Product recommendation method and device, equipment and storage medium

Abstract

The invention provides a product recommendation method and device, equipment and a storage medium, and the method comprises the steps: receiving strategy configuration sent by a configuration center, analyzing the strategy configuration, and obtaining strategy parameters of the strategy configuration; determining a filtering algorithm according to the strategy parameters; receiving user information of a target user, and filtering products according to the filtering algorithm and the user information to obtain a plurality of first products meeting filtering conditions; a binary tree is generated according to the strategy parameters and the first products, and nodes of the binary tree are the strategy parameters or the first products; and traversing each node of the binary tree based on a recursive mode to obtain a recommended product.

Description

Technical Field

[0001] The present invention relates to the field of Internet technology, and in particular to a product recommendation method, device, equipment and storage medium. Background Art

[0002] With the development of products, there are more and more product types, and users are often faced with many product choices. However, due to their lack of understanding of the products, they are unable to choose the product that best suits their needs. At the same time, existing technologies often recommend products to users based on the highest price or maximizing sales targets, rather than based on the users' own needs. Therefore, the existing market lacks an objective and professional product recommendation mechanism to ensure that users can choose the products that best suit them.

[0003] Therefore, in view of the shortcomings of the prior art solutions, the present invention provides a product recommendation method. Summary of the invention

[0004] Based on this, it is necessary to provide a product recommendation method, device, equipment and storage medium to address the above technical problems.

[0005] On the one hand, a product recommendation method is provided, the method comprising: receiving a policy configuration sent by a configuration center, parsing the policy configuration, and obtaining policy parameters of the policy configuration; determining a filtering algorithm based on the policy parameters; receiving user information of a target user, filtering products based on the filtering algorithm and the user information, and obtaining a plurality of first products that meet the filtering conditions; generating a binary tree based on the policy parameters and the plurality of first products, wherein the nodes of the binary tree are policy parameters or first products; and recursively traversing each node of the binary tree to obtain recommended products.

[0006] Optionally, before receiving the policy configuration sent by the configuration center, the method further includes: receiving policy parameters configured by the configuration user in the visual interface; and storing the policy parameters as the policy configuration in the configuration center.

[0007] Optionally, filtering the products according to the filtering algorithm and the user information to obtain multiple first products that meet the filtering conditions includes: obtaining a score assigned to each product by the target user according to the filtering algorithm and the user information; and comparing the score assigned to each product by the target user with a preset standard score to obtain multiple first products whose scores are greater than or equal to the standard score.

[0008] Optionally, obtaining the score of the target user for each product according to the filtering algorithm includes: obtaining the first score of the target product by the target user through collaborative filtering; obtaining the matching score of the target user and the target product according to the user information of the target user and the product information of the target product; obtaining the score of the target product by the target user according to the first score of the target product by the target user, the matching score and a preset weight coefficient.

[0009] Optionally, obtaining the first score of the target product by the target user through collaborative filtering includes: determining, according to the strategy parameters, that the collaborative filtering is user collaborative filtering or product collaborative filtering; when the collaborative filtering is user collaborative filtering, calculating the similarity between the target user and historical users, obtaining similar users who are similar to the target user among the historical users, and determining the first score of the target product by the target user according to the scores of the target product by the similar users, wherein the similar users may be multiple users; when the collaborative filtering is product collaborative filtering, calculating the similarity between the target product and historically evaluated products, obtaining similar products that are similar to the target product among the historically evaluated products, and determining the first score of the target product by the target user according to the score of the similar products by the target user, wherein the similar products may be multiple products.

[0010] Optionally, generating a binary tree according to the strategy parameters and the multiple first products includes: determining a root node of the binary tree according to the strategy parameters; adding child nodes layer by layer according to an order of recommended strategies in the strategy parameters, wherein the newly added child nodes are the recommended strategies or the first products.

[0011] Optionally, the method further includes: monitoring the target user's processing of the recommended product; analyzing the policy configuration according to the processing; and adjusting the policy configuration according to the analysis result.

[0012] On the other hand, a product recommendation device is provided, which includes: a receiving module, used to receive a policy configuration sent by a configuration center, parse the policy configuration, and obtain policy parameters of the policy configuration; a determination module, used to determine a filtering algorithm based on the policy parameters; a filtering module, used to receive information of target users, filter products according to the filtering algorithm, and obtain multiple first products that meet the filtering conditions; a generation module, used to generate a binary tree according to the policy parameters and the multiple first products, wherein the nodes of the binary tree are policy parameters or first products; and a recommendation module, used to traverse each node of the binary tree in a recursive manner to obtain recommended products.

