Vehicle insurance product recommendation method, device, equipment and storage medium

Abstract

The embodiments of the present application belong to the field of big data, and relate to a method, device, equipment, and storage medium for recommending vehicle insurance products. The method includes obtaining refueling data, claims data, and historical vehicle insurance product sales data of a refueling customer group; fusing the refueling data, and obtaining individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm; comprehensively analyzing the claims data and the historical vehicle insurance product sales data with the individual behavior pattern characteristics, and labeling the refueling customer group based on the analysis results; and recommending vehicle insurance products to the refueling customer group based on a preset association rule mining algorithm, a preset label recommendation algorithm based on the labels of the refueling customer group, and a preset collaborative filtering algorithm. The present application can improve the recommendation effect of vehicle insurance products.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a method, device, equipment and storage medium for recommending vehicle insurance products. Background Art

[0002] With the continuous development of insurance e-commerce, the marketing of insurance products has become an important task for major insurance companies. Compared with traditional insurance sales, the popular Internet sales now lower the threshold for users to buy insurance, and make up for the shortcomings of traditional agent services through the combination of finance and technology. In the era of rapid development of big data, how to effectively combine large amounts of data with insurance sales and recommend different insurance products to different customers, thereby improving marketing quality and increasing customer loyalty, is the focus of insurance companies. When making personalized recommendations for customers, improving customer value is the key, so the recommended products are not about quantity, but about refinement. At present, when making product recommendations, historical sales data is generally used. The data source is relatively single and cannot be fully recommended, resulting in low customer satisfaction and poor recommendation results. Summary of the invention

[0003] The purpose of the embodiments of the present application is to propose a method, device, equipment and storage medium for recommending vehicle insurance products to solve the problem of poor recommendation effect of existing recommendation methods.

[0004] In order to solve the above technical problems, the embodiment of the present application provides a method for recommending vehicle insurance products, which adopts the following technical solutions:

[0005] Obtain refueling data, claims data, and historical vehicle insurance product sales data of refueling customers;

[0006] Performing data fusion on the refueling data, and obtaining individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm;

[0007] Comprehensively analyzing the claims data and the historical vehicle insurance product sales data with the individual behavior pattern characteristics, and labeling the refueling customer group according to the analysis results;

[0008] The vehicle insurance product is recommended to the refueling customer group based on a preset association rule mining algorithm, a preset tag recommendation algorithm based on tags of the refueling customer group, and a preset collaborative filtering algorithm.

[0009] In order to solve the above technical problems, the embodiment of the present application further provides a vehicle insurance product recommendation device, which adopts the following technical solution:

[0010] The data acquisition module is used to obtain the refueling data, claims data and historical vehicle insurance product sales data of the refueling customer group;

[0011] A data fusion module, used to perform data fusion on the refueling data and obtain individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm;

[0012] An analysis module, for comprehensively analyzing the claims data and the historical vehicle insurance product sales data with the individual behavior pattern characteristics, and labeling the refueling customer group according to the analysis results;

[0013] The recommendation module is used to recommend vehicle insurance products to the refueling customer group based on a preset association rule mining algorithm, a preset tag recommendation algorithm based on the tags of the refueling customer group, and a preset collaborative filtering algorithm.

[0014] In order to solve the above technical problems, an embodiment of the present application also provides a computer device, including a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the vehicle insurance product recommendation method described in the above embodiment are implemented.

[0015] In order to solve the above technical problems, an embodiment of the present application also provides a computer-readable storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor, the steps of the vehicle insurance product recommendation method described in the above embodiment are implemented.

[0016] Compared with the prior art, the embodiment of the present application has the following beneficial effects: by obtaining the refueling data, claims data and historical vehicle insurance product sales data of the refueling customer group; performing data fusion on the refueling data, and obtaining the individual behavior pattern characteristics of the refueling customer group based on the preset ant colony clustering algorithm; comprehensively analyzing the claims data and historical vehicle insurance product sales data and individual behavior pattern characteristics, and labeling the refueling customer group according to the analysis results; recommending vehicle insurance products to the refueling customer group based on the preset association rule mining algorithm, the preset label recommendation algorithm based on the labels of the refueling customer group, and the preset collaborative filtering algorithm. In this way, the vehicle insurance products are personalized recommended by combining the refueling data, claims data and historical vehicle insurance product sales data of the refueling customer group, and the combination of multiple algorithms of the preset association rule mining algorithm, the preset label recommendation algorithm and the preset collaborative filtering algorithm. The diversity of the recommended data source and the recommended algorithm is guaranteed, and the characteristics of the insurance product itself, the customer's refueling data, the claims data and other data can be fully utilized to recommend products, and the continuous updating of customer labels makes the recommendation more real-time and reliable, further improving the recommendation effect of vehicle insurance products. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flowchart of an embodiment of a method for recommending a vehicle insurance product according to the present application;

