Product recommendation method and device, electronic equipment, chip and storage medium

Abstract

The invention provides a product recommendation method and device, electronic equipment, a chip and a storage medium, and the method comprises the steps: generating a cue word according to static information and behavior information of a user, the behavior information being obtained by carrying out feature extraction on the operation of the user by using a first network; according to the prompt word, a recommendation list corresponding to the user is generated, and the recommendation list comprises at least one recommended product. The prompt word can be dynamically generated according to the static information and the behavior information of the user, and the product recommendation accuracy can be improved.

Description

Technical Field

[0001] This disclosure relates to the field of data processing, and more particularly to a product recommendation method, apparatus, electronic device, chip, and storage medium. Background Technology

[0002] Existing life insurance recommendation systems are typically based on fixed recommendation algorithms, combining basic customer information (such as age, gender, occupation, etc.) and historical behavioral data (such as life insurance products viewed, consultation records, etc.) to make recommendations. These systems use a preset prompt word library and cannot adjust according to the customer's real-time needs and context during the recommendation process. Summary of the Invention

[0003] This disclosure provides a product recommendation method, apparatus, electronic device, chip, and storage medium to solve problems in the related art.

[0004] A first aspect of this disclosure provides a product recommendation method, the method comprising: generating prompt words based on a user's static information and behavioral information, wherein the behavioral information is obtained by feature extraction of the user's operations using a first network; and generating a recommendation list corresponding to the user based on the prompt words, wherein the recommendation list includes at least one recommended product.

[0005] In some embodiments of this disclosure, generating prompt words based on the user's static and behavioral information includes: obtaining industry knowledge information and market dynamic information from a knowledge base; performing data analysis on the static and behavioral information based on the industry knowledge information and market dynamic information to determine the user's contextual and demand information; determining a target template from a set of preset prompt word templates based on the contextual and demand information; and generating prompt words based on the target template, contextual and demand information.

[0006] In some embodiments of this disclosure, generating a recommendation list for a user based on a prompt word includes: retrieving customer cases related to the prompt word from a knowledge base; and generating a recommendation list for the user based on the customer cases and information on at least one candidate product included in the knowledge base.

[0007] In some embodiments of this disclosure, the method further includes: in response to an automatic update instruction or a user update instruction, obtaining updated content, wherein the automatic update instruction is triggered when preset conditions are met, and the user update instruction is triggered in response to a user's update operation, wherein the preset conditions include at least one of detecting an update to product information on a first website and determining that the fluctuation range of market data exceeds a preset threshold; updating the knowledge base according to the updated content; and updating the version information of the knowledge base.

[0008] In some embodiments of this disclosure, in response to an automatic update instruction or a user update instruction, obtaining updated content includes at least one of the following: in response to an automatic update instruction, obtaining updated content from a first website, the updated content including updated product information; in response to an automatic update instruction, comparing current market data with data in a knowledge base to obtain updated content; in response to a user update instruction, obtaining updated content from information input by the user.

[0009] In some embodiments of this disclosure, the knowledge base includes a cache area and a non-cache area. The cache area includes at least one of the following: information on at least one hot candidate product, at least one historical user profile information, and at least one historical customer case.

[0010] A second aspect of this disclosure provides a product recommendation device, comprising: a first processing unit for generating prompt words based on a user's static information and behavioral information, wherein the behavioral information is obtained by feature extraction of the user's operations using a first network; and a second processing unit for generating a recommendation list corresponding to the user based on the prompt words, wherein the recommendation list includes at least one recommended product.

[0011] A third aspect of this disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the methods described in the first aspect of this disclosure.

[0012] A fourth aspect of this disclosure provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods described in the first aspect of this disclosure.

[0013] A fifth aspect of this disclosure provides a chip including at least one processor and a communication interface; the communication interface is used to receive signals input to the chip or signals output from the chip, and the processor communicates with the communication interface and implements the method described in the first aspect of this disclosure through logic circuits or executing code instructions.

[0014] In summary, the product recommendation method proposed in this disclosure can dynamically generate prompts based on users' static and behavioral information. Subsequently, a recommendation list corresponding to the user can be generated based on the dynamically generated prompts, which can improve the accuracy of product recommendations and enhance user experience.

[0015] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0017] Figure 1 A flowchart illustrating a product recommendation method provided in this embodiment of the disclosure. Figure 1 ;

[0018] Figure 2 A flowchart illustrating a product recommendation method provided in this embodiment of the disclosure. Figure 2 ;

[0019] Figure 3 A flowchart illustrating a product recommendation method provided in this embodiment of the disclosure. Figure 3 ;

[0020] Figure 4A An architecture diagram of a life insurance intelligent recommendation system that integrates a large model, dynamic prompts, and a knowledge base is provided for embodiments of this disclosure;

[0021] Figure 4B A flowchart illustrating an intelligent life insurance recommendation method that integrates a large model, dynamic prompts, and a knowledge base, provided as an embodiment of this disclosure;

[0022] Figure 4C A flowchart illustrating real-time knowledge base construction provided in this embodiment of the disclosure;

[0023] Figure 5 This is a schematic diagram of the structure of a product recommendation device provided in an embodiment of the present disclosure;

[0024] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure;

[0025] Figure 7 This is a schematic diagram of the chip structure provided in an embodiment of this disclosure. Detailed Implementation

[0026] Embodiments of this disclosure are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0027] Existing life insurance recommendation systems typically rely on fixed recommendation algorithms, combining basic customer information (such as age, gender, and occupation) with historical behavioral data (such as viewed life insurance products and consultation records) to make recommendations. These systems use pre-set suggestion word libraries and cannot adjust to customers' real-time needs and contexts during the recommendation process. Furthermore, the knowledge base is not updated in a timely manner, often requiring regular manual maintenance. This prevents the system from reflecting dynamic changes in the life insurance market (such as new product launches, discontinuation of old products, and rate adjustments), resulting in limited accuracy and personalization in recommendations.