[0013] On the other hand, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: receiving a policy configuration sent by a configuration center, parsing the policy configuration, and obtaining policy parameters of the policy configuration; determining a filtering algorithm based on the policy parameters; receiving user information of a target user, filtering products based on the filtering algorithm and the user information, and obtaining a plurality of first products that meet the filtering conditions; generating a binary tree based on the policy parameters and the plurality of first products, wherein the nodes of the binary tree are policy parameters or first products; and recursively traversing each node of the binary tree to obtain recommended products.

[0014] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: receiving a policy configuration sent by a configuration center, parsing the policy configuration, and obtaining policy parameters of the policy configuration; determining a filtering algorithm based on the policy parameters; receiving user information of a target user, filtering products based on the filtering algorithm and the user information, and obtaining multiple first products that meet the filtering conditions; generating a binary tree based on the policy parameters and the multiple first products, wherein the nodes of the binary tree are policy parameters or first products; and recursively traversing each node of the binary tree to obtain recommended products.

[0015] The above-mentioned product recommendation method, device, equipment and storage medium, the method includes: receiving the policy configuration sent by the configuration center, parsing the policy configuration, and obtaining the policy parameters of the policy configuration; determining the filtering algorithm according to the policy parameters; receiving the user information of the target user, filtering the products according to the filtering algorithm and the user information, and obtaining multiple first products that meet the filtering conditions; generating a binary tree according to the policy parameters and the multiple first products, wherein the nodes of the binary tree are policy parameters or first products; traversing each node of the binary tree in a recursive manner to obtain recommended products; in this way, based on user information, through policy configuration, accurate matching of users and products is achieved, and suitable product selection results are recommended to users. At the same time, the strategy can be flexibly and simply configured according to the needs of business operations, thereby achieving more efficient and more accurate product recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a process of a product recommendation method in one embodiment;

[0017] Figure 2 A binary tree diagram of a product recommendation method in one embodiment;

[0018] Figure 3A schematic diagram of execution code of a product recommendation method in one embodiment;

[0019] Figure 4 A schematic diagram of an interface configuration of a product recommendation method in one embodiment;

[0020] Figure 5 A schematic diagram of rule codes for a product recommendation method in one embodiment;

[0021] Figure 6 A schematic diagram of a process of a product recommendation method in one embodiment;

[0022] Figure 7 A schematic diagram of the structure of the interaction between the configuration management background and the business system of the product recommendation method in one embodiment;

[0023] Figure 8 A schematic diagram of a service architecture of a product recommendation method in one embodiment;

[0024] Fig. 9 A schematic diagram of an artificial intelligence recommendation process of a product recommendation method in one embodiment;

[0025] Fig.10 A schematic diagram of prompt words recommended by artificial intelligence in a product recommendation method in one embodiment;

[0026] Fig.11 is a structural block diagram of a product recommendation device in one embodiment;

[0027] Fig.12 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0029] It should be understood that in the description of the present invention, unless the context clearly requires otherwise, words such as “include”, “including” and the like throughout the specification should be interpreted as including rather than being exclusive or exhaustive; that is, as including but not limited to “including”.

[0030] It should also be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0031] It should be noted that the terms "S1", "S2", etc. are only used for the purpose of describing the steps, and do not specifically refer to the order or sequence, nor are they used to limit the present invention. They are only for the convenience of describing the method of the present invention, and cannot be understood as indicating the order of the steps. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0032] In one embodiment, Figure 1 As shown, a product recommendation method is provided, comprising the following steps:

[0033] S101: receiving a policy configuration sent by a configuration center, parsing the policy configuration, and obtaining policy parameters of the policy configuration.

[0034] Here, the policy parameters may include policy ID, policy name, usage scenario, policy type, recommended policy, filtering algorithm, etc.

[0035] Specifically, the configuration center receives the policy configuration selected by the user, and sends the selected policy configuration to the server. The server receives the policy configuration and parses it to obtain policy parameters.

[0036] S102: Determine a filtering algorithm according to the policy parameters.

[0037] Here, the policy parameter can be recommendation based on user groups, and the corresponding filtering algorithm is filtering based on user information; the policy parameter can also be recommendation based on product type, and the corresponding filtering algorithm is filtering based on product information. For example, when selecting recommendation by product type, the collaborative filtering in the filtering algorithm can be product collaborative filtering, and when selecting recommendation by user attributes, the collaborative filtering in the filtering algorithm can be user collaborative filtering.

[0038] S103: receiving user information of a target user, filtering products according to the filtering algorithm and the user information, and obtaining a plurality of first products that meet the filtering conditions.

[0039] Here, the user information of the target user may be the user information obtained by collecting data on the target user's browsing history on a web page, or may be the personal information filled in by the target user.

[0040] Specifically, from the massive amount of products, based on the filtering algorithm, the products are roughly sorted, and a portion of products with high similarity are screened out to narrow the product range.