[0019] Figure 2 yes Figure 1 A flowchart of a specific implementation of step S102;

[0020] Figure 3 yes Figure 1 A flowchart of a specific implementation of step S103;

[0021] Figure 4 yes Figure 3 A flowchart of a specific implementation of step S303;

[0022] Figure 5 yes Figure 1 A flowchart of a specific implementation of step S104;

[0023] Figure 6 is a structural schematic diagram of an embodiment of a recommendation device for a vehicle insurance product according to the present application;

[0024] Figure 7 yes Figure 6 A structural schematic diagram of a specific implementation of the data fusion module shown;

[0025] Figure 8 yes Figure 6 A schematic structural diagram of a specific implementation of the analysis module shown;

[0026] Fig. 9 yes Figure 8 A schematic structural diagram of a specific implementation of the tag unit shown;

[0027] Fig.10 yes Figure 6 A schematic diagram of the structure of a specific implementation of the recommendation module shown;

[0028] Fig.11 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0030] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0031] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0032] Continue to refer Figure 1 , showing a flow chart of an embodiment of a method for recommending vehicle insurance products according to the present application. The method for recommending vehicle insurance products comprises the following steps:

[0033] Step S101, obtaining refueling data, claims data and historical vehicle insurance product sales data of refueling customers.

[0034] Among them, refueling data includes static refueling data and dynamic refueling data. Static refueling data includes user information, vehicle information and basic information of gas stations of refueling customers. User information may include user account, contact information, etc. Vehicle information may include license plate number, vehicle type, vehicle model, etc. Gas station information includes gas station name, gas station location, oil type and oil price, etc.

[0035] Dynamic refueling data includes time series data, trajectory data, and some external factors (such as weather and holidays) generated by the refueling behavior of refueling customers. Time series refers to the time series of each refueling of refueling customers. Trajectory data refers to the refueling trajectory of refueling customers, such as which gas stations refueling customers often refuel at and the refueling trajectory formed by these gas stations.

[0036] Static refueling data and dynamic refueling data may be stored in a database of the refueling data system.

[0037] The claims data is the past claims data of the refueling customers. The claims data can be stored in the database of the claims data system.

[0038] The historical vehicle insurance product sales data includes the time characteristics and consumption characteristics of the vehicle insurance products purchased by the refueling customer group in the past. The above-mentioned vehicle insurance product may refer to a vehicle insurance product. The historical vehicle insurance product sales data may be stored in a database of a vehicle insurance product data system.

[0039] Specifically, the corresponding refueling data, claims data and historical vehicle insurance product sales data can be obtained from the database of the corresponding big data system.

[0040] Step S102, performing data fusion on the refueling data, and obtaining individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm.

[0041] Among them, individual behavior pattern characteristics include refueling consumption characteristics and refueling time characteristics.

[0042] In this embodiment, if Figure 2 As shown, step S102 specifically includes:

[0043] Step S201, fusing static refueling data and dynamic refueling data to obtain fused data.

[0044] Specifically, the user information, vehicle information and basic information of the refueling customer group in the static refueling data of the refueling customer group, as well as the time series data, trajectory data and some external factors (such as weather, holidays) generated by the refueling behavior of the refueling customer group in the dynamic refueling data of the refueling customer group can be fused. The specific data fusion method is: feature extraction is performed on the static refueling data and the dynamic refueling data respectively, and the extracted feature information should be a full representation of the original data. Then, according to the feature information, the static refueling data and the dynamic refueling data are classified and clustered respectively, and some feature extraction methods (such as PCA principal component analysis, which finds the most dispersed features from high-dimensional data along this feature dimension) are used to extract valuable features as the input of the preset ant colony clustering algorithm. For example, the fusion is performed from the dimensions of each refueling customer's sensitivity to price, oil product preference, refueling on holidays, etc. Of course, it can also be fused from the dimensions of gas stations, weather, holidays, gas station trajectories, etc. Thus, the fused data corresponding to each refueling customer is obtained.

[0045] Step S202: performing cluster analysis on the fused data based on a preset ant colony clustering algorithm to obtain a clustering result.