[0028] The recommendation methods in related technologies have the following problems:

[0029] Lack of dynamic adaptability: Fixed recommendation algorithms and preset suggestion word libraries cannot adapt to dynamic changes in customer needs. Customers have significantly different needs for life insurance products at different life stages and economic environments, and the existing system cannot flexibly adjust recommendation strategies.

[0030] Lagging knowledge base updates: Knowledge base updates require manual intervention and have a long cycle, making it impossible to reflect market dynamics and product changes in a timely manner. This can easily lead to recommending outdated or inaccurate life insurance products to customers.

[0031] Insufficient personalization: The accuracy and personalization of recommendations are limited, failing to meet the diverse needs of customers. Existing systems often make recommendations based on a limited set of features, ignoring many potential customer needs.

[0032] Therefore, in order to solve the above problems, this disclosure proposes a product recommendation method that integrates large-scale models, dynamic prompt word generation, and knowledge base technology to achieve personalized, accurate, and intelligent recommendations for life insurance products. By dynamically generating prompt words and updating the knowledge base in real time, the system's responsiveness to customer needs and the accuracy of recommendations are improved.

[0033] The specific details of this method are as follows.

[0034] Figure 1 A flowchart illustrating a product recommendation method provided in this embodiment of the disclosure. Figure 1 .like Figure 1 As shown, the method may include the following steps.

[0035] Step 101: Generate prompt words based on the user's static and behavioral information.

[0036] In some embodiments, when it is necessary to recommend products (such as insurance products) to a user, static information and behavioral information of the user can be obtained. The static information may include information related to the user, such as the user's historical order information. The user's personal information includes, for example, the user's age, occupation, medical examination results, user query information, user input information, and user needs. The user's historical order information includes, for example, the user's historical insurance purchase information, historical policies, and coverage data. For example, when a user needs to obtain a recommendation list, the user can enter some static information in the query box. Optionally, the user can enter demand information or keywords. Optionally, the user's personal information, such as age and occupation, can also be entered. Then, based on the information entered by the user, a query can be performed in the database (storage database) to determine whether other information of the user, such as historical order information and historical reimbursement information, exists in the database.

[0037] In other words, it can receive customer input, including basic information, descriptions of needs, and real-time inquiries. This can be done through a web interface or mobile application for easy customer access. A front-end framework can be used to build the user interface, and a RESTful API can be used to interact with back-end services. Static information can be determined based on user input. For example, natural language processing can be used to extract key information such as age, income, family situation, and insurance needs. Optionally, static information can also be determined based on historical user information stored in a database. Static information can be used to indicate a user's basic conditions, such as their health status, financial situation, and preferences.

[0038] In some embodiments, user behavior information can be obtained by extracting features from user operations using a first network, such as a Temporal Convolutional Network (TCN). User operations include click operations, browsing operations, etc. Optionally, when users select or browse insurance product introductions on the official website, they usually browse products of interest for a longer period of time or click more times. Therefore, features can be extracted from user operations, such as dwell time, click stream, etc. User behavior information can be used to indicate user needs and user preferences.

[0039] In some embodiments, prompt words can be used to prompt the model to determine recommended products that meet the user's expectations. Prompt words can help the model better understand the user's needs and preferences, thereby recommending products that better meet the user's expectations. For example, the prompt word could be "Recommend a high-coverage, multi-protection life insurance product for a 45-year-old housewife who wants to upgrade her life insurance coverage".

[0040] In some embodiments, prompt words can be dynamically generated based on the user's static and behavioral information. The solution disclosed herein does not require the use of preset prompt words and can generate more targeted prompt words based on the user's real-time static and behavioral information. That is, this application adopts dual feature input when generating prompt words: explicit user input text + implicit behavioral sequence (click flow, dwell time, etc.), which can more comprehensively understand the user's needs and preferences.

[0041] Step 102: Generate a recommendation list for the user based on the prompt words.

[0042] In some embodiments, a recommendation list corresponding to a user can be generated based on the prompt word. The recommendation list includes at least one recommended product, that is, the recommendation list includes at least one product recommended to the user. Optionally, generating a recommendation list corresponding to a user based on the prompt word includes: retrieving customer cases related to the prompt word from a knowledge base; and generating a recommendation list corresponding to the user based on the customer cases and information on at least one candidate product included in the knowledge base.

[0043] In some embodiments, a customer case may include historical transaction order information and user profile information. For example, a customer case may be: User A, age 35, female, occupation: teacher, needs to be protected against accident and illness risks, and ultimately purchases product A, including the sum insured, premium, coverage, effective period, number of reimbursements, reimbursement amount, etc. of product A.

[0044] In some embodiments, the information of at least one candidate product can be information related to all products offered by the insurance company, such as the sum insured, premium, coverage, and duration of coverage. That is, when generating the recommendation list, customer cases that are similar to the user can be selected based on prompts. Then, the recommended products can be determined by referring to customer cases, user needs, and product-related information, which can make the recommended products more in line with the user's expectations.

[0045] In some embodiments, the knowledge base can store customer cases, product information, industry knowledge (such as regulatory requirements for the insurance industry, insurance-related knowledge, etc.), user historical information, user profiles, etc. In other words, the knowledge base includes life insurance product information (such as product name, coverage, premium rates, claims conditions, etc.), industry dynamics (such as new policy releases, market trends, etc.), and customer cases. The knowledge base can store the above data in a structured manner. For example, various types of life insurance-related knowledge collected can be structured, and knowledge graph technology can be used to associate and represent the knowledge, thereby determining the relationships between different pieces of knowledge and generating structured data. This structured data can be stored in the knowledge base. This structured data storage method improves the efficiency of knowledge retrieval and reasoning capabilities when querying data from the knowledge base.

[0046] In some embodiments, after determining the prompt words, similar customer cases can be obtained from the knowledge base based on the prompt words. At least one candidate product can be initially determined by referring to the similar customer cases. Then, at least one candidate product related to the user's needs can be determined from the knowledge base based on the prompt words. Then, semantic similarity matching can be performed on the multiple candidate products determined based on the customer cases and the prompt words, that is, the similarity between the multiple candidate products and the user's needs can be determined. For example, the similarity (semantic similarity) between the multiple candidate products and the prompt words can be determined. Then, at least one candidate product with a high similarity can be selected as the recommended product.