[0041] S104: Generate a binary tree according to the strategy parameters and the multiple first products, wherein the nodes of the binary tree are strategy parameters or first products.

[0042] Among them, a binary tree is an ordered tree in which the degree of each node in the tree is not greater than 2.

[0043] S105: Traverse each node of the binary tree in a recursive manner to obtain recommended products.

[0044] Here, we traverse each node recursively to determine the rule execution result until we encounter a leaf node, then terminate and return the leaf node. A leaf node is a node with a degree of 0.

[0045] For example, Figure 2 Schematic diagram of a binary tree in an embodiment of the present invention, such as Figure 2 As shown, the binary tree nodes include: 15 nodes from A to O, and the traversal results are A, B, D, H, I, E, J, K, C, F, L, M, G, N, O.

[0046] The traversal execution statement may use an Aviator expression, for example, Figure 3 Schematic diagram of the execution code in the embodiment of the present invention, such as Figure 3 As shown, the execution code includes: Private static Object traverse(NoderuleNode,Map<String,Object> param){

[0047] if(ObjectUtil.isNull(ruleNode)){

[0048] log.info(msg:"traverse param ruleNode is empty!");

[0049] return null;

[0050] }

[0051] String ruleExec=ruleNode.getRuleExec();

[0052] if(StringUtils.isBlank(ruleExec)){

[0053] log.warn(format:"traverse param ruleNode:{}ruleExec is empty!",GsonUtil.toJsonString(ruleNode));

[0054] return null;

[0055] }

[0056] boolean matched=(boolean)AviatorEvaluator.execute(ruleExec,param,true);

[0057] if(matched){

[0058] if(ObjectUtil.isNull(ruleNode.getLeft())){

[0059] return ruleNode.getResult();

[0060] }

[0061] return traverse(ruleNode.getLeft(),param);

[0062] }else{

[0063] return traverse(ruleNode.getRight(),param);

[0064] }

[0065] }

[0066] Specifically, from a small range of products, according to operating rules and strategies, and based on the target users' own attributes, labels, behaviors, etc., accurately select products that best suit the target users.

[0067] It should be noted that the present invention realizes accurate matching of users and products based on user information through strategy configuration, and recommends suitable product selection results to users. At the same time, the strategy can be flexibly and simply configured according to the needs of business operations, so as to achieve more efficient and accurate product recommendations, and exchange data through the configuration center, which can ensure the decoupling between the configuration background and the high QPS (Queries Per Second) business system.

[0068] In some specific implementations, before receiving the policy configuration sent by the configuration center, the method further includes:

[0069] Receive the policy parameters configured by the configuration user in the visual interface;

[0070] The policy parameters are stored in the configuration center as policy configuration.

[0071] Here, the configuration user can configure policy parameters such as policy ID, policy name, usage scenario, policy type and recommended policy in the visual interface. Policy types can include recommendation by product type, recommendation by specific product, recommendation by policyholder attributes and AIGC (Artificial Intelligence Generative Content).

[0072] The policy ID or policy name may be used as a policy identifier, and by inputting the policy identifier, a corresponding policy configuration pre-configured and stored in the configuration center can be matched.

[0073] When the configuration user updates some policy parameters in the policy configuration in the visual interface, the corresponding policy configuration in the configuration center is updated synchronously.

[0074] In some embodiments, assuming that the product is an insurance product, when the strategy type selects recommendation by product type, the corresponding recommendation strategy includes order and weight, where the order can include multiple product types: medical insurance, health insurance, critical illness insurance, accident insurance, property insurance, endowment insurance, life insurance and auto insurance, etc. After selecting a product type, you can select the weight of at least one specific product; when the strategy type selects recommendation by specific product, the corresponding recommendation strategy includes order, where the order can include specific products, such as outpatient medical insurance, personal hospitalization medical insurance of 10,000 yuan, etc.; when the strategy type selects recommendation by policyholder attributes, the corresponding recommendation strategy can include age, gender, permanent residence, work, etc., and you can choose to select by product, by type or by proportion, and select the weight of the product; when the strategy type selects recommendation by AIGC, the corresponding recommendation strategy includes order.

[0075] Here, a mapping table between products and logos can be pre-set, and different products can be represented by letters. For example, HNJCA can be used to represent Youbao·Family Comprehensive Protection (Basic Edition).

[0076] Specifically, Figure 4 is a schematic diagram of the interface configuration in an embodiment of the present invention, such as Figure 4As shown, the interface configuration includes: strategy ID, strategy name, usage scenario, and strategy type. For strategy type, select Recommend by policyholder attributes, select Age for recommended strategy, fill in 0 for starting age and 40 for ending age, select By product, select HNJCA product for weight 1, select GRZJA product for weight 2, select ZHZJA product for weight 3, and select ZHYWA product for weight 4.