[0046] Among them, the preset ant colony clustering algorithm is a pre-set ant colony clustering algorithm. The ant colony clustering algorithm performs clustering analysis based on the foraging behavior of ants. It regards data as ants with different attributes and the clustering results as food sources. Each ant moves with a certain probability and gathers at different food sources to achieve clustering. Through this clustering algorithm, customers with similar data features extracted earlier can be clustered together, and customers with large differences in data features can be separated as much as possible, which is helpful for in-depth analysis of customers belonging to the same category with the help of other data sources in the future, and enhances the accuracy of recommendations.

[0047] Specifically, the fusion data of multiple customers in the refueling customer group is analyzed by using a preset ant colony clustering algorithm, so as to obtain the refueling situation of different refueling customers in the refueling customer group. The refueling situation includes the refueling consumption situation and the refueling time situation.

[0048] For example, there are N customers, each of whom has the data features extracted above, a total of n. The purpose of the ant colony clustering algorithm is to minimize the sum of the distances from each customer to the cluster center. The main process of the algorithm is: 1. Initialize the ant colony parameters; 2. Construct the pheromone matrix and the objective function; 3. Each ant selects the one with the highest pheromone according to the current pheromone matrix when judging the belonging of the samples in its own solution set, and then calculates the target value of each ant according to the objective function, selects the best ant for local search, updates the pheromone table, and reaches the global optimal solution after multiple iterations.

[0049] Step S203, obtaining the refueling consumption characteristics and refueling time characteristics of the refueling customer group according to the clustering result.

[0050] Specifically, the refueling consumption characteristics and refueling time characteristics corresponding to each refueling customer are obtained according to the refueling consumption situation and refueling time situation of each refueling customer in the refueling customer group.

[0051] For example, through the preset ant colony clustering algorithm, cluster analysis is performed on the payment methods preferred by refueling customers, such as WeChat Pay or Alipay, and whether the consumption trend of the total amount of refueling is gradually increasing, gradually decreasing, or constantly fluctuating. Then, the consumption characteristics of the refueling customers are analyzed from the clustering results to obtain the corresponding refueling consumption characteristics.

[0052] By presetting the ant colony clustering algorithm, it is possible to identify the refueling time characteristics of the refueling customer groups, such as those who like to refuel at night or in the early morning, on weekdays or non-working days, and thus obtain the corresponding refueling time characteristics of the refueling customer groups.

[0053] These two types of features are the results of clustering based on the ant colony clustering algorithm. By analyzing the customers in the same customer group, the refueling consumption characteristics and refueling time characteristics of this refueling customer group are obtained.

[0054] In this embodiment, the consumption characteristics and time characteristics of the refueling customer group can be combined according to the multi-dimensional data in the refueling data to obtain the corresponding individual behavior pattern characteristics, so as to accurately recommend vehicle insurance products to the refueling customer group, thereby improving the recommendation effect of vehicle insurance products.

[0055] Step S103, comprehensively analyzing the claims data and historical vehicle insurance product sales data with individual behavior pattern characteristics, and labeling the refueling customer group based on the analysis results.

[0056] Among them, the claims data includes claims time characteristics and claims consumption characteristics; the historical vehicle insurance product sales data includes historical vehicle insurance product time characteristics and historical vehicle insurance product consumption characteristics.

[0057] Specifically, we can predict the refueling customer group based on their claims and the types of vehicle insurance products they have purchased in the past, and evaluate the probability of accidents, so as to label them accordingly. For example, customers who like to refuel at night are likely to be online ride-hailing vehicles, and their probability of accidents is generally high. We can label them as high-risk customers, so that when recommending vehicle insurance products, we can recommend accident insurance products to them.

[0058] In the embodiment of the present invention, Figure 3 As shown, step S103 specifically includes:

[0059] Step S301 , compare and analyze the claim settlement time characteristics, the historical vehicle insurance product time characteristics and the refueling time characteristics to obtain the time characteristic analysis result.

[0060] The claim time feature may be the time interval feature of each claim, for example, the length of the time interval between adjacent claims. Specifically, the risk may be graded according to the time interval between claims and the number of claims, and may be divided into high risk, medium risk, and low risk, etc. When the number of claims and the time interval between adjacent claims both meet the high risk range, it means that the customer is a high risk customer. Similarly, when the number of claims and the time interval between adjacent claims both meet the medium risk range, it means that the customer is a medium risk customer. When the number of claims and the time interval between adjacent claims both meet the low risk range, it means that the customer is a low risk customer.

[0061] At the same time, it is also necessary to analyze the time characteristics of the customer's purchase of vehicle insurance products based on the customer's claim time characteristics. This can determine whether the vehicle insurance product purchased by the customer is within the validity period.

