197] Recommend insurance products with a risk level higher than or equal to the highest target risk level to the user.
[0198] The embodiments of this application also provide a computer-readable storage medium storing a computer program, the computer program including program instructions, and the processor executing the program instructions to implement any of the insurance product recommendation methods provided in the embodiments of this application.
[0199] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device.
[0200] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A method for recommending insurance products, characterized in that, The method for recommending insurance products includes the following steps: Obtain occupational classification standard documents and internal occupational information, and generate colloquial terms for occupations and industry terms based on the occupational classification standard documents; The internal occupational information is supplemented based on the colloquial terms for the occupations and the colloquial terms for the industries to obtain an internal occupational standard document. The internal occupational standard document includes occupational code, occupational name, occupational category, occupational description, colloquial terms for the occupation, colloquial terms for the industry, and risk level. The system acquires dialogue information between the user and the claims adjuster, and identifies the user's occupation from the dialogue information based on the internal occupational standards document; it then recommends insurance products to the user based on the user's occupation. The step of generating colloquial terms for occupations and colloquial terms for industries based on the occupational classification standard document includes: Extract occupational descriptions from the occupational classification standard file; use the occupational descriptions as prompt words for the deep learning generative model, and generate colloquial terms for the occupations and industry terms based on the deep learning generative model; The step of identifying the user's occupation from the dialogue information based on the internal occupational standard document includes: Each candidate occupation corresponding to the dialogue information is segmented to obtain a first spoken occupation name keyword and a first industry keyword; a second spoken occupation name keyword and a second industry keyword are determined from at least one text keyword. Calculate the name similarity between the first spoken occupational name keyword and the second spoken occupational name keyword, and calculate the industry similarity between the first industry keyword and the second industry keyword; The similarity between each candidate occupation and the keyword is determined based on the name similarity and the industry similarity, and a score is calculated for multiple candidate occupations based on the similarity; the user's occupation is determined based on the scores of multiple candidate occupations. The dialogue information includes environmental information and emotional information during the dialogue. Identifying the user's occupation from the dialogue information based on the internal occupational standard document further includes: extracting environmental noise features from the environmental information; performing auxiliary verification or differentiation of the user's occupation based on the environmental noise features; analyzing the emotional information; and if the emotional information is negative, increasing the target risk level corresponding to the user's occupation.
2. The insurance product recommendation method as described in claim 1, characterized in that, The process of supplementing the internal occupational information based on the colloquial terms for the occupations and industry-specific colloquial terms to obtain an internal occupational standard document includes: Obtain the job title from the internal job information; Establish a mapping relationship between the occupational name and the colloquial terms for the occupation and the colloquial terms for the industry; The internal occupational information is supplemented based on the mapping relationship to obtain an internal occupational standard document.
3. The insurance product recommendation method as described in claim 1, characterized in that, The step of identifying the user's occupation from the dialogue information based on the internal occupational standard document includes: Convert the dialogue information into dialogue text; The dialogue information text is segmented to obtain at least one text keyword; Multiple candidate occupations are extracted from the internal occupational standard document based on at least one of the stated text keywords.
4. The insurance product recommendation method as described in claim 1, characterized in that, The internal occupational standards document includes a risk level for each occupation, and the process of determining a user's occupation based on the scores of multiple candidate occupations includes: The candidate occupation with the highest score is determined as the user's occupation; The step of recommending insurance products to the user based on the user's occupation includes: Determine the target risk level corresponding to the user's occupation, and recommend insurance products with a risk level higher than or equal to the target risk level for the user.
5. The insurance product recommendation method as described in claim 1, characterized in that, The method for recommending insurance products also includes: Determine the target risk level for each of the candidate occupations to obtain multiple target risk levels; Recommend insurance products with a risk level higher than or equal to the highest target risk level to the user.
6. An insurance product recommendation device, characterized in that, The insurance product recommendation device includes: The colloquial term generation unit is used to acquire occupational classification standard documents and internal occupational information, and to generate colloquial terms for occupations and industry colloquial terms based on the occupational classification standard documents. The occupational information supplementation unit is used to supplement the internal occupational information based on the colloquial terms of the occupation and the colloquial terms of the industry to obtain an internal occupational standard document. The internal occupational standard document includes occupational code, occupational name, occupational category, occupational description, colloquial terms of the occupation, colloquial terms of the industry, and risk level. The user occupation identification unit is used to acquire dialogue information between the user and the claims adjuster, and to identify the user's occupation from the dialogue information according to the internal occupational standard document; the insurance product recommendation unit is used to recommend insurance products to the user based on the user's occupation. The colloquial term generation unit is further configured to: extract occupational descriptions from the occupational classification standard file; use the occupational descriptions as prompts for the deep learning generative model; and generate colloquial terms for occupations and industry terms based on the deep learning generative model. The user occupation identification unit is further configured to: perform word segmentation processing on each candidate occupation corresponding to the dialogue information to obtain a first spoken occupation name keyword and a first industry keyword; and determine a second spoken occupation name keyword and a second industry keyword from at least one text keyword. Calculate the name similarity between the first spoken occupational name keyword and the second spoken occupational name keyword, and calculate the industry similarity between the first industry keyword and the second industry keyword; The similarity between each candidate occupation and the keyword is determined based on the name similarity and the industry similarity, and a score is calculated for multiple candidate occupations based on the similarity; the user's occupation is determined based on the scores of multiple candidate occupations. The dialogue information also includes environmental information and emotional information during the dialogue; the user occupation identification unit is further used to: extract environmental noise features from the environmental information; perform auxiliary verification or differentiation of the user's occupation based on the environmental noise features; analyze the emotional information; if the emotional information is negative, then increase the target risk level corresponding to the user's occupation.
7. A computer device, characterized in that, The computer device includes a processor, a memory, and an insurance product recommendation program stored in the memory and executable by the processor, wherein when the insurance product recommendation program is executed by the processor, it implements the steps of the insurance product recommendation method as described in any one of claims 1 to 5.
8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an insurance product recommendation program, wherein when the insurance product recommendation program is executed by a processor, it implements the steps of the insurance product recommendation method as described in any one of claims 1 to 5.