[0047] In some embodiments, the knowledge base may optionally include a cache area and a non-cache area. The cache area includes at least one of the following: information on at least one hot candidate product, at least one historical user profile, and at least one historical customer case. In other words, the knowledge base may include a cache area that can cache information related to hot products. Hot products may be, for example, products that are purchased in large quantities, products with high sales volume over a period of time, or products with high sales volume in the current geographical area, such as travel insurance.

[0048] Optionally, the cache can cache user image information, such as user historical order information, user static feature information, user preference information, etc. The cache can also cache customer case information and the results of the above semantic similarity matching, that is, it can cache the correspondence between user needs and products. Optionally, when determining the recommendation list based on prompt words, the recommended products can be determined first based on the data in the cache of the knowledge base. When the similarity between the product information in the cache and the user needs is low, the recommended products can be determined from the non-cache area of ​​the knowledge base. Optionally, when determining the recommended products based on the data in the cache of the knowledge base, the cached customer cases, user images, semantic similarity matching results, etc. can be directly referenced to determine the recommended products, which can reduce the amount of computation.

[0049] In some embodiments, after determining at least one recommended product, the at least one recommended product can be sorted and filtered according to its similarity to user needs and its cost-effectiveness to obtain a recommendation list. Optionally, the recommendation list may include information about at least one recommended product. For example, the recommendation list may provide detailed information on the features, coverage, and rates of each recommended product, and compare and analyze it with industry average data in the knowledge base to allow customers to more intuitively understand the advantages of the recommended products.

[0050] In some embodiments, the recommendation list can be presented to users in a visual manner, allowing customers to more intuitively understand the advantages of the recommended products. Then, user feedback can be received, such as users modifying keywords, changing requirements, adding requirements and keywords, etc. The system can receive user-modified information and then adjust the prompt words based on the user's changes, i.e., regenerate prompt words. After that, the recommended products can be re-determined based on the newly generated prompt words to optimize the recommendation results.

[0051] In some embodiments, the method further includes: in response to an automatic update instruction or a user update instruction, obtaining updated content, wherein the automatic update instruction is triggered when preset conditions are met, and the user update instruction is triggered in response to a user's update operation, wherein the preset conditions include at least one of detecting an update to product information on a first website and determining that the fluctuation range of market data exceeds a preset threshold; updating the knowledge base according to the updated content; and updating the version information of the knowledge base.

[0052] In some embodiments, in response to an automatic update instruction or a user update instruction, obtaining updated content includes at least one of the following: in response to an automatic update instruction, obtaining updated content from a first website, the updated content including updated product information; in response to an automatic update instruction, comparing current market data with data in a knowledge base to obtain updated content; in response to a user update instruction, obtaining updated content from information input by the user.

[0053] In other words, a data crawling and monitoring mechanism can be established to regularly obtain the latest life insurance product information and industry trends from data sources such as official websites and industry media, and automatically update the knowledge base. At the same time, a manual update interface is provided to facilitate manual intervention and error correction.

[0054] In some embodiments, the first website may be, for example, an official website containing product information. This means the system can detect whether product information has been updated on the official website, such as adding, deleting, or modifying products. When an update is detected, the updated product information can be retrieved from the first website, and then the product information stored in the knowledge base can be automatically updated based on the updated information. Optionally, market data can be detected. Market data refers to data related to the insurance product industry, such as market regulatory requirements, sales volume of various types of insurance products in the market, national policies, attention to various types of insurance products, economic environment, employment environment, etc. When market data fluctuates significantly, an automatic update instruction can be triggered. At this time, the detected market data can be compared with the relevant data stored in the knowledge base to determine the discrepancies. Optionally, the discrepancies are the content that needs to be updated, so the knowledge base can be updated based on the discrepancies. Optionally, the first website and market data can be periodically detected and updated when changes occur.

[0055] In some embodiments, a user can trigger a user update command, meaning the user can actively update the database. The user can trigger the update command through operations such as adding, deleting, or modifying data, and then input the content to be updated, such as information to be added or deleted. After receiving the content to be updated, the knowledge base can be updated accordingly. Optionally, an incremental update method can be used to update the knowledge base.

[0056] Optionally, the version information of the knowledge base can be managed, that is, after updating the knowledge base, the version information of the knowledge base can be updated, and the update history information of the knowledge base can be managed.

[0057] In summary, the embodiments disclosed above can dynamically generate prompt words based on the user's static and behavioral information. Subsequently, a recommendation list corresponding to the user can be generated based on the dynamically generated prompt words, which can improve the accuracy of recommended products. Furthermore, the data in the knowledge base can be updated in a timely manner as needed, ensuring the effectiveness of the knowledge base data and improving the user experience.

[0058] Figure 2 A flowchart illustrating a product recommendation method provided in this embodiment of the disclosure. Figure 2 .like Figure 2 As shown, based on Figure 1 The illustrated embodiment shows that the method includes the following steps.

[0059] Step 201: Obtain industry knowledge information and market dynamic information from the knowledge base.

[0060] In some embodiments, the knowledge base can store customer cases, product information, industry knowledge (such as regulatory requirements of the insurance industry, insurance-related knowledge, etc.), user history information, user profiles, etc. That is, the knowledge base includes life insurance product information (such as product name, coverage, premium rate, claims conditions, etc.), industry dynamics (such as market dynamics, new policy introductions, market trends, etc.) and customer cases.

[0061] In some embodiments, optionally, industry knowledge information and market dynamics information related to the user can be obtained based on the user's static information and behavioral information.

[0062] Step 202: Based on industry knowledge and market dynamics, perform data analysis on static and behavioral information to determine the user's contextual and demand information.

[0063] In some embodiments, a large model can be used to perform data analysis on static and behavioral information based on industry knowledge and market dynamics to determine the user's contextual and demand information. Contextual information can be information about the user's current situation or the user's background information. For example, contextual information could be that the user is a 25-year-old office worker with mortgage pressure. Demand information can be the user's needs, i.e., information related to the products the user expects. For example, demand information could be life insurance products that protect against accidental and disease risks.