[0062] We also need to analyze the customer's refueling time. For example, if the customer refuels at a relatively short interval, the customer uses the vehicle frequently. Or, customers who like to refuel at night are likely to be online ride-hailing vehicles, and the probability of accidents is generally higher. We can recommend accident insurance products to them.

[0063] In this way, the customer's time characteristic analysis results can be obtained by combining the analysis of claim time characteristics, historical vehicle insurance product time characteristics and refueling time characteristics.

[0064] Step S302, comprehensively analyzing the claim settlement consumption characteristics, historical vehicle insurance product consumption characteristics, and refueling consumption characteristics to obtain consumption characteristic analysis results.

[0065] The claim consumption feature may be the amount feature of each claim, such as the reimbursement amount feature of each claim. The above historical vehicle insurance product consumption feature may be the amount feature of the vehicle insurance product purchased and the corresponding insured amount feature. The above refueling consumption feature may be the amount feature of each refueling.

[0066] Specifically, the probability of a customer's accident can be determined by analyzing the customer's claim amount each time, the amount and insured amount of vehicle insurance products purchased, and the amount of each refueling, thereby obtaining the customer's consumption characteristic analysis results.

[0067] Step S303, labeling different refueling customer groups based on the time feature analysis results and the consumption feature analysis results.

[0068] Specifically, after determining the risk type of a customer, the customer can be labeled with a corresponding risk label.

[0069] In the embodiment of the present invention, Figure 4 As shown, step S303 specifically includes:

[0070] Step S401, classify the refueling customer group based on the time feature analysis result and the consumption feature analysis result to obtain the classification result.

[0071] Step S402: assign different labels to different refueling customer groups according to the classification results.

[0072] Specifically, based on the time feature analysis results and consumption feature analysis results, customers in the refueling customer group can be classified into high-risk customers, medium-risk customers, and low-risk customers, and then corresponding labels can be attached to the customers.

[0073] In this way, suitable vehicle insurance products can be recommended to different groups of refueling customers, thereby improving the recommendation effect of vehicle insurance products.

[0074] Step S104, recommending vehicle insurance products to the refueling customer group based on a preset association rule mining algorithm, a preset tag recommendation algorithm based on tags of the refueling customer group, and a preset collaborative filtering algorithm.

[0075] The above-mentioned preset association rule mining algorithm is a preset association rule mining algorithm based on FP Tree (Frequent Pattern tree).

[0076] In the embodiment of the present invention, Figure 5 As shown, step S104 specifically includes:

[0077] Step S501, extracting the association relationship between the historical vehicle insurance products and vehicle insurance products of the refueling customer group based on a preset association rule mining algorithm, taking the vehicle insurance product that meets the association relationship as the first candidate recommended vehicle insurance product, and performing a recommendation score.

[0078] Specifically, through the association rule mining algorithm based on FP Tree, the frequent item sets between the historical vehicle insurance products and vehicle insurance products of the refueling customer group are obtained, that is, which vehicle insurance products have been purchased more than the specified threshold, and then through FP Tree, the conditional pattern base, conditional FP tree and frequent patterns corresponding to each vehicle insurance product can be obtained, so as to obtain the association relationship between the historical vehicle insurance products and vehicle insurance products of the refueling customer group. Thus, the corresponding association rules are obtained, that is, the possibility of purchasing several vehicle insurance products at the same time is very high, so the vehicle insurance products that meet the association rules can be used as candidate recommended vehicle insurance products. At the same time. These candidate recommended vehicle insurance products can be scored separately. The scoring can be a comprehensive scoring based on the purchase rate, usage rate, etc.

[0079] Step S502: Based on a preset tag recommendation algorithm, the vehicle insurance product corresponding to the tag of the refueling customer group is selected as a second candidate recommended vehicle insurance product and a recommendation score is performed.

[0080] Among them, the preset tag recommendation algorithm is a pre-set tag-based recommendation algorithm. The preset tag recommendation algorithm specifically extracts key features from different dimensions based on some behavioral feature data of customers and labels them. In essence, it is actually a user portrait, but the features are more specific, detailed, and diverse. For example, based on refueling data, customer loyalty (the customer's loyalty is judged based on the distribution of the number of refueling times) and consumption trends (the total refueling amount shows an increasing, unstable, or decreasing trend) and other label dimensions can be extracted. Similar customers can be identified based on the similarity of different customer labels, so as to recommend vehicle insurance products.

