Non-vehicle property product combination output method, device, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence and the field of financial science and technology, and discloses a non-vehicle property product combination output method, which comprises the following steps: constructing a customer portrait, and extracting customer portrait features; extracting insurance characteristics of each non-vehicle property insurance product; particles are initialized in a particle swarm optimization method, a plurality of particles are generated randomly, and each particle represents a non-vehicle property insurance combination scheme; performing matching degree analysis on the insurance feature of each particle and the customer portrait feature, and determining an initial individual optimal position of each particle; the speed and position of the particles are updated through iteration of a particle swarm optimization method, matching degree analysis is carried out again, and the individual optimal position and the global optimal position of each particle are updated; and selecting the non-vehicle property insurance combination scheme corresponding to the global optimal position as output. According to the method, the multi-dimensional features in the customer portraits are accurately analyzed, the actual demands of customers are matched, various non-vehicle property insurance are combined together, and individuation and natural connection of the combination scheme are ensured.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence technology and financial technology, and in particular to a method, device, equipment and storage medium for outputting a non-automotive property product portfolio. Background Art

[0002] As the domestic automobile-related insurance market gradually becomes saturated, the growth of automobile insurance premium income has entered a bottleneck period, and many property insurance companies are facing growth difficulties. Therefore, insurance companies are in urgent need of finding new insurance products as a breakthrough to maintain the sustained growth of their business. Against this background, non-auto property insurance has gradually become an important development direction for the insurance industry.

[0003] Non-auto property insurance covers a wide range of categories, including property insurance, health insurance, travel insurance, liability insurance, accident insurance, etc. Customers of different ages and with different needs have different needs for these insurance types. Compared with auto insurance, non-auto property insurance can not only meet the diverse needs of customers in their daily lives, but also cover a wider market group. Therefore, the rapid development of non-auto property insurance has provided insurance companies with new growth opportunities and has become a key area that the insurance industry urgently needs to break through.

[0004] However, existing property insurance services are mostly concentrated on a single category of insurance products, and the multi-insurance combination services that meet customer needs are not perfect.

[0005] Current insurance recommendations are mostly based on a single insurance type, which fails to fully consider the multi-dimensional needs of customers, resulting in customers having to purchase multiple insurances in different scenarios and lacking one-stop comprehensive protection services. The recommendation of a single insurance type not only increases the complexity of customer selection, but may also miss the customer's potential demand for other insurance types.

[0006] Even though a small number of insurance companies have tried to provide multi-insurance package services, due to differences in service staff's skills and inconsistent wording, the recommended insurance packages often fail to connect naturally, making customers feel that the products are rigid or repetitive, thereby reducing their trust and willingness to buy.

[0007] Existing property insurance recommendation systems usually rely on customers to manually select insurance products, and fail to fully utilize customers' personal information, living habits and behavioral data for intelligent analysis and matching. This lack of personalized recommendation can neither accurately meet customer needs nor cause customers to miss out on more suitable insurance products, affecting customer experience and business conversion rates. Summary of the invention

[0008] The main purpose of the present invention is to provide a non-auto property product portfolio output method, device, equipment and storage medium, aiming to solve the technical problem that the prior art lacks personalized non-auto property insurance combinations based on customer portraits and naturally connected multiple insurance types, making it difficult to effectively meet the multi-dimensional needs of customers.

[0009] To achieve the above object, the present invention provides a method for outputting a non-vehicle property product portfolio, comprising:

[0010] Analyze customers’ personal information, historical purchase records, and browsing behaviors, build customer profiles, and extract customer profile features including age, income level, and risk tolerance;

[0011] Extracting insurance features of each insurance product from the non-auto property insurance product database, wherein the insurance features include suitable population categories, insurance identification values, and coverage;

[0012] Initialize particles in the particle swarm optimization method and randomly generate multiple particles, each particle represents a non-auto property insurance combination plan;

[0013] Compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination plan, and determine the initial individual optimal position of each particle based on the matching degree analysis result;

[0014] In each iteration of the particle swarm optimization method, the speed and position of each particle in the particle swarm are updated, and the matching degree analysis is re-performed for each updated particle position;

[0015] According to the re-determined matching degree analysis results, the individual optimal position of each particle and the global optimal position of the particle group are updated;

[0016] After the iteration is completed, the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm is selected as the final output.

[0017] Furthermore, in order to achieve the above-mentioned purpose, the present invention provides a non-car property product combination output device, comprising:

[0018] The customer profile building module is used to analyze the customer's personal information, historical purchase records and browsing behavior, build a customer profile, and extract customer profile features including age, income level and risk tolerance;

[0019] An insurance feature extraction module is used to extract the insurance features of each insurance product from the non-auto property insurance product library, wherein the insurance features include the category of suitable people, the insurance identification value and the coverage;

[0020] A particle initialization module is used to initialize particles in the particle swarm optimization method and randomly generate multiple particles, each particle representing a non-auto property insurance combination plan;

[0021] A matching degree analysis module, used to compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination plan, and determine the initial individual optimal position of each particle based on the matching degree analysis result;

[0022] The particle swarm iteration update module is used to update the speed and position of each particle in the particle swarm during each iteration of the particle swarm optimization method, and re-analyze the matching degree of each updated particle position;

[0023] The optimal position updating module is used to update the individual optimal position of each particle and the global optimal position of the particle group according to the re-determined matching degree analysis result;

[0024] The final combination output module is used to select the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm as the final output after the iteration is completed.

[0025] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer device, which includes a memory, a processor, and a non-vehicle property product combination output program stored in the memory and executable on the processor, and the non-vehicle property product combination output program, when executed by the processor, implements the steps of the non-vehicle property product combination output method as described above.

[0026] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a non-vehicle property product combination output program is stored, and when the non-vehicle property product combination output program is executed by a processor, the steps of the non-vehicle property product combination output method described above are implemented.

[0027] Beneficial effects: The present invention relates to the fields of artificial intelligence technology and financial technology, and discloses a method for outputting a combination of non-auto property products. By analyzing the customer's personal information, historical purchase records and browsing behavior, a customer portrait is constructed and customer portrait features are extracted; the insurance features of each insurance product are extracted from the non-auto property insurance product library, including suitable population categories, insurance identification values ​​and coverage; particles are initialized in the particle swarm optimization method, and multiple particles are randomly generated, each particle representing a non-auto property insurance combination plan; the insurance features of each particle are matched with the customer portrait features to determine the initial individual optimal position of each particle; the speed and position of the particles are updated through the iteration of the particle swarm optimization method, and the matching degree analysis is re-performed to update the individual optimal position and the global optimal position of each particle; and finally the non-auto property insurance combination plan corresponding to the global optimal position is selected as the output. The present invention accurately analyzes the multi-dimensional features in the customer portrait, matches the actual needs of the customer, effectively combines a variety of non-auto property insurance together, ensures the personalization and natural connection of the combination plan, avoids product duplication and unnecessary selection complexity in the prior art, and thus significantly improves customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0029] Figure 1 A schematic diagram of an application environment of a method for outputting a combination of non-car property products in an embodiment of the present invention;

[0030] Figure 2 A flowchart of an embodiment of a method for outputting a non-car property product combination according to the present invention;

[0031] Figure 3 It is a functional module diagram of a preferred embodiment of the non-car property product combination output device of the present invention;

[0032] Figure 4 A schematic diagram of the structure of a computer device in one embodiment of the present invention;

[0033] Figure 5 FIG. 4 is another schematic diagram of the structure of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0034] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0035] The non-car property product combination output method provided by the embodiment of the present invention can be applied in the following aspects: Figure 1In the application environment, the user end communicates with the server end through the network. The server end can analyze the customer's personal information, historical purchase records and browsing behavior through the user end, build a customer portrait, and extract customer portrait features; extract the insurance features of each insurance product from the non-auto property insurance product library, including the suitable population category, insurance identification value and coverage; initialize particles in the particle swarm optimization method, randomly generate multiple particles, and each particle represents a non-auto property insurance combination plan; perform matching degree analysis on the insurance features of each particle and the customer portrait features to determine the initial individual optimal position of each particle; update the speed and position of the particles through the iteration of the particle swarm optimization method, and re-perform matching degree analysis to update the individual optimal position and the global optimal position of each particle; finally select the non-auto property insurance combination plan corresponding to the global optimal position as the output. The present invention accurately analyzes the multi-dimensional features in the customer portrait, matches the actual needs of the customer, effectively combines a variety of non-auto property insurance together, ensures the personalization and natural connection of the combination plan, avoids product duplication and unnecessary selection complexity in the prior art, and thus significantly improves customer satisfaction. Among them, the user end can be but not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The server side can be implemented by an independent server or a server cluster composed of multiple servers. The present invention is described in detail below through specific embodiments.

[0036] See also Figure 2 , Figure 2 This is a flowchart of an embodiment of a method for outputting a non-vehicle property portfolio provided by the present invention. It should be noted that although a logical sequence is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0037] like Figure 2 As shown, the non-car property product portfolio output method proposed by the present invention includes the following steps:

[0038] S10, analyzing the customer’s personal information, historical purchase records, and browsing behavior, constructing a customer profile, and extracting customer profile features including age, income level, and risk tolerance;

[0039] In this embodiment, the customer's personal information includes basic information such as age, gender, occupation, income level, marital status, etc. By analyzing this data, we can preliminarily judge the customer's risk preference and financial ability, and then provide basic information for insurance product recommendations. This process requires the information provided by the customer to be structured and stored in the system's database.