[0064] Step 203: Based on the contextual information and the requirement information, determine the target template from a set of preset prompt word templates.

[0065] In some embodiments, a target template that is highly similar to the contextual information and the demand information can be determined from a plurality of preset prompt word templates based on the contextual information and the demand information.

[0066] Step 204: Generate prompt words based on the target template, contextual information, and requirement information.

[0067] In some embodiments, contextual information and demand information can be optionally filled into a preset prompt word template to dynamically generate prompt words. For example, the generated prompt word could be "Recommend a life insurance product that covers accident and illness risks for a 25-year-old working professional with mortgage pressure." In other words, based on the results of information extraction and context analysis, a suitable template can be selected from the predefined prompt word templates, and the corresponding information can be filled in to generate the final prompt word.

[0068] In summary, the above embodiments of this application can dynamically generate prompt words based on the user's static and behavioral information, and then generate a recommendation list corresponding to the user based on the dynamically generated prompt words, which can improve the accuracy of recommended products and enhance the user experience.

[0069] Figure 3 A flowchart illustrating a product recommendation method provided in this embodiment of the disclosure. Figure 3 .like Figure 3 As shown, based on Figure 1 The illustrated embodiment shows that the method includes the following steps.

[0070] Step 301: Based on the prompt words, retrieve customer cases related to the prompt words from the knowledge base.

[0071] In some embodiments, the knowledge base may store customer case information, which may include, for example, the customer's historical transaction order information, and may include user profile information. For example, a customer case may be: User A, age 35, female, occupation: teacher, needs to be protected against accident and illness risks, and ultimately purchases product A, including the sum insured, premium, coverage, effective period, and number of reimbursements, reimbursement amount, etc. of product A.

[0072] In some embodiments, after determining the prompt words, similar customer cases can be obtained from the knowledge base based on the prompt words, and at least one candidate product can be initially determined by referring to the similar customer cases.

[0073] Step 302: Generate a recommendation list for the user based on the customer case and information on at least one candidate product included in the knowledge base.

[0074] In some embodiments, after determining the prompt words, similar customer cases can be obtained from the knowledge base based on the prompt words. At least one candidate product can be initially determined by referring to the similar customer cases. Then, at least one candidate product related to the user's needs can be determined from the knowledge base based on the prompt words. Then, semantic similarity matching can be performed on the multiple candidate products determined based on the customer cases and the prompt words, that is, the similarity between the multiple candidate products and the user's needs can be determined. For example, the similarity (semantic similarity) between the multiple candidate products and the prompt words can be determined. Then, at least one candidate product with a high similarity can be selected as the recommended product.

[0075] Optionally, the knowledge base includes a cache area and a non-cache area. The cache area includes at least one of the following: information on at least one hot candidate product, at least one historical user profile, and at least one historical customer case. In other words, the knowledge base may include a cache area, which can cache information related to hot products and user image information, such as users' historical order information, users' static feature information, users' preference information, etc. Optionally, when determining the recommendation list based on prompts, recommended products can be determined first based on the data in the cache area of ​​the knowledge base. When the similarity between the product information in the cache area and the user's needs is low, recommended products can be determined from the non-cache area of ​​the knowledge base. Optionally, when determining recommended products based on the data in the cache area of ​​the knowledge base, information such as customer cases, user images, and semantic similarity matching results cached in the cache area can be directly referenced to determine recommended products, which can reduce the amount of computation.

[0076] In some embodiments, after determining at least one recommended product, the at least one recommended product can be sorted and filtered according to its similarity to user needs and its cost-effectiveness to obtain a recommendation list. Optionally, the recommendation list may include information about at least one recommended product. For example, the recommendation list may provide detailed information on the features, coverage, and rates of each recommended product, and compare and analyze it with industry average data in the knowledge base to allow customers to more intuitively understand the advantages of the recommended products.

[0077] In summary, the embodiments disclosed above can dynamically generate prompt words based on the user's static and behavioral information. Subsequently, a recommendation list corresponding to the user can be generated based on the dynamically generated prompt words, which can improve the accuracy of recommended products. Furthermore, the data in the knowledge base can be updated in a timely manner as needed, ensuring the effectiveness of the knowledge base data and improving the user experience.

[0078] The technical solutions of this disclosure will be further described in detail below with reference to specific application embodiments.

[0079] The following is an embodiment of the present disclosure of a life insurance intelligent recommendation system and method that integrates a large model, dynamic prompts, and a knowledge base. The method includes the following:

[0080] Dynamic prompt generation: Dynamically generates prompts that match the customer's context, improving the accuracy and personalization of recommendations. Based on real-time customer inquiries and behaviors, the prompts are adjusted in real time, guiding the large model to output recommendations that better meet the customer's needs.

[0081] Real-time knowledge base construction: Build a structured, easily searchable, and real-time updated knowledge base to provide accurate data support for the recommendation system. Integrate the latest information on various life insurance products and industry trends in the market to ensure the system can obtain accurate information in a timely manner.

[0082] Intelligent recommendation implementation: Combining large-scale models, dynamic prompts, and knowledge base technology, we can achieve intelligent recommendations for life insurance products, providing customers with more accurate and personalized life insurance product recommendation solutions.

[0083] like Figure 4A The diagram shown is an architecture diagram of a life insurance intelligent recommendation system that integrates a large model, dynamic prompts, and a knowledge base. Figure 4B The diagram shows a flowchart of a life insurance intelligent recommendation method that integrates a large model, dynamic prompts, and a knowledge base.

[0084] 1. User interaction module.

[0085] Function: Receives customer input, including basic information, description of needs, and real-time queries, and passes this information to the dynamic prompt word generation module. Simultaneously, it outputs recommendation results to the customer.

[0086] Implementation: A web interface or mobile application will be used for easy customer operation. A front-end framework will be used to build the user interface and interact with the back-end service via a RESTful API.

[0087] 2. Dynamic prompt word generation module.