[0081] Specifically, by using a preset tag recommendation algorithm, vehicle insurance products purchased by similar customers can be used as candidate recommended vehicle insurance products based on the customer's tag, and the obtained candidate recommended vehicle insurance products can be scored respectively. The scoring method can be to score the vehicle insurance products purchased by the customer according to the degree of similarity of the customer's tag. The vehicle insurance products purchased by the customer with the highest tag similarity have relatively high scores, and the vehicle insurance products purchased by the customer with low tag similarity have relatively low scores, etc.

[0082] Step S503: Based on a preset collaborative filtering algorithm, a vehicle insurance product having the same historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics is selected as a third candidate recommended vehicle insurance product, and a recommendation score is performed.

[0083] The preset collaborative filtering algorithm is a pre-set collaborative filtering algorithm. The collaborative filtering algorithm is based on mining the customer's historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics to discover the customer's preference and predict the products that the customer may like for recommendation.

[0084] When recommendations are made based on preset collaborative filtering algorithms, they are mainly divided into two categories, namely customer-based collaborative filtering algorithms and item-based collaborative filtering algorithms. The former discovers customers' preferences for products through their historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics, measures and scores these preferences, calculates the relationship between customers based on their attitudes and preferences for the same product, and recommends products to customers with the same preferences. The latter obtains the relationship between products by calculating the scores of different customers for different products, and recommends similar items to customers based on the relationship between products. For example, if customer A buys product 1 and product 2, it means that the two products have a high degree of correlation. Therefore, when customer B buys product 1, product 2 can be recommended to him.

[0085] Specifically, the preset collaborative filtering algorithm can be used to find vehicle insurance products with the same preferences of the same customers as candidate recommended vehicle insurance products. The obtained candidate recommended vehicle insurance products are scored. The scoring method can be to score according to the scores of the customers on the vehicle insurance products.

[0086] Step S504, obtain the final recommendation scores corresponding to the first candidate recommended vehicle insurance product, the second candidate recommended vehicle insurance product, and the third candidate recommended vehicle insurance product by weighted combination, and select the top n candidate recommended vehicle insurance products based on the final recommendation scores, and recommend them to the refueling customer group as a recommendation list.

[0087] Specifically, the three algorithms recommend candidate vehicle insurance products and their scores respectively, and then obtain the final scores through weighted combination, thereby obtaining the final vehicle insurance product recommendation list. The weights are equal at the beginning, and the weights are continuously adjusted through multiple combined recommendations and actual customer purchases to meet the needs of most customers.

[0088] The intersection of the products to be recommended obtained by the three algorithms is taken to obtain the corresponding candidate products and the scores given by the corresponding recommendation algorithms. Then all the scores are weighted and combined to obtain the final score of each product to be recommended. The top n scores are selected from them and recommended to the refueling customers as a recommendation list.

[0089] In the embodiment of the present invention, the results obtained by three recommendation algorithms, namely, a preset association rule mining algorithm, a preset tag recommendation algorithm, and a preset collaborative filtering algorithm, can be combined to achieve personalized recommendation of vehicle insurance products.

[0090] This application obtains the refueling data, claims data and historical vehicle insurance product sales data of the refueling customer group; performs data fusion on the refueling data, and obtains the individual behavior pattern characteristics of the refueling customer group based on the preset ant colony clustering algorithm; comprehensively analyzes the claims data and historical vehicle insurance product sales data and individual behavior pattern characteristics, and labels the refueling customer group according to the analysis results; recommends vehicle insurance products to the refueling customer group based on the preset association rule mining algorithm, the preset label recommendation algorithm based on the labels of the refueling customer group, and the preset collaborative filtering algorithm. In this way, the refueling data, claims data and historical vehicle insurance product sales data of the refueling customer group are combined with multiple data sources, as well as the preset association rule mining algorithm, the preset label recommendation algorithm and the preset collaborative filtering algorithm. The combination of multiple algorithms makes personalized recommendations for vehicle insurance products. It ensures the diversity of the recommended data sources and recommendation algorithms, and can make full use of the characteristics of the insurance product itself, the customer's refueling data, claims data and other data to recommend products, and with the help of the continuous updating of customer labels, the recommendation is more real-time and reliable, further improving the recommendation effect of vehicle insurance products.

[0091] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0092] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0093] Further references Figure 6 , as a response to the above Figure 1 The present application provides an embodiment of a vehicle insurance product recommendation device, and the device embodiment is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0094] like Figure 6 As shown, the vehicle insurance product recommendation device 600 described in this embodiment includes: a data acquisition module 601, a data fusion module 602, an analysis module 603 and a recommendation module 604. Among them:

[0095] The data acquisition module 601 is used to acquire the refueling data, claim settlement data and historical vehicle insurance product sales data of the refueling customer group.