[0040] Personal information can be collected by customers registering on the insurance platform or filling out questionnaires, and the system stores this information in the database through the API interface. During the implementation process, it is necessary to ensure the accuracy and completeness of the data, and take data security measures to protect customer privacy.

[0041] Historical purchase records refer to the insurance products that customers have purchased, the time of purchase, the amount of premiums, the coverage, and other data. These records can reflect the customer's past consumption behavior and risk preferences, and help predict their future insurance needs. For example, if a customer has purchased health insurance and travel insurance more frequently in the past, the system can preliminarily determine that they have a greater demand for these types of insurance.

[0042] By accessing the customer's historical purchase records on the insurance platform or third-party platform, the system can automatically collect and analyze this data. Technically, the customer's purchase records can be queried through the database, and compared and analyzed with the existing insurance product database to find out the customer's preferred insurance type and coverage.

[0043] Browsing behavior includes the customer's search history on the insurance platform, the type of insurance products browsed, the length of stay, etc. These behaviors can reflect the customer's current interests and demand tendencies. By analyzing browsing behavior, the system can identify which insurance products the customer has potential demand for.

[0044] Utilizing data tracking technology, we record customers’ browsing paths on the platform through tracking and log analysis, extract the types of products that customers are interested in and their behavioral characteristics (such as click frequency, browsing time, etc.), and analyze the demand intentions behind them through machine learning models.

[0045] Customer portraits are generated based on multi-dimensional data such as the customer's personal information, historical purchase records, browsing behavior, etc. The portrait contains information such as the customer's age, gender, income level, occupation, marital status, and past insurance product purchasing habits. Through customer portraits, we can better match and recommend insurance products suitable for customers.

[0046] After summarizing various customer information, a customer portrait is generated through data modeling and feature extraction technology. The image usually includes labels such as the customer's risk tolerance and spending capacity. This process can use a big data analysis platform to classify customer data through algorithms to generate a comprehensive multi-dimensional customer portrait.

[0047] After building a customer profile, it is necessary to further extract key features from the customer profile, such as age, income level, risk tolerance, etc. Different insurance products are suitable for customers with different characteristics. For example, customers with higher income and stronger risk tolerance may be more suitable for insurance products with higher insured amounts, while customers with lower income levels may prefer products with lower premiums.

[0048] Using feature extraction algorithms, the system extracts important feature information from customer profiles and matches these features with the attributes of insurance products. The extraction process can be implemented through rule definition or based on machine learning models, and the results of feature extraction will be used for subsequent insurance product recommendations.

[0049] Example: In the financial sector, customers have diverse and complex needs. Insurance companies can recommend property insurance products more accurately based on customer personal data and behavior analysis. For example, if a customer with a high income level has browsed travel insurance and health insurance many times and has a purchase record, the system can analyze his income and risk preferences to recommend a combination of travel insurance and health insurance with high coverage to better meet his needs.

[0050] By comprehensively analyzing customers' personal information, historical purchase records, and browsing behaviors, we can accurately build customer portraits and extract key features, such as age, income level, and risk tolerance. These features provide strong data support for subsequent insurance product recommendations, making the recommended insurance products more in line with customer needs and improving customer satisfaction. Through personalized portrait analysis, insurance companies can better understand customers' insurance needs, thereby improving the accuracy of product matching, increasing policy sales, and reducing customer dissatisfaction and churn.

[0051] S20, extracting insurance features of each insurance product from the non-auto property insurance product database, wherein the insurance features include suitable population categories, insurance identification values, and coverage;

[0052] In this embodiment, the non-auto property insurance product library is a database that stores all non-auto property insurance products, covering product information such as property insurance, health insurance, travel insurance, liability insurance, accident insurance, etc. The system needs to extract various insurance products from this product library for subsequent feature analysis and recommendation.

[0053] The product database can obtain all relevant insurance product data from the insurance company's internal system or external database through standard database query statements (such as SQL queries). The product database is usually classified and stored by product category, terms, premium, insurance amount and other characteristics. The system queries the data in the product database through the API interface or direct database connection.

[0054] Each insurance product has a suitable population category, which is usually defined based on the customer's age, occupation, income level, lifestyle, etc. For example, some health insurance is suitable for young people, while some liability insurance is more suitable for business owners or high-income people. The extraction of suitable population categories is to ensure that the recommended insurance products match the customer profile.

[0055] The suitable population category can be determined by the terms of the insurance product or the specific product definition. The system reads the description of the suitable population in the product terms and maps it to the relevant characteristics in the customer profile. For example, if the insurance product description clearly states that it is suitable for "high-income people aged 30-45", the system will match it with the customer's age and income level.

[0056] The insurance identification value is used to distinguish and mark specific information of different insurance products, including basic identification values ​​and additional identification values. The basic identification value includes the core characteristics of the insurance product, such as the type of insurance, premium range, etc.; the additional identification value can cover specific insurance terms, coverage or special conditions. The extraction of insurance identification values ​​helps to accurately match the specific information of insurance products with customer needs.

[0057] The insurance identification value can be directly extracted from the insurance product terms and database. The system extracts the identification value and generates a set of data tags that can be used to identify product characteristics based on the product category and specific terms. These identification values ​​can be used in the subsequent matching and recommendation process to help screen out products that are highly relevant to customer needs.

[0058] The coverage refers to the insurance liability content, liability exemption clauses and claim limits covered by each insurance product. The coverage of different insurance products can involve multiple aspects such as medical care, property, and liability. Therefore, when recommending insurance products, accurate matching of the coverage is crucial. The purpose of extracting the coverage is to ensure that the recommended insurance products can provide comprehensive protection for customers.

[0059] The coverage information is stored in the terms field of the product library. The system can parse the content of the insurance terms to extract relevant insurance liability, exemption clauses, claim limits and other data. Through natural language processing (NLP) technology or rule matching technology, the system can automatically identify the specific content of the coverage and store it in a structured manner to facilitate subsequent matching and recommendations.

[0060] Example description: In the insurance field, suppose a customer's profile shows that he is 35 years old, has a high income and a strong risk tolerance. By extracting product features from the non-auto property insurance product library, the system identifies a combination of health insurance and liability insurance that is suitable for this customer. These products have higher premiums and a wider coverage, which can meet the customer's high protection needs. At the same time, the matching of the insurance identification value indicates that the combination product is suitable for high-income people and provides additional supplementary protection terms, such as high medical compensation and liability exemption. This intelligent matching process improves the customer experience, enabling customers to quickly obtain insurance products that meet their needs, and insurance companies have achieved higher policy sales as a result.

[0061] By accurately extracting the key features of insurance products from the non-auto property insurance product library, such as the appropriate population category, insurance identification value and coverage, it helps to achieve personalized insurance product recommendations. By extracting and matching these features, the system can provide customers with a more suitable insurance product portfolio based on customer portraits, reduce product duplication and redundancy, and improve the efficiency and accuracy of insurance recommendations, thereby enhancing customer satisfaction and increasing the sales conversion rate of insurance policies.

[0062] S30, initializing particles in a particle swarm optimization method, and randomly generating a plurality of particles, each particle representing a non-auto property insurance combination plan;

[0063] In this embodiment, the particle swarm optimization method (PSO) is an optimization algorithm that simulates group behavior in nature and is used to solve complex optimization problems. The particle swarm optimization method is used to optimize and select the optimal non-auto property insurance combination plan. The core of the particle swarm optimization method is to search in the solution space through multiple particles and gradually approach the optimal solution. Each particle represents a possible solution, which is an insurance combination plan in this embodiment.

[0064] The particle swarm is initialized by defining its initial parameters, such as the number of particles, inertia weight, learning factor, etc. Each particle corresponds to a specific insurance portfolio. The implementation of the algorithm usually uses an optimization library or a custom PSO algorithm module in a programming language (such as Python, Matlab, etc.) to handle the insurance product portfolio optimization problem in the solution space.

[0065] Particle initialization means that in the particle swarm optimization method, the system generates multiple initial particles, each of which represents a possible insurance combination. The "position" of each particle represents a specific configuration of the insurance combination, while the "speed" indicates how the particle moves in the solution space to find a better solution. Initializing particles is the first step of the particle swarm optimization algorithm, providing a starting point for subsequent iterations and optimizations.

[0066] During the system initialization process, the position and velocity of the particles are determined by the random distribution function. The system will randomly select different non-auto property insurance products as the initial value of the particle position and assign an initial velocity to each particle. The position can be represented by a data vector, with each dimension corresponding to the selection of an insurance product. The velocity represents the direction and step size of the particle movement. Generally, the number and velocity of the initialized particles depend on the complexity of the problem and the size of the search space.

[0067] The random generation of multiple particles is to ensure the diversity of the initial solution and prevent the algorithm from falling into the local optimal solution. Each particle represents a possible non-auto property insurance combination plan. The random generation of particles means that these combination plans are initially scattered in the solution space, thereby increasing the chance of finding the optimal solution. Randomness plays a role in expanding the search scope in the algorithm and helps to improve the quality of the final recommended solution.