[0088] Function: Based on customer information transmitted from the user interaction module and relevant knowledge in the knowledge base, dynamically generate prompts that match the customer context. These prompts will serve as input to guide the large model in reasoning.

[0089] Implementation method:

[0090] Information extraction: Perform natural language processing on the information input by customers to extract key information, such as customer age, income, family situation, insurance needs, etc.

[0091] Contextual analysis: Combining industry knowledge and market dynamics from the knowledge base, we analyze the client's current context and needs to determine the direction and focus of recommendations.

[0092] Prompt generation: Based on the results of information extraction and context analysis, a suitable template is selected from predefined prompt templates, and the corresponding information is filled in to generate the final prompt. For example, if the customer is a young working professional with mortgage pressure, the prompt could be "Recommend a life insurance product that covers accident and illness risks for a working professional around 25 years old with mortgage pressure."

[0093] Key points for implementation:

[0094] A dual-feature input approach is employed: explicit user input text plus implicit behavioral sequences (clickstream, dwell time, etc.). Knowledge graph-based association queries utilize the Cypher language to establish the correlation dimensions between user profiles and product features. The risk scoring model employs XGBoost, with training data including historical policy payout rates and user feature correlation data.

[0095] 3. Large model reasoning module.

[0096] Function: Receive prompts from the dynamic prompt generation module, use a large model for reasoning, and generate a list of recommended life insurance products.

[0097] Implementation: Select a suitable large language model and use the API interface to send prompt words to the model for inference. For open-source large models, local deployment and fine-tuning can be performed to improve the model's specialization and adaptability to the life insurance field. The model output may be a preliminary recommendation list and a brief description of the rationale.

[0098] Key technologies:

[0099] The RAG (Retrieval Enhanced Generation) architecture is adopted to reduce the illusion of large models.

[0100] The fine-tuned dataset contains 500,000 conversation records in the insurance field plus actuary-annotated data.

[0101] Constraint generation uses the GSLM framework to ensure that recommended products meet affordability rules.

[0102] 4. Knowledge base management module.

[0103] Function: Responsible for building, updating, and querying the knowledge base. The knowledge base contains information such as life insurance product information (e.g., product name, coverage, premium rates, claims conditions), industry news (e.g., new policies, market trends), and customer case studies.

[0104] Implementation method:

[0105] Knowledge base construction: The collected life insurance-related knowledge is structured and stored in a database. Knowledge graph technology is used to associate and represent the knowledge, improving query efficiency and reasoning capabilities.

[0106] Real-time updates: A data crawling and monitoring mechanism has been established to regularly obtain the latest life insurance product information and industry trends from data sources such as official websites and industry media, and automatically update the knowledge base. At the same time, a manual update interface is provided to facilitate manual intervention and error correction.

[0107] Knowledge Query: Based on the needs of the dynamic prompt word generation module and the large-scale model reasoning module, relevant knowledge is retrieved from the knowledge base to provide data support for the recommendation process. Optionally, the real-time knowledge base construction process is as follows: Figure 4C As shown.

[0108] Technical parameters:

[0109] Data update frequency: Regulatory policy changes are pushed out in real time, and product information is updated incrementally every hour.

[0110] Version control uses Git-LFS to manage the knowledge graph change history.

[0111] Layered cache design:

[0112] L1: Hot Product Information (TTL 5min);

[0113] L2: User profile feature cache (TTL 1h);

[0114] L3: Large model response cache (semantic similarity matching).

[0115] 5. Recommendation results output module.

[0116] Function: Organizes and optimizes the recommendation results output by the large model inference module, combines them with detailed information from the knowledge base, generates a detailed recommendation report, and provides feedback to the user interaction module.

[0117] Implementation: The recommended results are sorted and filtered, prioritizing life insurance products that meet customer needs and offer high cost-effectiveness. The recommendation report details the features, coverage, and premiums of each recommended product, and compares them with industry average data in the knowledge base, allowing customers to more intuitively understand the advantages of the recommended products.

[0118] System workflow:

[0119] Customers input basic information and insurance needs through the user interaction module.

[0120] The dynamic prompt generation module processes customer input and generates dynamic prompts by combining information from the knowledge base.

[0121] The large-scale model reasoning module performs inference based on the prompt words and generates a preliminary list of life insurance product recommendations.

[0122] The recommendation results output module organizes and optimizes the recommendation list, retrieves detailed information from the knowledge base, and generates a detailed recommendation report.

[0123] The user interaction module displays the recommendation report to the customer and receives customer feedback. Based on the feedback, the system can repeat the above steps to further optimize the recommendations.

[0124] In summary, the examples disclosed above significantly improve the accuracy and personalization of life insurance recommendations by integrating large models, dynamic prompt word generation, and knowledge base technologies. The dynamically generated prompt words and the real-time updated knowledge base make the recommendation system more flexible and intelligent, better able to meet the diverse needs of customers. The application of this invention will drive the life insurance industry towards a more intelligent and personalized direction, providing strong support and technological impetus for the innovative transformation of the life insurance industry.

[0125] The following are some scenario-based examples provided by this solution.

[0126] Scenario 1: Life insurance recommendations for young professionals.

[0127] Background: Mr. Li, 28 years old, has just entered the workforce and is under mortgage pressure. He has a basic understanding of life insurance products, but not an in-depth one.

[0128] Workflow:

[0129] User interaction: Mr. Li enters his basic information (age, gender, occupation, income, etc.) and insurance needs (such as wanting to be covered for accident and illness risks) through the web interface or mobile application.

[0130] Dynamic prompt word generation: After receiving Mr. Li's information, the system performs natural language processing to extract key information.

[0131] By combining industry knowledge from the knowledge base (such as common insurance needs of young professionals and insurance configuration suggestions under mortgage pressure), we can analyze Mr. Li's situation and needs.

[0132] The system generates a dynamic suggestion: "Recommend a comprehensive life insurance product that covers accident and critical illness for a 28-year-old young professional with mortgage pressure."

[0133] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0134] The large model generates a recommendation list containing various life insurance products based on prompts and built-in algorithms, along with a brief introduction and reasons for recommending each product.