[0096] Among them, refueling data includes static refueling data and dynamic refueling data. Static refueling data includes user information, vehicle information and basic information of gas stations of refueling customers. User information may include user account, contact information, etc. Vehicle information may include license plate number, vehicle type, vehicle model, etc. Gas station information includes gas station name, gas station location, oil type and oil price, etc.

[0097] Dynamic refueling data includes time series data, trajectory data, and some external factors (such as weather and holidays) generated by the refueling behavior of refueling customers. Time series refers to the time series of each refueling of refueling customers. Trajectory data refers to the refueling trajectory of refueling customers, such as which gas stations refueling customers often refuel at and the refueling trajectory formed by these gas stations.

[0098] Static refueling data and dynamic refueling data may be stored in a database of the refueling data system.

[0099] The claims data is the past claims data of the refueling customers. The claims data can be stored in the database of the claims data system.

[0100] The historical vehicle insurance product sales data includes the time characteristics and consumption characteristics of the vehicle insurance products purchased by the refueling customer group in the past. The above-mentioned vehicle insurance product may refer to a vehicle insurance product. The historical vehicle insurance product sales data may be stored in a database of a vehicle insurance product data system.

[0101] Specifically, the corresponding refueling data, claims data and historical vehicle insurance product sales data can be obtained from the database of the corresponding big data system.

[0102] The data fusion module 602 is used to perform data fusion on the refueling data and obtain individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm.

[0103] Among them, individual behavior pattern characteristics include refueling consumption characteristics and refueling time characteristics.

[0104] In this embodiment, if Figure 7 As shown, the data fusion module 602 specifically includes:

[0105] The data fusion unit 6021 is used to fuse the static refueling data and the dynamic refueling data to obtain fused data.

[0106] Specifically, the user information, vehicle information and basic information of the refueling customer group in the static refueling data of the refueling customer group, as well as the time series data, trajectory data and some external factors (such as weather, holidays) generated by the refueling behavior of the refueling customer group in the dynamic refueling data of the refueling customer group can be fused. The specific data fusion method is: feature extraction is performed on the static refueling data and the dynamic refueling data respectively, and the extracted feature information should be a full representation of the original data. Then, according to the feature information, the static refueling data and the dynamic refueling data are classified and clustered respectively, and some feature extraction methods (such as PCA principal component analysis, which finds the most dispersed features from high-dimensional data along this feature dimension) are used to extract valuable features as the input of the preset ant colony clustering algorithm. For example, the fusion is performed from the dimensions of each refueling customer's sensitivity to price, oil product preference, refueling on holidays, etc. Of course, it can also be fused from the dimensions of gas stations, weather, holidays, gas station trajectories, etc. Thus, the fused data corresponding to each refueling customer is obtained.

[0107] The clustering analysis unit 6022 is used to perform cluster analysis on the fusion data based on the preset ant colony clustering algorithm to obtain a clustering result.

[0108] Among them, the preset ant colony clustering algorithm is a pre-set ant colony clustering algorithm. The ant colony clustering algorithm performs clustering analysis based on the foraging behavior of ants. It regards data as ants with different attributes and the clustering results as food sources. Each ant moves with a certain probability and gathers at different food sources to achieve clustering. Through this clustering algorithm, customers with similar data features extracted earlier can be clustered together, and customers with large differences in data features can be separated as much as possible, which is helpful for in-depth analysis of customers belonging to the same category with the help of other data sources in the future, and enhances the accuracy of recommendations.

[0109] Specifically, the fusion data of multiple customers in the refueling customer group is analyzed by using a preset ant colony clustering algorithm, so as to obtain the refueling situation of different refueling customers in the refueling customer group. The refueling situation includes the refueling consumption situation and the refueling time situation.

[0110] For example, there are N customers, each of whom has the data features extracted above, a total of n. The purpose of the ant colony clustering algorithm is to minimize the sum of the distances from each customer to the cluster center. The main process of the algorithm is: 1. Initialize the ant colony parameters; 2. Construct the pheromone matrix and the objective function; 3. Each ant selects the one with the highest pheromone according to the current pheromone matrix when judging the belonging of the samples in its own solution set, and then calculates the target value of each ant according to the objective function, selects the best ant for local search, updates the pheromone table, and reaches the global optimal solution after multiple iterations.

[0111] The feature acquisition unit 6023 is used to acquire the refueling consumption characteristics and refueling time characteristics of the refueling customer group according to the clustering result.

[0112] Specifically, the refueling consumption characteristics and refueling time characteristics corresponding to each refueling customer are obtained according to the refueling consumption situation and refueling time situation of each refueling customer in the refueling customer group.