[0068] The random generation of the particle swarm algorithm can be achieved through a pseudo-random number generator or a random distribution function. The system will randomly combine different non-auto property insurance products according to the characteristics and configuration rules of the insurance products to generate multiple particles. These particles represent different insurance combination schemes, and the initial number of particles can be set by empirical values ​​or according to the scale of the solution space. The randomly generated particles are distributed in different areas of the solution space to ensure the breadth of exploration.

[0069] In the particle swarm optimization method, each particle corresponds to a specific non-auto property insurance combination plan. Different particles are combined in different ways, and they will try to meet customer needs with different configurations. Each particle contains multiple non-auto property insurance product combinations, and is scored and optimized based on their match with the customer profile. The role of particles is to gradually approach the optimal insurance combination plan through repeated iterations.

[0070] The position vector of each particle represents a non-auto property insurance product portfolio. Specifically, each dimension of the vector represents an insurance product category, and the value of the dimension represents the selected insurance product. The system will iteratively adjust the "position" of these particles, that is, continuously adjust the insurance product portfolio to find the best solution that best meets customer needs. This process can combine the characteristics of the customer portrait and the product characteristics, and finally select the solution with the highest matching degree by repeatedly updating the particle position.

[0071] Example: In the financial field, suppose a customer needs to purchase non-auto property insurance. The system randomly generates multiple particles through the particle swarm optimization method, and each particle corresponds to a different insurance combination plan. For example, a particle may represent a combination of health insurance, liability insurance, and accident insurance, and another particle may represent a combination of health insurance and travel insurance. Through the particle swarm optimization algorithm, the system will gradually optimize these combination plans and finally recommend the plan that best meets the customer's needs.

[0072] By initializing particles in the particle swarm optimization method and randomly generating multiple particles, each particle represents a non-auto property insurance combination plan, the diversity of the initial plan can be guaranteed, and the optimal insurance combination plan can be gradually found through iterative optimization. This avoids the local optimal solution problem that traditional recommendation systems may fall into, and improves the recommendation accuracy and efficiency.

[0073] S40, comparing the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyzing the matching degree of each non-auto property insurance combination plan, and determining the initial individual optimal position of each particle based on the matching degree analysis result;

[0074] In this embodiment, the non-auto property insurance combination represented by each particle usually includes multiple insurance products, such as health insurance, liability insurance, accident insurance, etc. Each insurance product has its own independent insurance characteristics, including suitable population categories, insurance identification values, coverage, etc. During the matching process, the system will analyze the characteristics of each insurance product one by one and match them with the characteristics in the customer profile.

[0075] First, the system extracts the corresponding insurance features (such as suitable age, income level, coverage, etc.) from each insurance product, and then compares these features with the specific features in the customer profile (such as age, income, risk tolerance, etc.). The matching degree of each insurance product can be calculated through rule matching or machine learning-based models to ensure that the characteristics of each insurance product are fully analyzed.

[0076] The overall matching degree of the combination is the weighted sum of the matching degrees of all insurance products. The matching degree of each insurance product contributes differently to the overall combination. The system needs to calculate the matching degree of the overall combination based on the matching degree and importance of different insurance products. For example, the matching weight of some core insurance products may be higher, while the matching weight of other secondary products may be relatively lower.

[0077] The system will calculate the matching degree of each insurance product separately, and then summarize it according to the preset weight to get the overall matching degree of the whole combination plan. For example, the matching degree weight of health insurance may be higher, and the weight of travel insurance may be lower. The final matching degree score will be used as the overall score of the combination plan to determine whether it can become the individual best position of the customer.

[0078] The initial individual optimal position of each particle is determined based on the matching analysis results of all insurance products. If the combination scheme in a particle has a high overall matching degree, the system will set its current position as the initial individual optimal position of the particle. This process ensures that the particle swarm optimization can start from a good initial position and continuously optimize the scheme through iteration.

[0079] The system first calculates the matching score of each insurance product, and then calculates the matching of the entire combination plan based on these scores. Based on the matching results, the system will determine whether to update the individual optimal position of the particle, and iteratively adjust according to this optimal position during the particle swarm optimization process.

[0080] Example description: In the financial field, suppose the customer profile shows that the customer is 45 years old, has a high income level, and has a medium risk tolerance. The system generates multiple non-auto property insurance combination plans for the customer, which include health insurance, liability insurance, and travel insurance. Each insurance product has different insurance characteristics, such as health insurance is suitable for people aged 40-50, liability insurance is suitable for high-income families, and travel insurance has a wider coverage. The system will match and analyze the characteristics of these three insurance products with the customer profile separately, and calculate the matching degree of each product. For example, health insurance is highly matched with the customer's age, and liability insurance is matched with his income level. Finally, the matching score of the entire combination plan is obtained through weighted calculation. The system updates the initial individual optimal position of the particle based on the matching result, and further improves the accuracy of the recommended plan in the subsequent optimization process.

[0081] By matching the features of multiple insurance products in each particle one by one, the system can more accurately analyze the overall matching degree of the combination plan. Determining the initial individual optimal position of the particle based on the matching analysis results can ensure that the optimization process starts from a reasonable initial position and avoid falling into the local optimal solution during the search process. Through the optimized recommendation plan, customers can obtain a more personalized and comprehensive insurance product portfolio, which improves the accuracy of insurance recommendations and customer satisfaction.

[0082] S50, in each iteration of the particle swarm optimization method, updating the speed and position of each particle in the particle swarm, and re-performing matching degree analysis on each updated particle position;

[0083] In this embodiment, the core of the particle swarm optimization method (PSO) lies in the iterative process, which searches for the optimal solution by continuously updating the speed and position of the particles. In each iteration, the system is updated according to the current state of each particle (i.e., the current speed and position) and moves towards the optimal solution. The iterative process allows the particle swarm to gradually approach the optimal solution.

[0084] At the beginning of each iteration, the system calculates the speed and position of each particle through a formula and gradually updates the state of the particle. This process stops when a preset number of iterations or when a certain convergence condition is reached. In practice, the system uses an optimization algorithm library in a programming language or a custom PSO algorithm to manage the iteration process.

[0085] In the particle swarm optimization method, the speed and position of the particles are constantly updated for global search. The speed represents the direction and speed of the particle's movement in the solution space, and the position represents the current insurance product portfolio. In each iteration, the speed and position are adjusted based on the inertia weight, the individual optimal position, and the global optimal position to ensure that the particles continue to move towards a better solution.

[0086] The particle's velocity and position are updated according to the following formula:

[0087] Speed ​​update formula:

[0088] in, represents the velocity of particle i in the dth dimension at the tth iteration; w is the inertia weight, which controls the influence of the original velocity of the particle; c1 and c2 are learning factors, which represent the influence coefficients of moving closer to the individual optimal position and the global optimal position of the particle respectively; r1 and r2 are random numbers in the range of [0,1] used to introduce randomness and ensure the diversity of particles; is the current individual optimal position of particle i; is the global optimal position of the entire particle swarm; represents the position of particle i at the tth iteration.

[0089] Position update formula:

[0090] It means that the particle's position update is based on the sum of the previous position and the new velocity, and the new position represents the particle's current new insurance combination plan.

[0091] Each time the particle position is updated, the system needs to re-analyze the particle's current insurance combination plan. The matching analysis is to determine the degree of fit between the plan and the customer's needs by comparing the insurance characteristics of the current combination plan with the characteristics of the customer profile. Through the re-matching analysis, the system can determine whether the updated particle position is closer to the optimal plan.

[0092] For each updated particle position, the system will re-extract the features of the insurance product portfolio corresponding to that position, and then calculate the matching degree between it and the features in the customer profile. The matching degree calculation can be based on rule matching or machine learning models, and the matching degree between the new position and customer needs is analyzed through each iteration. After re-analyzing the matching degree, the system will record the current matching degree result for optimization in the next iteration.

[0093] During each iteration of the particle swarm optimization method, the speed and position of the particles are constantly updated. After each update, the system will re-analyze the matching degree. Through multiple iterations, the particles gradually move closer to the position with higher matching degree, and finally find the global optimal solution. The update process of each iteration helps to search for the optimal insurance combination solution in a wide solution space and avoid falling into the local optimal solution.

[0094] In each iteration, the system updates the speed and position of all particles and re-analyzes the matching degree of each particle one by one. The system records the matching degree between the individual optimal position and the global optimal position in each iteration, determines whether to further adjust the speed and direction of the particles, and finally finds the optimal insurance combination solution.

[0095] Example description: In the financial field, the system recommends non-auto property insurance combination plans for customers through particle swarm optimization method. Assuming that a customer's portrait shows that he is 50 years old and has a medium income, the system initially generates multiple combination plans including health insurance, accident insurance and liability insurance. As the iteration proceeds, the system continuously updates the speed and position of the particles according to the current global optimal position and individual optimal position. After each update, the system re-analyzes the matching degree of the new combination plan with customer needs, for example, whether the coverage of health insurance is suitable for the customer's age group, and whether the premium of liability insurance is in line with his income level. After many iterations, the system finally recommended an insurance combination plan that highly meets the customer's needs, achieving higher recommendation accuracy and conversion rate.