[0135] Recommendation results output: The system sorts and filters the recommendation list, prioritizing products with high cost-performance ratios. A detailed recommendation report is generated and displayed to Mr. Li through the user interaction module.

[0136] Outcome: Through this detailed recommendation report, Mr. Li quickly understood the life insurance products that suited him and successfully purchased a comprehensive life insurance policy that covers accidents and critical illnesses.

[0137] Scenario 2: Life insurance upgrade for middle-aged housewives.

[0138] Background: Ms. Zhang, 45 years old, a housewife, already has basic life insurance coverage, but with changes in her family's financial situation, she hopes to upgrade her life insurance coverage.

[0139] Workflow:

[0140] User interaction: Ms. Zhang inputs her basic information (age, occupation, family situation, etc.) and her needs for upgrading insurance (such as increasing the coverage amount, expanding the scope of coverage, etc.) through the mobile application.

[0141] Dynamic prompt generation: The system analyzes Ms. Zhang's information and combines it with market dynamics in the knowledge base (such as newly launched life insurance products, rate adjustments, etc.).

[0142] Generate dynamic prompts: "Recommend a high-coverage, multi-protection life insurance product for a 45-year-old housewife who wants to upgrade her life insurance coverage."

[0143] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0144] Based on prompts and built-in algorithms, the large model generates a recommended list of various upgraded life insurance products, along with a detailed comparison and upgrade suggestions for each product.

[0145] Recommendation Results Output: The system sorts the recommendation list and prioritizes suitable products based on Ms. Zhang's budget and needs. A detailed upgrade plan report for life insurance is generated and displayed to Ms. Zhang through the user interaction module.

[0146] Outcome: Through this life insurance upgrade plan report, Ms. Zhang successfully upgraded her life insurance coverage, increased the sum insured, and expanded the scope of coverage.

[0147] Scenario 3: Life insurance consultation for elderly retirees.

[0148] Background: Mr. Wang, 65 years old, retired, hopes to reduce the financial burden on his children through life insurance.

[0149] Workflow:

[0150] User interaction: Mr. Wang inputs his basic information (age, occupation, economic status, etc.) and insurance consultation needs (such as wanting to learn about life insurance products suitable for the elderly) through the web interface.

[0151] Dynamic prompt word generation: The system analyzes Mr. Wang's information and combines it with knowledge of elderly insurance and market dynamics from the knowledge base.

[0152] Generate dynamic prompts: "Recommend a life insurance product suitable for the financial situation and needs of a 65-year-old retired senior citizen."

[0153] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0154] The large model generates a recommendation list of various life insurance products suitable for the elderly based on prompts and built-in algorithms, along with the features and purchase advice for each product.

[0155] Recommendation Results Output: The system filters the recommendation list, prioritizing products that offer good value for money and are suitable for seniors. A detailed life insurance consultation report is generated and displayed to Mr. Wang through the user interaction module.

[0156] Outcome: Through this life insurance consultation report, Mr. Wang learned about various life insurance products suitable for the elderly and chose the appropriate product to purchase based on his needs.

[0157] Scenario 4: Comprehensive protection plan for entrepreneurs.

[0158] Background: Mr. Zhao, 35 years old, is an entrepreneur whose company is in a phase of rapid development. He has a need for comprehensive life insurance coverage and also wants to take into account potential future risks.

[0159] Workflow:

[0160] User interaction: Mr. Zhao inputs his basic information (age, occupation, company status, income, etc.) and comprehensive protection needs (such as wanting to cover multiple risks such as illness, accidents, and death) through the system.

[0161] Dynamic prompt generation: The system analyzes Mr. Zhao's information, taking into account his special status as an entrepreneur and the risks he may face in the future (such as business failure, health problems, etc.).

[0162] Combining entrepreneur insurance knowledge and market dynamics from the knowledge base, a dynamic prompt is generated: "Recommend a life insurance product that comprehensively covers the risks of illness, accident, and death for a 35-year-old man in the startup phase, and consider possible future risk factors."

[0163] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0164] Based on prompts and built-in algorithms, the big model generates a recommended list of comprehensive life insurance products, along with detailed coverage information, premium comparisons, and future risk coverage suggestions for each product.

[0165] Recommendation Results Output: The system comprehensively analyzes the recommendation list, taking into account Mr. Zhao's financial situation and risk tolerance, and prioritizes recommending product combinations suitable for him. A comprehensive protection plan report is generated and presented to Mr. Zhao through the user interaction module, along with follow-up consultation and purchase services.

[0166] Outcome: Through this comprehensive protection planning report, Mr. Zhao learned about various life insurance product combinations suitable for entrepreneurs and chose the most suitable product to purchase according to his needs, providing solid protection for his entrepreneurial journey.

[0167] Scenario 5: Recommendations for combining life insurance and financial planning for the breadwinner of the family.

[0168] Background: Ms. Chen, 40 years old, is the breadwinner of her family. She has a stable job and income and hopes to purchase life insurance while also meeting her financial planning needs.

[0169] Workflow:

[0170] User interaction: Ms. Chen inputs her basic information (age, occupation, family situation, income, etc.) and her needs for combining life insurance with financial management (such as hoping that life insurance products have a certain investment return) through the system.

[0171] Dynamic prompt generation: The system analyzes Ms. Chen's information, taking into account her responsibilities as the breadwinner of the family and her financial needs.

[0172] By combining life insurance and financial planning product knowledge with market dynamics in the knowledge base, a dynamic prompt is generated: "Recommend a life insurance product that combines protection and investment returns for a 40-year-old woman who is the breadwinner of her family."

[0173] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0174] Based on prompts and built-in algorithms, the large model generates a recommendation list containing various life insurance and wealth management products, along with detailed information such as the coverage, investment return rate, and risk level of each product.

[0175] Recommendation Results Output: The system comprehensively evaluates the recommendation list, taking into account Ms. Chen's risk tolerance and financial goals, and prioritizes products suitable for her. A combined life insurance and wealth management recommendation report is generated and displayed to Ms. Chen through the user interaction module, providing subsequent purchase and wealth management services.