[0113] For example, through the preset ant colony clustering algorithm, cluster analysis is performed on the payment methods preferred by refueling customers, such as WeChat Pay or Alipay, and whether the consumption trend of the total amount of refueling is gradually increasing, gradually decreasing, or constantly fluctuating. Then, the consumption characteristics of the refueling customers are analyzed from the clustering results to obtain the corresponding refueling consumption characteristics.

[0114] By presetting the ant colony clustering algorithm, it is possible to identify the refueling time characteristics of the refueling customer groups, such as those who like to refuel at night or in the early morning, on weekdays or non-working days, and thus obtain the corresponding refueling time characteristics of the refueling customer groups.

[0115] These two types of features are the results of clustering based on the ant colony clustering algorithm. By analyzing the customers in the same customer group, the refueling consumption characteristics and refueling time characteristics of this refueling customer group are obtained.

[0116] In this embodiment, the consumption characteristics and time characteristics of the refueling customer group can be combined according to the multi-dimensional data in the refueling data to obtain the corresponding individual behavior pattern characteristics, so as to accurately recommend vehicle insurance products to the refueling customer group, thereby improving the recommendation effect of vehicle insurance products.

[0117] The analysis module 603 is used to comprehensively analyze the claim data and the historical vehicle insurance product sales data with the individual behavior pattern characteristics, and label the refueling customer group according to the analysis results.

[0118] Among them, the claims data includes claims time characteristics and claims consumption characteristics; the historical vehicle insurance product sales data includes historical vehicle insurance product time characteristics and historical vehicle insurance product consumption characteristics.

[0119] Specifically, we can predict the refueling customer group based on their claims and the types of vehicle insurance products they have purchased in the past, and evaluate the probability of accidents, so as to label them accordingly. For example, customers who like to refuel at night are likely to be online ride-hailing vehicles, and their probability of accidents is generally high. We can label them as high-risk customers, so that when recommending vehicle insurance products, we can recommend accident insurance products to them.

[0120] In the embodiment of the present invention, Figure 8 As shown, the analysis module 603 includes:

[0121] The first analysis unit 6031 is used to compare and analyze the claim settlement time characteristics, the historical vehicle insurance product time characteristics and the refueling time characteristics to obtain a time characteristic analysis result.

[0122] The claim time feature may be the time interval feature of each claim, for example, the length of the time interval between adjacent claims. Specifically, the risk may be graded according to the time interval between claims and the number of claims, and may be divided into high risk, medium risk, and low risk, etc. When the number of claims and the time interval between adjacent claims both meet the high risk range, it means that the customer is a high risk customer. Similarly, when the number of claims and the time interval between adjacent claims both meet the medium risk range, it means that the customer is a medium risk customer. When the number of claims and the time interval between adjacent claims both meet the low risk range, it means that the customer is a low risk customer.

[0123] At the same time, it is also necessary to analyze the time characteristics of the customer's purchase of vehicle insurance products based on the customer's claim time characteristics. This can determine whether the vehicle insurance product purchased by the customer is within the validity period.

[0124] We also need to analyze the customer's refueling time. For example, if the customer refuels at a relatively short interval, the customer uses the vehicle frequently. Or, customers who like to refuel at night are likely to be online ride-hailing vehicles, and the probability of accidents is generally higher. We can recommend accident insurance products to them.

[0125] In this way, the customer's time characteristic analysis results can be obtained by combining the analysis of claim time characteristics, historical vehicle insurance product time characteristics and refueling time characteristics.

[0126] The second analysis unit 6032 is used to comprehensively analyze the claim consumption characteristics, the historical vehicle insurance product consumption characteristics and the refueling consumption characteristics to obtain a consumption characteristic analysis result.

[0127] The claim consumption feature may be the amount feature of each claim, such as the reimbursement amount feature of each claim. The above historical vehicle insurance product consumption feature may be the amount feature of the vehicle insurance product purchased and the corresponding insured amount feature. The above refueling consumption feature may be the amount feature of each refueling.

[0128] Specifically, the probability of a customer's accident can be determined by analyzing the customer's claim amount each time, the amount and insured amount of vehicle insurance products purchased, and the amount of each refueling, thereby obtaining the customer's consumption characteristic analysis results.

[0129] The label unit 6033 is used to label different refueling customer groups based on the time feature analysis result and the consumption feature analysis result.

[0130] Specifically, after determining the risk type of a customer, the customer can be labeled with a corresponding risk label.