[0096] Through iterative updates of the particle swarm optimization method, the system can continuously adjust the speed and position of particles and dynamically explore better non-auto property insurance combination solutions. After each iteration, the matching degree analysis is re-performed to ensure that each updated particle position can reflect the current optimal solution. It can not only improve the accuracy of recommendations, but also avoid falling into local optimal solutions, so that the final recommended insurance combination is more in line with the actual needs of customers, improve customer satisfaction and promote policy conversion.

[0097] S60, updating the individual optimal position of each particle and the global optimal position of the particle group according to the re-determined matching degree analysis result;

[0098] In this embodiment, after each iteration to update the particle speed and position, the system needs to re-analyze the matching degree to evaluate whether the new insurance product combination plan is more in line with the customer's needs than the previous plan. The matching degree analysis is to calculate the degree of fit between the current particle position (insurance combination plan) and the customer profile. The higher the matching degree, the more suitable the insurance combination is for the customer.

[0099] Matching analysis can be achieved through rule matching or machine learning models. The system first extracts the current insurance product features of the particle, compares them with the features of the customer portrait, and obtains a new matching score. The calculation of the matching score may be based on multi-dimensional features, such as suitable population category, coverage, insurance identification value, etc.

[0100] The individual optimal position refers to the insurance combination that each particle finds during its movement that best matches the customer's needs. Based on the matching analysis results, the system will determine whether the current particle's matching degree exceeds its best matching degree in history. If it exceeds the previous best matching degree, the system will update the current position to the particle's individual optimal position.

[0101] Implementation: After each iteration, the system compares the current match with the particle's historical best match. If the current match is higher, the system will update the particle's individual best position. This process is completed through the following logic: If the match is higher, the system will update the particle's individual best position. Then update the individual optimal position of the particle

[0102] The global optimal position refers to the best insurance combination solution found by the entire particle swarm in all iterations. After each iteration, the system checks the individual optimal positions of all particles. If the matching degree of the individual optimal position of a particle is higher than the previous global optimal matching degree, the global optimal position is updated to the individual optimal position of the particle.

[0103] The system traverses the individual optimal positions of all particles Find the position with the highest matching degree. If the matching degree of the individual optimal position of a particle is higher than the global optimal matching degree, then update the global optimal position, indicating The update rule is: if the matching degree but

[0104] By updating the individual optimal positions and global optimal positions of particles, the system can ensure that the optimization process always moves towards the optimal solution and finds the insurance product portfolio that best meets customer needs.

[0105] S70, after the iteration is completed, the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm is selected as the final output.

[0106] In this embodiment, the particle swarm optimization method updates the speed and position of the particles in each iteration through multiple iterations, gradually approaching the optimal solution. The condition for the end of the iteration is usually to set a maximum number of iterations or to reach a certain convergence condition (for example, the global optimal position does not change in several consecutive iterations). When the iteration process is completed, it means that the system has been as close to the optimal solution as possible in the current search space.

[0107] The system updates the particle speed and position and evaluates the matching degree during each iteration. When the preset number of iterations is reached or the matching degree change is less than the preset threshold, the system determines that the iteration is over. At this time, the global optimal position in the particle swarm is the optimal solution found.

[0108] The global optimal position is the best solution among all particles, representing the insurance combination solution that has been found so far and has the highest degree of match with customer needs. Through multiple iterations in the solution space, the particle swarm optimization algorithm ensures that the system finds the insurance solution that best matches the characteristics of the customer profile. Selecting the global optimal position means selecting the insurance combination solution with the highest degree of match among all current particles.

[0109] At the end of the iteration, the system will compare the matching degree of all particles and finally select the global optimal position with the highest matching degree. The system will record the insurance product combination represented by this position and output it as the final recommendation result.

[0110] The non-auto property insurance combination plan is a combination generated by matching customer profile features with multiple insurance product features, including different types of non-auto property insurance, such as health insurance, liability insurance, travel insurance, etc. The system will provide the insurance product combination corresponding to the global optimal position as the final recommendation plan to the customer.

[0111] When the global optimal position of the particle swarm is determined, the system extracts the insurance combination plan corresponding to the position, including detailed information of each insurance product, such as coverage, premium, suitable population, etc. The system generates a report or recommendation content for the combination plan as the final output result and recommends it to the customer.

[0112] Example description: In the financial field, suppose the customer's profile shows that he is 40 years old, has a high income level, and prefers to choose insurance products with high protection. The system generates multiple insurance combination plans through particle swarm optimization method, including different combinations of health insurance, liability insurance and travel insurance. After multiple iterations, the system finds the global optimal position, and the insurance plan corresponding to this position is a combination that highly meets customer needs: including high-protection health insurance and additional travel insurance. The system uses this combination plan as the final output and generates a recommendation report to the customer to ensure that the customer obtains the most suitable insurance combination plan.

[0113] Through multiple iterations of the particle swarm optimization method, the system finally selects the non-auto property insurance combination plan corresponding to the global optimal position as the final output, ensuring that the recommended plan is the insurance combination that best matches the customer's profile characteristics. It can significantly improve the accuracy of insurance product recommendations, ensure that customers are provided with personalized and diversified insurance options, improve customer satisfaction, and thus improve the sales conversion rate of insurance companies. In addition, through the intelligent optimization process, the system can effectively avoid falling into the local optimal solution and ensure that the global optimal solution is found, thereby maximizing the effect of optimizing insurance recommendations.

[0114] The present invention relates to the fields of artificial intelligence technology and financial technology, and discloses a non-auto property product combination output method, which constructs a customer portrait and extracts customer portrait features by analyzing the customer's personal information, historical purchase records and browsing behavior; extracts the insurance features of each insurance product from the non-auto property insurance product library, including suitable population categories, insurance identification values ​​and coverage; initializes particles in a particle swarm optimization method, randomly generates multiple particles, and each particle represents a non-auto property insurance combination scheme; performs a matching degree analysis on the insurance features of each particle and the customer portrait features, and determines the initial individual optimal position of each particle; updates the speed and position of the particles through the iteration of the particle swarm optimization method, and re-performs the matching degree analysis to update the individual optimal position and the global optimal position of each particle; and finally selects the non-auto property insurance combination scheme corresponding to the global optimal position as the output. The present invention accurately analyzes the multi-dimensional features in the customer portrait, matches the actual needs of the customer, effectively combines a variety of non-auto property insurance together, ensures the personalization and natural connection of the combination scheme, avoids product duplication and unnecessary selection complexity in the prior art, and thus significantly improves customer satisfaction.

[0115] In one embodiment, the above S10 includes:

[0116] S101, collecting basic personal information of the customer, including age, gender and income;

[0117] S102, retrieving the customer's historical purchase records from the insurance database, wherein the historical purchase records include the type and quantity of insurance products purchased;

[0118] S103, tracking the browsing behavior of the customer on the insurance platform through the interaction data of the insurance platform, wherein the browsing behavior includes the type of insurance products browsed, the browsing frequency and the browsing duration;

[0119] S104, integrating the basic personal information, historical purchase records and browsing behaviors to generate a comprehensive customer profile;

[0120] S105, extracting customer profile features from the comprehensive customer profile, wherein the customer profile features include age, income level and risk tolerance.

[0121] In this embodiment, basic personal information refers to the basic information provided by the customer in the insurance platform or related systems, including the customer's age, gender and income level. This information is used to initially build a customer profile and provide a data basis for subsequent insurance product recommendations. For example, age can reflect the customer's health status and life expectancy, and income level is closely related to the customer's ability to bear premiums.

[0122] The system collects basic personal information through customer registration or filling out information forms. The front-end page will present a data collection form to the user, and the information entered by the user will be passed to the back-end database for storage through the API interface. If necessary, the customer's age can be verified through ID card information, or their income level can be confirmed through income proof.

[0123] Historical purchase records refer to information about insurance products that customers have purchased from insurance companies or related platforms, including the types and quantities of products purchased. These records can reflect customers' insurance consumption habits and preferences, and help the system understand customers' demand for different insurance products. For example, if a customer has purchased health insurance many times, it means that he or she has a high demand for health protection.

[0124] The system searches the insurance company's customer purchase database to retrieve the customer's historical purchase records. It extracts the customer's purchase product type, purchase time, premium, and insured amount from the database through SQL query statements or API interfaces, and combines it with the customer profile. This data can help the system infer the customer's preference for a certain type of product.

[0125] Browsing behavior refers to the customer's interactive activities on the insurance platform, such as the type of insurance products browsed, the frequency of browsing, and the duration of browsing. These behavioral data can show the customer's potential interest in a certain type of insurance product and help the system predict the customer's future insurance needs. For example, if a customer browses the travel insurance page many times, it may mean that he is about to travel and has a need to purchase travel insurance.

[0126] Through point-of-sale technology, log recording or the use of third-party data tracking tools, the system can collect customer behavior data on the insurance platform. Every user's click on the page, length of stay, access frequency, etc. will be recorded in the log file or analysis system. Through data analysis, the system can classify and model the customer's browsing behavior for use in subsequent recommendations.

[0127] Customer portrait is a model generated based on multi-dimensional data, which integrates the customer's basic information, purchase history and behavior data to form an overall customer portrait. This portrait can help the system understand the customer's insurance needs more deeply, so as to recommend the most suitable insurance products for them. For example, the system can comprehensively analyze the customer's risk tolerance and insurance preferences based on his income, purchase history and browsing behavior.