[0176] Outcome: Through this recommendation report combining life insurance and financial planning, Ms. Chen learned about a variety of life insurance products that combine protection and investment returns. She then chose the most suitable product based on her needs, achieving the dual goals of life insurance protection and financial planning.

[0177] Scenario 6: Joint life insurance planning for newlyweds.

[0178] Background: Mr. Lin and Ms. Lin, both 30 years old, are a newlywed couple who want to provide financial security for each other's future life and are considering purchasing joint life insurance.

[0179] Workflow:

[0180] User interaction: Newlyweds jointly input their basic information (age, occupation, income, family situation, etc.) and joint life insurance needs (such as wanting to protect their lives, children's education, etc.) through the system.

[0181] Dynamic prompt generation: The system analyzes the information of newlyweds, taking into account their special status as newlyweds and the life changes they may face in the future (such as the birth of children, changes in family income, etc.).

[0182] By combining knowledge of joint life insurance from the knowledge base with market dynamics, a dynamic prompt is generated: "Recommend a joint life insurance product for a newlywed couple aged 30 that protects both of their lives and takes into account their children's education."

[0183] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0184] Based on prompts and built-in algorithms, the big model generates a recommendation list containing multiple joint life insurance products, along with detailed information such as the coverage, premium comparison, and claims process for each product.

[0185] Recommendation Results Output: The system comprehensively evaluates the recommendation list, taking into account the newlyweds' financial situation and insurance needs, and prioritizes recommending suitable joint life insurance products for them. A joint life insurance planning report is generated and presented to the newlyweds through the user interaction module, along with follow-up consultation and purchase services.

[0186] Outcome: Through this joint life insurance planning report, the newlyweds learned about a variety of joint life insurance products suitable for them and chose the most appropriate product to purchase based on their needs, providing solid financial security for their future life together.

[0187] Scenario 7: Personalized life insurance plans for individuals with strong health awareness.

[0188] Background: Ms. Zhou, 38 years old, is health-conscious and values ​​quality of life. She hopes to customize a life insurance plan based on her health condition and lifestyle.

[0189] Workflow:

[0190] User interaction: Ms. Zhou inputs her basic information (age, occupation, income, family situation, etc.), health status (such as whether she has chronic diseases, exercise habits, etc.) and customized life insurance needs through the system.

[0191] Dynamic prompt generation: The system analyzes Ms. Zhou's information, taking into account the impact of her health condition and lifestyle habits on her life insurance needs.

[0192] By combining customized life insurance knowledge and market dynamics from the knowledge base, a dynamic prompt is generated: "Recommend a life insurance product customized based on the health condition and lifestyle of a 38-year-old woman with a strong health awareness."

[0193] Large-scale model inference: The system sends the prompt words to the large model for inference.

[0194] Based on prompts and built-in algorithms, the big model generates a recommendation list containing various customized life insurance products, along with detailed information such as the coverage, premium calculation method, and health disclosure requirements for each product.

[0195] Recommendation Results Output: The system comprehensively evaluates the recommended list, taking into account Ms. Zhou's financial situation, health condition, and insurance needs, and prioritizes recommending customized life insurance products suitable for her. A customized life insurance plan report is generated and presented to Ms. Zhou through the user interaction module, along with follow-up consultation and purchase services.

[0196] Outcome: Through this customized life insurance plan report, Ms. Zhou learned about a variety of life insurance products suitable for her and chose the most suitable product to purchase according to her needs, achieving a match between life insurance protection and her own health condition and lifestyle.

[0197] Figure 5 This is a schematic diagram of the structure of a product recommendation device 500 provided in an embodiment of this disclosure. Figure 5 As shown, the device includes: a first processing unit 510, configured to generate prompt words based on the user's static information and behavioral information, wherein the behavioral information is obtained by feature extraction of the user's operations using a first network; and a second processing unit 520, configured to generate a recommendation list corresponding to the user based on the prompt words, wherein the recommendation list includes at least one recommended product.

[0198] In some embodiments, the first processing unit is further configured to generate prompt words based on the user's static information and behavioral information, including: obtaining industry knowledge information and market dynamic information from a knowledge base; performing data analysis on the static information and behavioral information based on the industry knowledge information and market dynamic information to determine the user's contextual information and demand information; determining a target template from a plurality of preset prompt word templates based on the contextual information and demand information; and generating prompt words based on the target template, contextual information, and demand information.

[0199] In some embodiments, the second processing unit is further configured to retrieve customer cases related to the prompt words from the knowledge base; and generate a recommendation list corresponding to the user based on the customer cases and information on at least one candidate product included in the knowledge base.

[0200] In some embodiments, the product recommendation device further includes a third processing unit. The third processing unit is configured to: obtain updated content in response to an automatic update instruction or a user update instruction, wherein the automatic update instruction is triggered when preset conditions are met, and the user update instruction is triggered in response to a user's update operation, wherein the preset conditions include at least one of detecting an update to product information on a first website and determining that the fluctuation range of market data exceeds a preset threshold; update the knowledge base according to the updated content; and update the version information of the knowledge base.

[0201] In some embodiments, the third processing unit is further configured to, in response to an automatic update instruction, obtain updated content from a first website, the updated content including updated product information; in response to an automatic update instruction, compare current market data with data in a knowledge base to obtain updated content; and in response to a user update instruction, obtain updated content from information input by the user.

[0202] In some embodiments, the knowledge base includes a cache area and a non-cache area, wherein the cache area includes at least one of the following: information on at least one hot candidate product, at least one historical user profile information, and at least one historical customer case.

[0203] In summary, the product recommendation device 500 can dynamically generate prompts based on users' static and behavioral information. Then, it can generate a recommendation list for each user based on the dynamically generated prompts, which can improve the accuracy of product recommendations and enhance the user experience.

[0204] The methods and apparatus provided in the embodiments of this application have been described above. To implement the functions of the methods provided in the embodiments of this application, the electronic device may include a hardware structure and software modules, and may implement the above functions in the form of a hardware structure, software modules, or a hardware structure plus software modules. One of the above functions may be executed in the form of a hardware structure, software modules, or a hardware structure plus software modules.