[0077] Further, according to Figure 4 Configure the corresponding Aviator expression. Figure 5 is a schematic diagram of rule codes in an embodiment of the present invention, such as Figure 5 As shown, the rule code includes:

[0078]

[0079]

[0080] In this way, users can configure strategies from multiple dimensions on the visualization page.

[0081] In some specific implementations, filtering products according to the filtering algorithm and the user information to obtain a plurality of first products that meet the filtering conditions includes:

[0082] Obtaining the score of each product given by the target user according to the filtering algorithm and the user information;

[0083] The target user's score for each product is compared with a preset standard score to obtain a plurality of first products whose scores are greater than or equal to the standard score.

[0084] Here, the filtering algorithm may be determined according to the usage scenario in the configuration parameters, or may be manually determined by the user.

[0085] In some specific implementations, obtaining the target user's rating of each product according to the filtering algorithm includes:

[0086] Obtaining a first rating of the target product by the target user through collaborative filtering;

[0087] Obtaining a matching score between the target user and the target product according to the user information of the target user and the product information of the target product;

[0088] The score of the target product by the target user is obtained according to the first score of the target product by the target user, the matching score and a preset weight coefficient.

[0089] Here, the filtering algorithm may be a hybrid algorithm of collaborative filtering and content recommendation. Collaborative filtering may include user collaborative filtering and product collaborative filtering. User collaborative filtering is used to recommend products based on the behaviors of similar users, and product collaborative filtering is used to recommend products based on the popularity of similar products. Content recommendation is used to recommend products based on the matching degree between user features and product features.

[0090] Specifically, data preprocessing is performed on user data and product data. User data includes basic information of users (such as age, gender, income, etc.) and historical behavior data (such as browsing history, purchase history, etc.); product data includes basic information of products (such as product type, coverage, price, etc.) and product features (such as risk coverage, applicable population, etc.). Through the collaborative filtering algorithm, based on the preprocessed data, the collaborative filtering score of the user for the product is obtained. Through the content recommendation algorithm, based on the preprocessed data, feature vectors are established for the user and the product, and the matching degree between the user and the product is calculated as the content recommendation score. A weight coefficient is preset, and based on the weight, collaborative filtering score and content recommendation score, the user's final score for the product is obtained.

[0091] Furthermore, the matching degree between users and products can be calculated by the following formula:

[0092] score(u,p)=f u *f p (1)

[0093] Among them, f u represents the feature vector of user u, f p The feature vector representing the product p.

[0094] Furthermore, the results of collaborative filtering and content recommendation are combined to generate the final score, which can be calculated by the following formula:

[0095] final_score(u,p)=α*cf_score(u,p)+(1-α)*cb_score(u,p) (2)

[0096] Among them, α is the weight coefficient, cf_score(u,p) is the collaborative filtering score, cb_score(u,p) is the content recommendation score, and final_score(u,p) is the final score.

[0097] In this way, the advantages of collaborative filtering and content recommendation are combined, and the accuracy and diversity of recommendations are improved through a hybrid algorithm.

[0098] In some specific implementations, obtaining the first rating of the target product by the target user through collaborative filtering includes:

[0099] According to the strategy parameters, determining that the collaborative filtering is user collaborative filtering or product collaborative filtering;

[0100] When the collaborative filtering is user collaborative filtering, the similarity between the target user and the historical users is calculated to obtain similar users similar to the target user among the historical users, and a first score of the target user for the target product is determined according to the scores of the similar users for the target product, wherein the similar users may be multiple users;

[0101] When the collaborative filtering is product collaborative filtering, the similarity between the target product and the historically evaluated products is calculated to obtain similar products in the historically evaluated products that are similar to the target product, and the first score of the target product by the target user is determined based on the score of the similar product by the target user, wherein the similar product may be multiple products.

[0102] Here, the recommendation strategy in the strategy parameters can be determined as product collaborative filtering, in which the recommendation strategy is based on product recommendation, and the collaborative filtering can be determined as user collaborative filtering, in which the recommendation strategy is based on user recommendation.

[0103] Specifically, assuming that collaborative filtering is user collaborative filtering, the specific steps include calculating the similarity between users through cosine similarity values, and the calculation formula is as follows:

[0104]

[0105] Among them, r ui represents the rating of user u on product i, I uv Represents the set of products rated by user u and user v.

[0106] Based on the ratings of similar users, predict the target user's rating of a product. The calculation formula is as follows:

[0107]

[0108] in, represents the predicted rating of user u for product i, represents the average rating of user u, and N(u) represents the k users most similar to user u.