[0128] The system integrates the collected customer data and generates a comprehensive customer profile through big data analysis and feature extraction technology. Data cleaning and feature extraction are important steps in generating customer profiles. The system combines structured data (such as age, gender, income) with unstructured data (such as browsing behavior) to generate a customer profile containing multiple dimensions of information.

[0129] Customer profile features are key attributes extracted from comprehensive customer profiles, including age, income level, and risk tolerance. By extracting these key features, the system can match customer profile features with insurance product features, thereby providing personalized insurance recommendations for customers. Age can affect the choice of insurance products, income level determines the ability to bear premiums, and risk tolerance affects the demand for insurance coverage.

[0130] The system automatically extracts key features from customer profiles through feature extraction algorithms. Each feature can be calculated through rule setting or models based on historical data. The system scores each customer's profile and generates a feature vector that reflects their insurance needs for use by subsequent matching algorithms.

[0131] Example description: In the financial field, suppose a customer registers on an insurance platform and provides information that he is 35 years old and has a medium income. At the same time, the system extracts records of the customer's purchase of health insurance and liability insurance from historical data. Through browsing behavior analysis, the system finds that the customer has recently frequently viewed the travel insurance page. The system integrates this information to generate a comprehensive profile of the customer, judging that his risk tolerance is medium and he has a large demand for health insurance and travel insurance. Based on this profile, the system recommends a combination of a high-security health insurance and a travel insurance plan, which improves the accuracy of the recommendation and increases customer satisfaction.

[0132] In this embodiment, by comprehensively analyzing the customer's personal information, historical purchase records and browsing behavior, the system can build a complete and detailed customer profile and extract key features. Through these profile features, the system can more accurately recommend suitable insurance products to customers, thereby improving the accuracy of recommendations and customer satisfaction.

[0133] In one embodiment, the above S20 includes:

[0134] S201, obtaining an insurance product list from the non-auto property insurance product database, wherein the insurance product list includes information of all insurance products belonging to the non-auto property insurance category;

[0135] S202, classify each insurance product according to its nature and terms, and determine the corresponding suitable population category;

[0136] S203, extracting and recording a corresponding insurance identification value from the terms of each insurance product, wherein the insurance identification value includes a basic identification value and an additional identification value;

[0137] S204, determining the coverage of each insurance product based on the terms of each insurance product, wherein the coverage includes insurance liability content, liability exemption clauses, and claim identification value limits;

[0138] S205, feature labeling is performed on the suitable population category, insurance identification value and coverage scope of each insurance product to generate insurance features of each insurance product.

[0139] In this embodiment, the non-auto property insurance product library is a database that stores and manages product information of all non-auto property insurance categories. The system needs to obtain all insurance products that meet the non-auto property insurance category from the product library, including property insurance, health insurance, liability insurance, etc., and generate a product list. This product information includes but is not limited to insurance name, product number, terms, coverage content, etc.

[0140] The system extracts all non-auto property insurance product data from the product library through API interfaces or database queries to form a product list. SQL query statements or RESTful API-based interface calls can be used to extract this data to ensure that the list contains all non-auto insurance products.

[0141] The nature and terms of each insurance product define the target population it is suitable for. For example, some health insurance may be suitable for young people, while liability insurance may be suitable for people working at high risk. The system needs to classify and determine the suitable population category for each product based on the nature of the insurance product (such as product category, coverage) and terms and conditions.

[0142] The system reads the terms and conditions of each insurance product, parses the description in the product terms, and uses rule matching or natural language processing (NLP) technology to classify the products and determine the suitable population categories. For example, the system can automatically identify suitable age groups or occupational groups based on keywords (such as "suitable for people aged 30-45") and associate this information with product features.

[0143] The insurance identification value is used to identify the characteristics of different insurance products, including basic identification values ​​and additional identification values. The basic identification value may represent the core information of the product, such as the insured age, premium range, etc., while the additional identification value can reflect the specific terms or additional protection of the product. The extraction of insurance identification values ​​helps to structure the characteristics of insurance products, which is convenient for subsequent matching and recommendation.

[0144] The system automatically extracts key fields and records their insurance identification values ​​by parsing the terms of each insurance product. Basic identification values ​​such as insurance period and insured amount can be extracted through rule matching, and additional identification values ​​may be obtained through more detailed text analysis. The extracted identification values ​​will be stored in the database for subsequent recommendation.

[0145] The coverage of each insurance product refers to the insurance liability content, liability exemption clauses and claim limits it covers. Different insurance products may provide different levels of protection and claim conditions, which directly affect customers' choice of insurance products. Therefore, the system needs to extract and determine the coverage of each insurance product from the product terms.

[0146] The system parses the coverage information in the insurance terms through natural language processing or preset rules. The terms may clearly list the coverage responsibilities (such as "covering medical expenses"), liability exemptions (such as "except high-risk activities") and claim limits (such as "the upper limit of compensation is 500,000 yuan"). The system automatically extracts this information through text analysis technology, generates structured coverage fields, and stores them in the feature database of insurance products.

[0147] Feature labeling refers to labeling the suitable population category, insurance identification value and coverage of each insurance product in a specific format to generate structured insurance product features. These labels will be used for subsequent insurance product matching and recommendation to ensure that the system can quickly and accurately compare product features with customer profiles.

[0148] The system formats all extracted product features (suitable population, insurance identification value, coverage) and generates labels. The labels corresponding to each insurance product can be stored in the form of key-value pairs to mark the key characteristics of the insurance. These feature labels will be stored in the insurance product database for the recommendation algorithm to call.

[0149] Example description: In the financial field, the system needs to recommend a non-auto property insurance package for a certain customer. The customer profile shows that the customer is 45 years old, has a high income, and is concerned about health protection. The system extracts multiple related health insurance products from the non-auto property insurance product library, and by parsing their terms, determines the suitable population category (40-50 years old middle- and high-income people), insurance identification value (high insurance amount, high premium) and coverage (comprehensive health protection, claim limit of 1 million yuan). The system labels these product features and matches them with the customer profile features, and finally recommends a health insurance and liability insurance combination plan that suits the customer's needs.

[0150] This embodiment extracts the suitable population category, insurance identification value and coverage of each insurance product from the non-auto property insurance product library and performs feature labeling processing on them, so that the system can recommend the most suitable insurance product combination for each customer. Feature labeling processing ensures the standardization and structuring of data, facilitates the system to quickly compare the features of different insurance products, and improves the accuracy of insurance product recommendations.

[0151] In one embodiment, the above S30 includes:

[0152] S301, determining the initial number of particles in the particle swarm to set the particle swarm scale;

[0153] S302, randomly distributing the positions of particles in the defined solution space, wherein the position of each particle corresponds to a non-auto property insurance combination plan;

[0154] S303, assigning an initial velocity to each particle, wherein the initial velocity represents the moving direction and velocity of the particle in the solution space;

[0155] S304, initializing the individual optimal position of each particle as the initial position of each particle.

[0156] In this embodiment, the number of particles in the particle swarm optimization method determines the coverage of the solution space, and the setting of the initial number of particles affects the effect of the entire optimization process. If the size of the particle swarm is too small, it may lead to insufficient coverage of the solution space and difficulty in finding the optimal solution; if the size is too large, it may increase the computational cost. Therefore, the system needs to determine the initial number of particle swarms based on the complexity of the problem and the size of the solution space.

[0157] The system can set the initial number of particles by empirical rules or according to the complexity of the solution space. Usually, the number of particles is between tens and hundreds. The system sets the number of particle swarms through an initial parameter configuration to ensure that a sufficient range of the solution space is covered. Depending on the algorithm implementation used (such as Python, Matlab, etc.), the configuration parameters in the programming language can be used to determine the initial number of particles.

[0158] The position of a particle represents a possible solution of the particle in the solution space. The position of each particle corresponds to a combination of non-auto property insurance. During initialization, the system needs to randomly distribute the positions of these particles to ensure that the particle swarm has a good coverage of the solution space, thereby increasing the possibility of finding the global optimal solution.

[0159] The system randomly assigns the initial position of each particle through a random function or random number generator to ensure uniform particle distribution. The position of each particle is a multidimensional vector, representing different feature combinations of insurance products in the solution space. For example, the position of a particle may be composed of products such as health insurance and liability insurance, and represents attributes such as premiums and coverage under different combinations.

[0160] The speed of a particle determines its moving direction and step size in the solution space. The setting of the initial speed affects the search path of the particle during the iteration process. The system needs to randomly assign an initial speed to each particle to ensure that the particle can explore the solution space in different directions.

[0161] By randomly generating the initial velocity of the particles, the system ensures that the particles move in different steps and directions. The initial velocity can be represented by a randomly generated vector whose dimension is consistent with the dimension of the solution space. For example, if the solution space is three-dimensional (corresponding to the combined dimension of the three insurance products), the velocity of each particle is a three-dimensional vector. The system can generate the initial velocity vector through a random function to represent the movement trend of the particle in different directions.

[0162] The individual optimal position refers to the best position that a particle has found so far in the solution space. During initialization, the system sets the individual optimal position of a particle as its initial position, because the particle has not searched the solution space before iteration, so the current initial position is the known optimal solution.