[0205] Figure 6 This is a block diagram illustrating an electronic device 600 for implementing the above-described method according to an exemplary embodiment. For example, the electronic device 600 may be a mobile phone, computer, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc.

[0206] Reference Figure 6 The electronic device 600 may include one or more of the following components: a processing component 602, a memory 604, a power supply component 606, a multimedia component 608, an audio component 610, an input / output (I / O) interface 612, a sensor component 614, and a communication component 616.

[0207] Processing component 602 typically controls the overall operation of electronic device 600, such as operations associated with display, telephone calls, data communication, camera operation, and recording operations. Processing component 602 may include one or more processors 620 to execute instructions to perform all or part of the steps of the methods described above. Furthermore, processing component 602 may include one or more modules to facilitate interaction between processing component 602 and other components. For example, processing component 602 may include a multimedia module to facilitate interaction between multimedia component 608 and processing component 602.

[0208] Memory 604 is configured to store various types of data to support the operation of electronic device 600. Examples of this data include instructions for any application or method operating on electronic device 600, contact data, phonebook data, messages, pictures, videos, etc. Memory 604 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0209] Power supply component 606 provides power to various components of electronic device 600. Power supply component 606 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 600.

[0210] Multimedia component 608 includes a screen that provides an output interface between electronic device 600 and user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 608 includes a front-facing camera and / or a rear-facing camera. When electronic device 600 is in an operating mode, such as a shooting mode or video mode, the front-facing camera and / or rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0211] Audio component 610 is configured to output and / or input audio signals. For example, audio component 610 includes a microphone (MIC) configured to receive external audio signals when electronic device 600 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 604 or transmitted via communication component 616. In some embodiments, audio component 610 also includes a speaker for outputting audio signals.

[0212] I / O interface 612 provides an interface between processing component 602 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0213] Sensor assembly 614 includes one or more sensors for providing state assessments of various aspects of electronic device 600. For example, sensor assembly 614 may detect the on / off state of electronic device 600, the relative positioning of components such as the display and keypad of electronic device 600, changes in position of electronic device 600 or a component of electronic device 600, the presence or absence of user contact with electronic device 600, orientation or acceleration / deceleration of electronic device 600, and temperature changes of electronic device 600. Sensor assembly 614 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 614 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 614 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0214] Communication component 616 is configured to facilitate wired or wireless communication between electronic device 600 and other devices. Electronic device 600 can access wireless networks based on communication standards, such as WiFi, 2G or 3G, 4G LTE, 5G NR (NewRadio), or combinations thereof. In one exemplary embodiment, communication component 616 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 616 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0215] In an exemplary embodiment, the electronic device 600 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0216] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 604 including instructions, which can be executed by a processor 620 of an electronic device 600 to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0217] Embodiments of this disclosure also provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods described in the above embodiments of this disclosure.

[0218] Figure 7 This is a schematic diagram illustrating the structure of a chip 700 for implementing the above method according to an exemplary embodiment. (Refer to...) Figure 7 The chip 700 includes a communication interface 701 and at least one processor 702. The communication interface 701 is used to receive signals input to the chip 700 or signals output from the chip 700. The processor 702 communicates with the communication interface 701 and implements the methods described in the above embodiments of this disclosure through logic circuits or executing code instructions.

[0219] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0220] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in at least one embodiment or example.

[0221] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0222] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processing module, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having at least one wiring (control method), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0223] It should be understood that various parts of the embodiments of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0224] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0225] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc.

[0226] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A product recommendation method, characterized in that, The method further includes: Based on the user's static and behavioral information, prompt words are generated, wherein the behavioral information is obtained by feature extraction of the user's operations using a first network. Based on the prompt words, a recommendation list corresponding to the user is generated, and the recommendation list includes at least one recommended product.

2. The method according to claim 1, characterized in that, The process of generating prompts based on the user's static and behavioral information includes: Obtain industry knowledge and market dynamics information from the knowledge base; Based on the industry knowledge and market dynamics, data analysis is performed on the static and behavioral information to determine the contextual and demand information corresponding to the user. Based on the context information and the requirement information, a target template is determined from a set of preset prompt word templates; The prompt words are generated based on the target template, the context information, and the requirement information.

3. The method according to claim 2, characterized in that, The step of generating the recommendation list corresponding to the user based on the prompt words includes: Based on the prompt word, retrieve customer cases related to the prompt word from the knowledge base; Based on the customer case and information on at least one candidate product included in the knowledge base, a recommendation list corresponding to the user is generated.

4. The method according to claim 3, characterized in that, The method further includes: In response to an automatic update command or a user update command, the updated content is obtained. The automatic update command is triggered when preset conditions are met, and the user update command is triggered in response to a user's update operation. The preset conditions include at least one of the following: detecting an update to product information on a first website and determining that the fluctuation range of market data exceeds a preset threshold. The knowledge base is updated according to the updated content; Update the version information of the knowledge base.

5. The method according to claim 4, characterized in that, The response to an automatic update command or a user update command, the acquisition of update content includes at least one of the following: In response to the automatic update instruction, the updated content is obtained from the first website, the updated content including the updated product information; In response to the automatic update command, the current market data is compared with the data in the knowledge base to obtain the updated content; In response to the user update instruction, the update content is obtained from the information input by the user.

6. The method according to claim 3, characterized in that, The knowledge base includes a cache area and a non-cache area. The cache area includes at least one of the following: information on at least one hot candidate product, at least one historical user profile, and at least one historical customer case.

7. A product recommendation device, characterized in that, The device includes: The first processing unit is used to generate prompt words based on the user's static information and behavioral information, wherein the behavioral information is obtained by feature extraction of the user's operation using the first network; The second processing unit is used to generate a recommendation list corresponding to the user based on the prompt words, the recommendation list including at least one recommended product.

8. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

10. A chip, characterized in that, It includes at least one processor and a communication interface; the communication interface is used to receive signals input to the chip or signals output from the chip, and the processor communicates with the communication interface and implements the method as described in any one of claims 1 to 6 through logic circuits or executing code instructions.

Images