[0109] Specifically, assuming that the collaborative filtering is product collaborative filtering, the specific steps include calculating the similarity between products through the cosine similarity value, and the calculation formula is as follows:

[0110]

[0111] Among them, r ui represents the rating of user u on product i, U ijrepresents the set of users who rated both product i and product j.

[0112] Based on the ratings of similar products, predict the target user's rating of a product. The calculation formula is as follows:

[0113]

[0114] in, represents the predicted rating of user u for product i, and N(i) represents the k products most similar to product i.

[0115] In some embodiments, the performance can be evaluated based on indicators such as root mean square error (RMSE) and recall rate (Recall), and the specific formula is as follows:

[0116]

[0117] In some embodiments, after obtaining the user's rating of each product, the products are rearranged from high to low according to the rating.

[0118] In some specific implementations, generating a binary tree according to the strategy parameter and the plurality of first products includes:

[0119] Determining the root node of the binary tree according to the strategy parameters;

[0120] According to the order of the recommended strategies in the strategy parameters, sub-nodes are added layer by layer, wherein the newly added sub-nodes are the recommended strategies or the first product.

[0121] Specifically, assuming that the recommendation strategy includes multiple screening conditions, including condition 1, condition 2, condition 3 and condition 4, the root node of the binary tree is condition 1, the left child node in the second layer is condition 2, and the right child node does not meet condition 1. The left node under the condition 2 child node in the third layer can be condition 3, and the right node under the condition 2 child node can be not meeting condition 2. The left node under the child node that does not meet condition 1 in the third layer can be condition 4, and the left node under the condition 3 child node in the fourth layer can be the target product.

[0122] In some embodiments, the method further comprises:

[0123] Monitoring the target user's handling of the recommended product;

[0124] Analyzing the policy configuration according to the processing situation;

[0125] According to the analysis result, the policy configuration is adjusted.

[0126] Specifically, assuming the product is an insurance product, it will be summarized based on the user's subsequent bill of lading, payment, or underwriting behavior, provide strategy effect analysis capabilities, and be able to make real-time adjustments.

[0127] In one embodiment, Figure 6 Schematic diagram of the process in the embodiment of the present invention, as shown in Figure 6 As shown, the process of the present invention includes three parts. The first part is the insurance rule policy management background, the second part is the insurance rule policy core, namely the policy algorithm, policy execution, coarse sorting, fine sorting, and includes the recommendation capability based on the AIGC big model, and the third part is the recommendation result recovery and performance analysis.

[0128] Specifically, the insurance rule strategy management backend includes: 2.1 information synchronization in the product center, synchronizing the product configuration to the knowledge base, and in the configuration center includes: 3.1 associated strategy products, 3.2 configuration of auxiliary information and display pictures, and 3.3 configuration of recommended jump links.

[0129] Specifically, the core of insurance rule strategy includes AIGC recommendation and rule engine recommendation. For AIGC recommendation, the specific process includes 1.2 prompts processing, 1.2.1 AIGC customizes the plan based on the product information in its own knowledge base and the collected user data, 1.2.2 Generates customization reasons in multiple dimensions, 1.2.3 Summarizes the main product information such as terms and conditions based on the knowledge base; for rule engine recommendation, the specific process includes 1.3.1 pre-underwriting, 1.3.2 customer group labels, 1.3.3 in-insurance data statistics, 1.3.4 rule engine execution, and 1.3.5 generates recommendation reasons in combination with AIGC.

[0130] Specifically, the recommended result recovery and performance analysis include 1.4 displaying the reasons for customization and the best matching products, terms and conditions, etc., and 1.5 returning the insurance products; performing point embedding and log statistics, and returning to the data warehouse through XFLOW, or entering the data warehouse through nuclear insurance conversion.

[0131] In one embodiment, Figure 7 FIG. 1 is a schematic diagram of the structure of the interaction between the configuration management backend and the business system in an embodiment of the present invention. Figure 7 As shown in the figure, the system-level configuration management backend is isolated from the business system, and data is exchanged through the Zookeeper configuration center to ensure the decoupling between the configuration backend and the high QPS business system.

[0132] Specifically, CRUD (Create, Read, Update, Delete) operations are performed on the retail financial platform to configure rules, policies, products, and engines. The above configurations are stored in the local cache and database. At the same time, the synchronization strategy stores them in Zookeeper (configuration center). Zookeeper sends the configuration to the synchronization operation cache, which refreshes the rule cache, updates asynchronous loading, eviction strategy, refresh, and statistics.

[0133] Specifically, the business system uses the feign client to query the local proxy through the policy engine to perform tracking, degradation and verification, and calculates rules through AviatorEvaluator and DroolsEvaluator.