[0163] When the system initializes the particles, it stores the current position of each particle as its individual optimal position. In subsequent iterations, the system will update the optimal position, but at the time of initialization, the particle's own initial position is the optimal solution. The system records the initial position of each particle in the form of a vector and compares and updates it in subsequent iterations.

[0164] Example description: In the financial field, suppose the system needs to generate a non-auto property insurance combination plan for a certain customer. The system first initializes the particle swarm through the particle swarm optimization method and randomly generates 50 particles. Each particle represents a different insurance product combination plan, including different combinations of health insurance, liability insurance and travel insurance. The system randomly assigns a position (insurance product combination) and speed (search step) to each particle, and sets the individual optimal position of each particle as its initial position. Subsequently, the system begins to iterate these particles multiple times, and by continuously updating the position and speed of the particles, it finally finds the insurance combination plan that best meets the customer's needs.

[0165] In this embodiment, by initializing particles in the particle swarm optimization method and randomly generating multiple particles, the system can generate multiple non-auto property insurance combination solutions in the solution space, ensuring the wide coverage of the solution space and the effectiveness of the search. By assigning an initial velocity to each particle, the system can ensure that the particle swarm explores the solution space in different directions in subsequent iterations, increasing the probability of finding the global optimal solution. The individual optimal position of the particle is initialized to its current position, ensuring that each particle has an initial solution as a benchmark, helping the system to continuously optimize the insurance product portfolio during the iteration process.

[0166] In one embodiment, in the above S40, comparing the insurance characteristics of the non-auto property insurance combination solution in each particle with the customer portrait characteristics, and analyzing the matching degree of each non-auto property insurance combination solution, includes:

[0167] S401, analyzing the matching scores of the suitable population category characteristics of the non-auto property insurance combination solution in each particle and the customer portrait characteristics according to the preset population matching standard;

[0168] S402, analyzing the matching score between the insurance identification value feature of the non-auto property insurance combination plan in each particle and the customer portrait feature according to the preset insurance matching standard;

[0169] S403, analyzing the matching score between the coverage characteristics of the non-auto property insurance combination solution in each particle and the customer profile characteristics according to a preset risk tolerance matching standard;

[0170] S404, summarizing the matching scores of the non-auto property insurance combination scheme of each particle in terms of suitable population category, insurance identification value and coverage scope, and determining the overall matching score of each particle.

[0171] In this embodiment, each non-auto property insurance product has its own suitable population category characteristics, such as age group, occupation, income level, etc. These characteristics are used to describe the most suitable insured population for the product. By comparing the suitable population category of the non-auto property insurance combination plan with the customer profile characteristics (such as age, gender, income, etc.), the system can analyze whether the combination plan meets the customer's needs and derive a matching score.

[0172] The system uses preset rules or machine learning models to compare the characteristics of the suitable population of each insurance product with the customer portrait. For example, if the suitable age range of an insurance product is "30-45 years old" and the customer is 35 years old, the matching score of the product in this dimension may be high. The matching score can be achieved through a scoring mechanism or a similarity calculation function, and the result is usually expressed as a percentage or score.

[0173] The insurance identification value is the key data used to distinguish the characteristics of insurance products, such as the insured amount, premium level, insurance period, etc. The system needs to compare the insurance identification value in the non-auto property insurance combination plan with the relevant characteristics in the customer profile (such as affordable premium, expected insurance period, etc.), analyze whether it meets the customer's needs, and give a matching score.

[0174] The system analyzes the corresponding data of the insurance identification value and the customer profile item by item through the predefined insurance matching criteria. For example, if the customer's income is 200,000, and the premium of a certain insurance product is suitable for people with an income of more than 200,000, then the insurance matching score of this product is high. The matching score can be calculated based on the preset threshold range or rule algorithm.

[0175] Each non-auto property insurance product has a certain coverage, such as the insurance liabilities, liability exemption clauses, and claim limits. The customer's risk tolerance determines which coverage is suitable for them. The system needs to analyze the match between the coverage in the non-auto property insurance package and the risk tolerance in the customer profile. The higher the match, the more the insurance plan meets the customer's risk preference.

[0176] Through the preset risk tolerance matching criteria, the system compares the coverage of the insurance product with the risk tolerance in the customer profile (such as the customer's assets, job nature, etc.). If the customer's risk tolerance is high and the insurance product provides extensive protection, the matching score is high. The matching calculation can be scored using a risk scoring model or matching rules.

[0177] After obtaining the matching scores for each dimension (suitable population category, insurance identification value and coverage), the system needs to summarize these scores and calculate the overall matching score for each particle. The overall matching score is used to measure the degree of fit between the entire insurance package and the customer profile, and ultimately decide whether to recommend the plan.

[0178] The system weights and summarizes the scores of suitable population categories, insurance identification values, and coverage to obtain an overall matching score. For example, the weighted average method can be used to calculate the overall matching, or different weights can be assigned to different dimensions based on specific rules. The system sorts all particles according to the summarized matching scores to determine the optimal solution.

[0179] Example description: In the financial field, suppose a customer's profile shows that he is 40 years old, has a medium income, and has a low risk tolerance. The system generates multiple non-auto property insurance combination plans, and each particle contains a different combination of health insurance and liability insurance. Through the matching analysis of the population category, the system finds that a health insurance product is particularly suitable for people in this age group and has a high matching degree; in the analysis of the insurance identification value, the premium and coverage of a liability insurance product are in line with the customer's income level and have a high score; and the analysis of the coverage shows that the customer prefers low-risk insurance products. The system weights and summarizes the matching scores of each dimension, and finally obtains the most suitable insurance combination plan for the customer and recommends it to the customer.

[0180] In this embodiment, by performing a detailed matching analysis between the non-auto property insurance combination plan in each particle and the customer profile characteristics, the system can recommend the insurance combination plan that best meets the customer's needs. Based on the matching calculation of different dimensions, the system can provide personalized insurance product recommendations, improve customer satisfaction and the accuracy of insurance plan recommendations.

[0181] In one embodiment, the above S50 includes:

[0182] S501, at the beginning of each iteration, recording the current speed of each particle, where the current speed is the speed of the particle at the end of the previous iteration;

[0183] S502, adjusting the current speed of each particle using the inertia weight to maintain the original movement trend of each particle;

[0184] S503, adjusting the speed direction and speed magnitude of each particle according to the individual optimal position of each particle and the global optimal position of the particle group;

[0185] S504, determining a new position of each particle according to the adjusted velocity direction and velocity magnitude;

[0186] S505, based on the new position of each particle, the non-auto property insurance combination plan represented by each particle is adjusted, and the matching degree analysis is performed again.

[0187] In this embodiment, the speed of a particle determines its moving direction and speed in the solution space. At the beginning of each iteration, the system records the current speed of each particle, which is the speed at the end of the previous iteration. This process ensures that the motion information of each particle is maintained and updated during each iteration. At the beginning of each iteration, the system stores the velocity vector of each particle as the current state. The velocity vector contains the motion information of each particle in each dimension (such as the moving direction and step size in the solution space). After storage, the system will continue to update the velocity of the particle.

[0188] InertiaWeight is used to control the movement tendency of particles so that they maintain a certain inertial motion in the solution space. By adjusting the inertia weight, the system can control whether particles are more inclined to explore new solutions or to make fine adjustments near existing solutions. The larger the inertia weight, the stronger the tendency of particles to retain the original direction of movement; the smaller the weight, the particles tend to move faster in a new direction.

[0189] The system adjusts the current speed of each particle by introducing the inertia weight parameter w into the speed update formula. The system sets the inertia weight based on empirical values, generally in the range of [0,1]. The size of the inertia weight is dynamically adjusted according to the algorithm convergence requirements.

[0190] The speed of a particle is not only determined by the inertia weight, but also affected by the individual optimal position and the global optimal position. The individual optimal position is the best position found by the particle itself in history, and the global optimal position is the optimal solution found among all particles. Through the influence of these two positions, the system can adjust the speed direction and size of the particle to ensure that the particle moves in the direction of the optimal solution.

[0191] The position of a particle represents its current solution in the solution space. After the velocity is updated, the particle will move to the new position at the adjusted velocity. The system calculates the new position of the particle in the solution space based on the latest velocity vector. This new position corresponds to a new non-auto property insurance combination plan.

[0192] The position of each particle corresponds to a non-auto property insurance combination plan. The change of the particle's position will cause the insurance combination plan it represents to change. Therefore, after each position update, the system needs to re-analyze the match between the combination plan and the customer profile to determine whether the plan is more in line with customer needs than before.

[0193] After the particle position is updated, the system re-analyzes the matching degree of the particle's current non-auto property insurance combination plan with the customer portrait, and the process is the same as the above matching degree calculation. The matching degree analysis includes comparisons of suitable population categories, insurance identification values, and coverage. The system determines whether the particle has a better solution based on the latest position and records the current matching degree.

[0194] In this embodiment, through iterative updates in the particle swarm optimization method, the system can dynamically adjust the speed and position of each particle to continuously approach the global optimal solution. After each position update, the system will re-analyze the matching degree of the non-auto property insurance combination plan represented by the particle to ensure that the insurance combination plan meets the characteristics of the customer profile.

[0195] In one embodiment, the above S60 includes:

[0196] S601, comparing the re-determined current matching degree analysis result of each particle with the historical optimal matching degree analysis result of each particle;

[0197] S602, if the current matching degree analysis result is better than the historical optimal matching degree analysis result, updating the individual optimal position of the particle to the current particle position;

[0198] S603, comparing the updated individual optimal position of the particle with the global optimal position of the particle swarm;

[0199] S604: If the individual optimal position of the particle is better than the global optimal position of the particle swarm, the global optimal position of the particle swarm is updated to the individual optimal position of the particle.

[0200] In this embodiment, each particle has a current matching degree in the solution space, which reflects the degree of fit between its current position (corresponding non-auto property insurance combination plan) and the customer profile. At the same time, the particle also saves its historical optimal matching degree, which indicates the best matching solution found by the particle in the past iteration process. The system needs to compare the current matching degree of the particle with its historical optimal matching degree to determine whether the current solution is better than the optimal solution found previously.

[0201] The system recalculates the particle's matching degree at each iteration and compares it with the best matching degree recorded in history. If the current matching degree is higher, it means that the insurance combination solution of the particle's current position is better than the best solution in history.

[0202] If the current matching degree of a particle exceeds its historical optimal matching degree, it means that the current position of the particle (i.e. the current insurance combination solution) is more suitable for customer needs than the previous optimal solution. In this case, the system will update the individual optimal position of the particle to the current position of the particle to save the currently found optimal solution.

[0203] The system determines whether to update the individual optimal position by comparing the current matching degree with the historical matching degree. The specific operation is: if the current matching degree > the historical matching degree, the individual optimal position of the particle is updated to the current position.

[0204] The global optimal position in a particle swarm represents the best solution found among all particles. The system needs to compare the individual optimal position of each particle with the global optimal position to determine whether any particle has found a solution that is better than the current global optimal solution. Through this comparison, the system can continuously update the global optimal solution to ensure that the entire particle swarm converges in the optimal direction.

[0205] After each update of the individual optimal position of a particle, the system will compare the position with the global optimal position. If the match of the individual optimal position of the particle is better than the current global optimal match, the global optimal position will be updated.

[0206] The global optimal position is the best solution among all particle swarms, representing the current optimal insurance combination in the particle swarm optimization process. If the individual optimal position of a particle is better than the existing global optimal position, the system will update the individual optimal position of the particle to the new global optimal position, thus ensuring that the system is always optimized towards the global optimal solution.

[0207] The system decides whether to update the global optimal position by comparing the matching degree of the individual optimal position and the global optimal position. The specific operation is: if the matching degree of the individual optimal position of a particle is greater than the global optimal matching degree, the global optimal position is updated to the individual optimal position of the particle.

[0208] Example description: In the financial field, the system recommends non-auto property insurance combination plans to customers through the particle swarm optimization method. In each iteration, the system calculates the current matching degree of each particle and compares it with the historical optimal matching degree of the particle. For example, if the matching degree of a particle in the t+1th iteration is better than its matching degree in the tth iteration, the system will update its individual optimal position to the current particle position. At the same time, the system compares the individual optimal position of the particle with the global optimal position of the particle swarm and finds that its matching degree is higher, so the global optimal position is updated to the individual optimal position of the particle.

[0209] In this embodiment, by comparing the matching analysis results, the system can dynamically update the individual optimal position of each particle and the global optimal position of the particle swarm, ensuring that the entire particle swarm optimization process converges towards the global optimal solution. The matching comparison and position update in each iteration can effectively improve the accuracy of the system's recommendations, making the final recommended non-auto property insurance combination plan more in line with customer needs.

[0210] In one embodiment, a non-automotive product combination output device is provided, and the non-automotive product combination output device corresponds to the non-automotive product combination output method in the above embodiment. Figure 3 , Figure 3 This is a functional module diagram of a preferred embodiment of the non-car property product combination output device of the present invention. Customer portrait construction module 10, insurance feature extraction module 20, particle initialization module 30, matching degree analysis module 40, particle swarm iteration update module 50, optimal position update module 60 and final combination output module 70. The functional modules are described in detail as follows:

[0211] The customer profile building module 10 is used to analyze the customer's personal information, historical purchase records and browsing behavior, build a customer profile, and extract customer profile features including age, income level and risk tolerance;

[0212] The insurance feature extraction module 20 is used to extract the insurance features of each insurance product from the non-auto property insurance product library, wherein the insurance features include the category of suitable people, the insurance identification value and the coverage;

[0213] A particle initialization module 30 is used to initialize particles in a particle swarm optimization method and randomly generate a plurality of particles, each particle representing a non-auto property insurance combination scheme;

[0214] A matching degree analysis module 40 is used to compare the insurance characteristics of the non-auto property insurance combination scheme in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination scheme, and determine the initial individual optimal position of each particle based on the matching degree analysis result;

[0215] The particle swarm iteration update module 50 is used to update the speed and position of each particle in the particle swarm during each iteration of the particle swarm optimization method, and re-analyze the matching degree of each updated particle position;

[0216] The optimal position updating module 60 is used to update the individual optimal position of each particle and the global optimal position of the particle group according to the re-determined matching degree analysis result;

[0217] The final combination output module 70 is used to select the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm as the final output after the iteration is completed.

[0218] In one embodiment, the customer portrait building module 10 is specifically used to:

[0219] Collect basic personal information of customers, including age, gender and income;

[0220] Retrieving the customer's historical purchase records from an insurance database, wherein the historical purchase records include the type and quantity of insurance products purchased;

[0221] Tracking the browsing behavior of customers on the insurance platform through the interaction data of the insurance platform, the browsing behavior including the type of insurance products browsed, browsing frequency and browsing duration;

[0222] Integrate the basic personal information, historical purchase records and browsing behaviors to generate a comprehensive customer profile;

[0223] Customer profile features are extracted from the comprehensive customer profile, wherein the customer profile features include age, income level, and risk tolerance.

[0224] In one embodiment, the insurance feature extraction module 20 is specifically used to:

[0225] Obtaining an insurance product list from the non-auto property insurance product database, wherein the insurance product list includes information of all insurance products belonging to the non-auto property insurance category;

[0226] Classify each insurance product according to its nature and terms and determine the corresponding suitable population category;

[0227] Extract and record the corresponding insurance identification value from the terms of each insurance product, wherein the insurance identification value includes a basic identification value and an additional identification value;

[0228] Determine the coverage of each insurance product based on the terms of each insurance product, including the insurance liability content, liability exemption clauses and claim identification value limits;

[0229] The suitable population category, insurance identification value and coverage scope of each insurance product are feature-labeled to generate insurance features for each insurance product.

[0230] In one embodiment, the particle initialization module 30 is specifically used to:

[0231] Determine the initial number of particles in the particle swarm to set the particle swarm size;

[0232] The positions of particles are randomly distributed in the defined solution space, and the position of each particle corresponds to a non-auto property insurance combination plan;

[0233] Assigning an initial velocity to each particle, wherein the initial velocity represents the moving direction and speed of the particle in the solution space;

[0234] Initialize the individual optimal position of each particle as the initial position of each particle.

[0235] In one embodiment, the matching degree analysis module 40 is specifically used for:

[0236] According to the preset population matching standard, the matching score between the suitable population category characteristics of the non-auto property insurance combination plan in each particle and the customer profile characteristics is analyzed;

[0237] According to the preset insurance matching standard, the matching score between the insurance identification value characteristics of the non-auto property insurance combination plan in each particle and the customer portrait characteristics is analyzed;

[0238] According to the preset risk tolerance matching standard, the matching score between the coverage characteristics of the non-auto property insurance combination plan in each particle and the customer profile characteristics is analyzed;

[0239] The matching scores of each particle's non-auto property insurance combination plan in terms of suitable population category, insurance identification value and coverage scope are summarized to determine the overall matching score of each particle.

[0240] In one embodiment, the particle swarm iterative update module 50 is specifically used to:

[0241] At the beginning of each iteration, the current speed of each particle is recorded, and the current speed is the speed of the particle at the end of the previous iteration;

[0242] Use inertia weight to adjust the current speed of each particle to maintain the original movement trend of each particle;

[0243] According to the individual optimal position of each particle and the global optimal position of the particle group, the speed direction and speed magnitude of each particle are adjusted;

[0244] Determine the new position of each particle according to the adjusted velocity direction and velocity magnitude;

[0245] Based on the new position of each particle, the non-auto property insurance combination plan represented by each particle is adjusted, and the matching degree analysis is performed again.

[0246] In one embodiment, the optimal position updating module 60 is specifically configured to:

[0247] Compare the re-determined current matching degree analysis result of each particle with the historical optimal matching degree analysis result of each particle;

[0248] If the current matching degree analysis result is better than the historical optimal matching degree analysis result, the individual optimal position of the particle is updated to the current particle position;

[0249] Comparing the updated individual optimal position of the particle with the global optimal position of the particle swarm;

[0250] If the individual optimal position of a particle is better than the global optimal position of the particle swarm, the global optimal position of the particle swarm is updated to the individual optimal position of the particle.

[0251] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via 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 and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external user terminal through a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the service side of a non-vehicle property product portfolio output method.