[0134] In one embodiment, Figure 8 Schematic diagram of the service architecture in an embodiment of the present invention. Figure 8 As shown, the service architecture of the present invention includes: an API interface, a policy engine and rule execution. Configuration parameters are received through the API interface, and parameter verification, configuration query, process control and tracking, and exception handling are performed through the policy engine. The rule execution specifically includes parsing, compilation and calculation, wherein the parsing part performs expression parsing and format conversion, the compilation part performs pre-compilation, pre-loading and pre-processing, and the calculation part performs expression evaluation and value formatting.

[0135] In one embodiment, when the strategy type is selected as AIGC, AIGC relies on the GPT (Generative Pre-Training Transformer) large model technology, and the user information after natural language processing will be further processed through the prompts template. AIGC will combine the content of the knowledge base and the processed user information to perform a customized product recommendation process. The process of product recommendation through AIGC specifically includes: 1) User information processing: through natural language processing technology, parse and understand the user's input information, extract user information from the user's natural language description, and ensure the integrity and accuracy of the information. 2) Information processing: The extracted user information is further optimized and processed through a specific prompts template, and the user information is converted into a format suitable for the recommendation system by using the predefined template and the powerful capabilities of the large model. 3) Knowledge base combination: The processed user information is intelligently matched with the knowledge base content to ensure the accuracy and relevance of the recommendation. 4) Customized recommendation: Combine user information and knowledge base content to perform a customized product recommendation process, and generate highly matched product or service recommendation solutions based on the specific needs and preferences of users. 5) Result feedback and optimization: The recommendation results will be displayed to users and continuously optimized and improved through user feedback to ensure the continuous improvement of the recommendation system and maximize user satisfaction. Fig. 9 FIG. 1 is a schematic diagram of an artificial intelligence recommendation process in an embodiment of the present invention. Fig. 9 As shown, the artificial intelligence recommendation process of the present invention includes: inputting information, determining the knowledge base to be queried according to conditions, and obtaining output results by querying the knowledge base in combination with prompt word recommendations.

[0136] Specifically, Fig.10 Schematic diagram of prompt words recommended by artificial intelligence in an embodiment of the present invention, such as Fig.10 As shown, the prompt words of the present invention include: component words: medical insurance product recommendation, model type: GPT3.5, answer style: accurate, parsing method: text, output variable name: result, prompt words: 1. Expert: insurance product recommendation expert, 2. Language: Chinese, 3. Description: You are an insurance recommendation expert. Based on the provided user information and the insurance product information held, customize the recommendation of the insurance product that best meets the user's needs, 4. Skills: Data analysis ability: understand the user's financial and health data. Insurance knowledge: in-depth understanding of the characteristics of various insurance products. Logical reasoning: match user needs with insurance product characteristics. Scenario simulation: simulate possible risk scenarios for users. Communication and explanation skills: clearly explain the reasons for the recommendation. 5. Rules: provide personalized insurance recommendations. Optimize user insurance coverage.

[0137] It should be understood that although Figure 1-Figure 10The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1-Figure 10 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0138] In one embodiment, Fig.11 As shown, a product recommendation device is provided, and the device includes: a receiving module 1101, used to receive the policy configuration sent by the configuration center, parse the policy configuration, and obtain the policy parameters of the policy configuration; a determining module 1102, used to determine the filtering algorithm according to the policy parameters; a filtering module 1103, used to receive the information of the target user, filter the products according to the filtering algorithm, and obtain multiple first products that meet the filtering conditions; a generating module 1104, used to generate a binary tree according to the policy parameters and the multiple first products, wherein the nodes of the binary tree are policy parameters or first products; a recommending module 1105, used to traverse each node of the binary tree in a recursive manner to obtain recommended products.

[0139] As a preferred implementation, in an embodiment of the present invention, the device also includes a storage module, which is specifically used to: receive policy parameters configured by a configuration user in a visual interface; and store the policy parameters as policy configuration in the configuration center.

[0140] As a preferred implementation manner, in an embodiment of the present invention, the filtering module 1103 is specifically used to: obtain the score of each product given by the target user based on the filtering algorithm and the user information; compare the score of each product given by the target user with a preset standard score, and obtain multiple first products whose scores are greater than or equal to the standard score.

[0141] As a preferred implementation manner, in an embodiment of the present invention, the filtering module 1103 is specifically used to: obtain the first score of the target user for the target product through collaborative filtering; obtain the matching score of the target user and the target product based on the user information of the target user and the product information of the target product; obtain the score of the target user for the target product based on the first score of the target user for the target product, the matching score and a preset weight coefficient.

[0142] As a preferred implementation manner, in an embodiment of the present invention, the filtering module 1103 is specifically used to: determine whether the collaborative filtering is user collaborative filtering or product collaborative filtering based on the strategy parameters; when the collaborative filtering is user collaborative filtering, calculate the similarity between the target user and historical users, obtain similar users who are similar to the target user in the historical users, and determine the first score of the target product by the target user based on the scores of the target product by the similar users, wherein the similar users may be multiple users; when the collaborative filtering is product collaborative filtering, calculate the similarity between the target product and historically evaluated products, obtain similar products who are similar to the target product in the historically evaluated products, and determine the first score of the target product by the target user based on the score of the similar products by the target user, wherein the similar products may be multiple products.

[0143] As a preferred implementation, in an embodiment of the present invention, the generation module 1104 is specifically used to: determine the root node of the binary tree according to the strategy parameters; add child nodes layer by layer according to the order of recommended strategies in the strategy parameters, wherein the newly added child nodes are the recommended strategies or the first product.

[0144] As a preferred implementation, in an embodiment of the present invention, the device also includes an analysis module, which is specifically used to: monitor the target user's processing of the recommended product; analyze the policy configuration based on the processing; and adjust the policy configuration based on the analysis results.

[0145] For the specific definition of the product recommendation device, please refer to the definition of the product recommendation method above, which will not be repeated here. Each module in the above-mentioned product recommendation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0146] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.12As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a product recommendation method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0147] Those skilled in the art will understand that Fig.12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0148] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: S1: receiving a policy configuration sent by a configuration center, parsing the policy configuration, and obtaining policy parameters of the policy configuration; S2: determining a filtering algorithm based on the policy parameters; S3: receiving user information of a target user, filtering products based on the filtering algorithm and the user information, and obtaining a plurality of first products that meet the filtering conditions; S4: generating a binary tree based on the policy parameters and the first products, wherein the nodes of the binary tree are the policy parameters and the first products; S5: traversing each node of the binary tree recursively to obtain recommended products.

[0149] In one embodiment, when the processor executes the computer program, the following steps are also implemented: receiving the policy parameters configured by the configuration user in the visual interface; and storing the policy parameters as the policy configuration in the configuration center.

[0150] In one embodiment, when the processor executes the computer program, the following steps are also implemented: according to the filtering algorithm and the user information, the score of each product given by the target user is obtained; and the score of each product given by the target user is compared with a preset standard score to obtain multiple first products whose scores are greater than or equal to the standard score.

[0151] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining the first score of the target product by the target user through collaborative filtering; obtaining the matching score of the target user and the target product based on the user information of the target user and the product information of the target product; obtaining the score of the target product by the target user based on the first score of the target product by the target user, the matching score and a preset weight coefficient.

[0152] In one embodiment, when the processor executes the computer program, the following steps are also implemented: according to the strategy parameters, determining whether the collaborative filtering is user collaborative filtering or product collaborative filtering; when the collaborative filtering is user collaborative filtering, calculating the similarity between the target user and historical users, obtaining similar users who are similar to the target user among the historical users, and determining the first score of the target product by the target user according to the scores of the target product by the similar users, wherein the similar users may be multiple users; when the collaborative filtering is product collaborative filtering, calculating the similarity between the target product and historically evaluated products, obtaining similar products who are similar to the target product among the historically evaluated products, and determining the first score of the target product by the target user according to the score of the similar products by the target user, wherein the similar products may be multiple products.

[0153] In one embodiment, when the processor executes the computer program, the following steps are also implemented: determining the root node of the binary tree according to the strategy parameters; adding child nodes layer by layer according to the order of the recommended strategies in the strategy parameters, wherein the newly added child nodes are the recommended strategies or the first product.

[0154] In one embodiment, when the processor executes the computer program, the following steps are also implemented: monitoring the target user's processing of the recommended product; analyzing the policy configuration based on the processing; and adjusting the policy configuration based on the analysis results.

[0155] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: S1: receiving a policy configuration sent by a configuration center, parsing the policy configuration, and obtaining policy parameters of the policy configuration; S2: determining a filtering algorithm based on the policy parameters; S3: receiving user information of a target user, filtering products based on the filtering algorithm and the user information, and obtaining a plurality of first products that meet the filtering conditions; S4: generating a binary tree based on the policy parameters and the first products, wherein the nodes of the binary tree are the policy parameters and the first products; S5: recursively traversing each node of the binary tree to obtain recommended products.

[0156] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: receiving policy parameters configured by a configuration user in a visual interface; and storing the policy parameters as policy configuration in the configuration center.

[0157] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: according to the filtering algorithm and the user information, the score of each product given by the target user is obtained; and the score of each product given by the target user is compared with a preset standard score to obtain multiple first products whose scores are greater than or equal to the standard score.