[0252] In one embodiment, a computer device is provided. The computer device may be a user terminal, and its internal structure diagram may be as follows: Figure 5 As 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 server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the user side of a non-vehicle property product portfolio output method

[0253] 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:

[0254] Analyze customers’ personal information, historical purchase records, and browsing behaviors, build customer profiles, and extract customer profile features including age, income level, and risk tolerance;

[0255] Extracting insurance features of each insurance product from the non-auto property insurance product database, wherein the insurance features include suitable population categories, insurance identification values, and coverage;

[0256] Initialize particles in the particle swarm optimization method and randomly generate multiple particles, each particle represents a non-auto property insurance combination plan;

[0257] Compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination plan, and determine the initial individual optimal position of each particle based on the matching degree analysis result;

[0258] In each iteration of the particle swarm optimization method, the speed and position of each particle in the particle swarm are updated, and the matching degree analysis is re-performed for each updated particle position;

[0259] According to the re-determined matching degree analysis results, the individual optimal position of each particle and the global optimal position of the particle group are updated;

[0260] After the iteration is completed, the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm is selected as the final output.

[0261] 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:

[0262] Analyze customers’ personal information, historical purchase records, and browsing behaviors, build customer profiles, and extract customer profile features including age, income level, and risk tolerance;

[0263] Extracting insurance features of each insurance product from the non-auto property insurance product database, wherein the insurance features include suitable population categories, insurance identification values, and coverage;

[0264] Initialize particles in the particle swarm optimization method and randomly generate multiple particles, each particle represents a non-auto property insurance combination plan;

[0265] Compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination plan, and determine the initial individual optimal position of each particle based on the matching degree analysis result;

[0266] In each iteration of the particle swarm optimization method, the speed and position of each particle in the particle swarm are updated, and the matching degree analysis is re-performed for each updated particle position;

[0267] According to the re-determined matching degree analysis results, the individual optimal position of each particle and the global optimal position of the particle group are updated;

[0268] After the iteration is completed, the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm is selected as the final output.

[0269] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the user side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0270] 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 in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. 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).

[0271] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0272] It should be noted that if software tools or components other than those of the Company appear in the embodiments of the present application, they are only used for illustration and do not represent actual use. The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the above-mentioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the above-mentioned embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents; and these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for outputting a non-vehicle property product combination, characterized in that: The following steps are involved: Analyze customers’ personal information, historical purchase records, and browsing behaviors, build customer profiles, and extract customer profile features including age, income level, and risk tolerance; Extracting insurance features of each insurance product from the non-auto property insurance product database, wherein the insurance features include suitable population categories, insurance identification values, and coverage; Initialize particles in the particle swarm optimization method and randomly generate multiple particles, each particle represents a non-auto property insurance combination plan; Compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination plan, and determine the initial individual optimal position of each particle based on the matching degree analysis result; In each iteration of the particle swarm optimization method, the speed and position of each particle in the particle swarm are updated, and the matching degree analysis is re-performed for each updated particle position; According to the re-determined matching analysis results, the individual optimal position of each particle and the global optimal position of the particle group are updated; After the iteration is completed, the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm is selected as the final output.

2. The method for outputting a non-vehicle property product combination as claimed in claim 1, characterized in that: Analyze customers’ personal information, historical purchase records, and browsing behaviors to build customer profiles and extract customer profile features including age, income level, and risk tolerance, including: Collect basic personal information of customers, including age, gender and income; Retrieving the customer's historical purchase records from an insurance database, wherein the historical purchase records include the type and quantity of insurance products purchased; Tracking the browsing behavior of customers on the insurance platform through the interaction data of the insurance platform, the browsing behavior including the type of insurance products browsed, browsing frequency and browsing duration; Integrate the basic personal information, historical purchase records and browsing behaviors to generate a comprehensive customer profile; Customer profile features are extracted from the comprehensive customer profile, wherein the customer profile features include age, income level, and risk tolerance.

3. The method for outputting a non-vehicle property product combination as claimed in claim 1, characterized in that: The insurance characteristics of each insurance product are extracted from the non-auto property insurance product database, wherein the insurance characteristics include suitable population categories, insurance identification values ​​and coverage, including: Obtaining an insurance product list from the non-auto property insurance product database, wherein the insurance product list includes information of all insurance products belonging to the non-auto property insurance category; Classify each insurance product according to its nature and terms and determine the corresponding suitable population category; Extract and record the corresponding insurance identification value from the terms of each insurance product, wherein the insurance identification value includes a basic identification value and an additional identification value; Determine the coverage of each insurance product based on the terms of each insurance product, including the insurance liability content, liability exemption clauses and claim identification value limits; The suitable population category, insurance identification value and coverage scope of each insurance product are feature-labeled to generate insurance features for each insurance product.

4. The method for outputting a non-vehicle property product combination as claimed in claim 1, characterized in that: In the particle swarm optimization method, particles are initialized and multiple particles are randomly generated. Each particle represents a non-auto property insurance combination scheme, including: Determine the initial number of particles in the particle swarm to set the particle swarm size; The positions of particles are randomly distributed in the defined solution space, and the position of each particle corresponds to a non-auto property insurance combination plan; Assigning an initial velocity to each particle, wherein the initial velocity represents the moving direction and speed of the particle in the solution space; Initialize the individual optimal position of each particle as the initial position of each particle.

5. The method for outputting a non-vehicle property product combination as claimed in claim 1, characterized in that: Compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer profile characteristics, and analyze the matching degree of each non-auto property insurance combination plan, including: According to the preset population matching standard, the matching score between the suitable population category characteristics of the non-auto property insurance combination plan in each particle and the customer profile characteristics is analyzed; According to the preset insurance matching standard, the matching score between the insurance identification value characteristics of the non-auto property insurance combination plan in each particle and the customer portrait characteristics is analyzed; According to the preset risk tolerance matching standard, the matching score between the coverage characteristics of the non-auto property insurance combination plan in each particle and the customer profile characteristics is analyzed; The matching scores of each particle's non-auto property insurance combination plan in terms of suitable population category, insurance identification value and coverage scope are summarized to determine the overall matching score of each particle.

6. The method for outputting a non-vehicle property product combination as claimed in claim 1, characterized in that: In each iteration of the particle swarm optimization method, the speed and position of each particle in the particle swarm are updated, and the matching degree analysis is re-performed for each updated particle position, including: At the beginning of each iteration, the current speed of each particle is recorded, and the current speed is the speed of the particle at the end of the previous iteration; Use inertia weight to adjust the current speed of each particle to maintain the original movement trend of each particle; According to the individual optimal position of each particle and the global optimal position of the particle group, the speed direction and speed magnitude of each particle are adjusted; Determine the new position of each particle according to the adjusted velocity direction and velocity magnitude; Based on the new position of each particle, the non-auto property insurance combination plan represented by each particle is adjusted, and the matching degree analysis is performed again.

7. The method for outputting a non-vehicle property product combination as claimed in claim 1, characterized in that: According to the re-determined matching analysis results, the individual optimal position of each particle and the global optimal position of the particle swarm are updated, including: Compare the re-determined current matching degree analysis result of each particle with the historical optimal matching degree analysis result of each particle; If the current matching degree analysis result is better than the historical optimal matching degree analysis result, the individual optimal position of the particle is updated to the current particle position; Comparing the updated individual optimal position of the particle with the global optimal position of the particle swarm; If the individual optimal position of a particle is better than the global optimal position of the particle swarm, the global optimal position of the particle swarm is updated to the individual optimal position of the particle.

8. A non-automotive product combination output device, characterized in that: The non-vehicle property product combination output device includes: The customer profile building module is used to analyze the customer's personal information, historical purchase records and browsing behavior, build a customer profile, and extract customer profile features including age, income level and risk tolerance; An insurance feature extraction module is used to extract the insurance features of each insurance product from the non-auto property insurance product library, wherein the insurance features include the category of suitable people, the insurance identification value and the coverage; A particle initialization module is used to initialize particles in the particle swarm optimization method and randomly generate multiple particles, each particle representing a non-auto property insurance combination plan; A matching degree analysis module, used to compare the insurance characteristics of the non-auto property insurance combination plan in each particle with the customer portrait characteristics, analyze the matching degree of each non-auto property insurance combination plan, and determine the initial individual optimal position of each particle based on the matching degree analysis result; The particle swarm iteration update module is used to update the speed and position of each particle in the particle swarm during each iteration of the particle swarm optimization method, and re-analyze the matching degree of each updated particle position; The optimal position updating module is used to update the individual optimal position of each particle and the global optimal position of the particle group according to the re-determined matching degree analysis result; The final combination output module is used to select the non-auto property insurance combination plan corresponding to the global optimal position of the particle swarm as the final output after the iteration is completed.

9. A computer device, characterized in that: The computer device includes a memory, a processor, and a non-vehicle property product combination output program stored in the memory and executable on the processor. When the non-vehicle property product combination output program is executed by the processor, the steps of the non-vehicle property product combination output method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The storage medium stores a non-vehicle property product combination output program, and when the non-vehicle property product combination output program is executed by the processor, the steps of the non-vehicle property product combination output method as described in any one of claims 1-7 are implemented.

Images