[0158] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining the first score of the target product by the target user through collaborative filtering; obtaining the matching score of the target user and the target product based on the user information of the target user and the product information of the target product; obtaining the score of the target product by the target user based on the first score of the target product by the target user, the matching score and a preset weight coefficient.

[0159] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: according to the strategy parameters, determining whether the collaborative filtering is user collaborative filtering or product collaborative filtering; when the collaborative filtering is user collaborative filtering, calculating the similarity between the target user and historical users, obtaining similar users who are similar to the target user among the historical users, and determining the first score of the target product by the target user according to the scores of the target product by the similar users, wherein the similar users may be multiple users; when the collaborative filtering is product collaborative filtering, calculating the similarity between the target product and historically evaluated products, obtaining similar products who are similar to the target product among the historically evaluated products, and determining the first score of the target product by the target user according to the score of the similar products by the target user, wherein the similar products may be multiple products.

[0160] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: determining the root node of the binary tree according to the strategy parameters; adding child nodes layer by layer according to the order of the recommended strategies in the strategy parameters, wherein the newly added child nodes are the recommended strategies or the first product.

[0161] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: monitoring the target user's processing of the recommended product; analyzing the policy configuration based on the processing; and adjusting the policy configuration based on the analysis results.

[0162] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0163] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0164] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A product recommendation method, characterized in that: The method comprises: Receive the policy configuration sent by the configuration center, parse the policy configuration, and obtain policy parameters of the policy configuration; Determining a filtering algorithm according to the strategy parameters; receiving user information of a target user, filtering products according to the filtering algorithm and the user information, and obtaining a plurality of first products that meet the filtering conditions; Generate a binary tree according to the strategy parameters and the plurality of first products, wherein the nodes of the binary tree are strategy parameters or first products; Each node of the binary tree is traversed recursively to obtain recommended products.

2. The product recommendation method according to claim 1, characterized in that: Before receiving the policy configuration sent by the configuration center, the method further includes: Receive the policy parameters configured by the configuration user in the visual interface; The policy parameters are stored in the configuration center as policy configuration.

3. The product recommendation method according to claim 1, characterized in that: The filtering of products according to the filtering algorithm and the user information to obtain a plurality of first products meeting the filtering conditions includes: Obtaining the score of each product given by the target user according to the filtering algorithm and the user information; The target user's score for each product is compared with a preset standard score to obtain a plurality of first products whose scores are greater than or equal to the standard score.

4. The product recommendation method according to claim 3, characterized in that: The step of obtaining the target user's rating of each product according to the filtering algorithm includes: Obtaining a first rating of the target product by the target user through collaborative filtering; Obtaining a matching score between the target user and the target product according to the user information of the target user and the product information of the target product; The score of the target product by the target user is obtained according to the first score of the target product by the target user, the matching score and a preset weight coefficient.

5. The product recommendation method according to claim 4, characterized in that: The step of obtaining the first rating of the target product by the target user through collaborative filtering includes: According to the strategy parameters, determining that the collaborative filtering is user collaborative filtering or product collaborative filtering; When the collaborative filtering is user collaborative filtering, the similarity between the target user and the historical users is calculated to obtain similar users similar to the target user among the historical users, and a first score of the target user for the target product is determined according to the scores of the similar users for the target product, wherein the similar users may be multiple users; When the collaborative filtering is product collaborative filtering, the similarity between the target product and the historically evaluated products is calculated to obtain similar products in the historically evaluated products that are similar to the target product, and the first score of the target product by the target user is determined based on the score of the similar product by the target user, wherein the similar product may be multiple products.

6. The product recommendation method according to claim 1, characterized in that: The step of generating a binary tree according to the strategy parameter and the plurality of first products includes: Determining the root node of the binary tree according to the strategy parameters; According to the order of the recommended strategies in the strategy parameters, sub-nodes are added layer by layer, wherein the newly added sub-nodes are the recommended strategies or the first product.

7. The product recommendation method according to claim 1, characterized in that: The method further comprises: Monitoring the target user's handling of the recommended product; Analyzing the policy configuration according to the processing situation; According to the analysis result, the policy configuration is adjusted.

8. A product recommendation device, characterized in that: The device comprises: A receiving module, used to receive the policy configuration sent by the configuration center, parse the policy configuration, and obtain the policy parameters of the policy configuration; A determination module, used to determine a filtering algorithm according to the policy parameters; A filtering module, used to receive information of target users, filter products according to the filtering algorithm, and obtain a plurality of first products that meet the filtering conditions; A generating module, configured to generate a binary tree according to the strategy parameters and the plurality of first products, wherein the nodes of the binary tree are strategy parameters or first products; The recommendation module is used to traverse each node of the binary tree in a recursive manner to obtain recommended products.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images