[0131] In the embodiment of the present invention, Fig. 9 As shown, the label unit 6033 includes:

[0132] The classification subunit 60331 is used to classify the refueling customer group in combination with the time feature analysis result and the consumption feature analysis result to obtain a classification result.

[0133] The label subunit 60332 is used to label different refueling customer groups with different labels according to the classification results.

[0134] Specifically, based on the time feature analysis results and consumption feature analysis results, customers in the refueling customer group can be classified into high-risk customers, medium-risk customers, and low-risk customers, and then corresponding labels can be attached to the customers.

[0135] In this way, suitable vehicle insurance products can be recommended to different groups of refueling customers, thereby improving the recommendation effect of vehicle insurance products.

[0136] The recommendation module 604 is used to recommend vehicle insurance products to the refueling customer group based on a preset association rule mining algorithm, a preset tag recommendation algorithm based on tags of the refueling customer group, and a preset collaborative filtering algorithm.

[0137] The above-mentioned preset association rule mining algorithm is a preset association rule mining algorithm based on FP Tree (Frequent Pattern tree).

[0138] In the embodiment of the present invention, Fig.10 As shown, the recommendation module 604 specifically includes:

[0139] The first recommendation unit 6041 is used to extract the association relationship between the historical vehicle insurance products of the refueling customer group and the vehicle insurance products based on a preset association rule mining algorithm, and take the vehicle insurance product that meets the association relationship as the first candidate recommended vehicle insurance product and perform a recommendation score.

[0140] Specifically, through the association rule mining algorithm based on FP Tree, the frequent item sets between the historical vehicle insurance products and vehicle insurance products of the refueling customer group are obtained, that is, which vehicle insurance products have been purchased more than the specified threshold, and then through FP Tree, the conditional pattern base, conditional FP tree and frequent patterns corresponding to each vehicle insurance product can be obtained, so as to obtain the association relationship between the historical vehicle insurance products and vehicle insurance products of the refueling customer group. Thus, the corresponding association rules are obtained, that is, the possibility of purchasing several vehicle insurance products at the same time is very high, so the vehicle insurance products that meet the association rules can be used as candidate recommended vehicle insurance products. At the same time. These candidate recommended vehicle insurance products can be scored separately. The scoring can be a comprehensive scoring based on the purchase rate, usage rate, etc.

[0141] The second recommendation unit 6042 is used to select the vehicle insurance product corresponding to the tag of the refueling customer group as the second candidate recommended vehicle insurance product based on a preset tag recommendation algorithm and perform a recommendation score.

[0142] Among them, the preset tag recommendation algorithm is a pre-set tag-based recommendation algorithm. The preset tag recommendation algorithm specifically extracts key features from different dimensions based on some behavioral feature data of customers and labels them. In essence, it is actually a user portrait, but the features are more specific, detailed, and diverse. For example, based on refueling data, customer loyalty (the customer's loyalty is judged based on the distribution of the number of refueling times) and consumption trends (the total refueling amount shows an increasing, unstable, or decreasing trend) and other label dimensions can be extracted. Similar customers can be identified based on the similarity of different customer labels, so as to recommend vehicle insurance products.

[0143] Specifically, by using a preset tag recommendation algorithm, vehicle insurance products purchased by similar customers can be used as candidate recommended vehicle insurance products based on the customer's tag, and the obtained candidate recommended vehicle insurance products can be scored respectively. The scoring method can be to score the vehicle insurance products purchased by the customer according to the degree of similarity of the customer's tag. The vehicle insurance products purchased by the customer with the highest tag similarity have relatively high scores, and the vehicle insurance products purchased by the customer with low tag similarity have relatively low scores, etc.

[0144] The third recommendation unit 6043 is used to recommend the vehicle insurance product having the same historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics as the third candidate vehicle insurance product based on a preset collaborative filtering algorithm, and to perform a recommendation score.

[0145] The preset collaborative filtering algorithm is a pre-set collaborative filtering algorithm. The collaborative filtering algorithm is based on mining the customer's historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics to discover the customer's preference and predict the products that the customer may like for recommendation.

[0146] When recommendations are made based on preset collaborative filtering algorithms, they are mainly divided into two categories, namely customer-based collaborative filtering algorithms and item-based collaborative filtering algorithms. The former discovers customers' preferences for products through their historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics, measures and scores these preferences, calculates the relationship between customers based on their attitudes and preferences for the same product, and recommends products to customers with the same preferences. The latter obtains the relationship between products by calculating the scores of different customers for different products, and recommends similar items to customers based on the relationship between products. For example, if customer A buys product 1 and product 2, it means that the two products have a high degree of correlation. Therefore, when customer B buys product 1, product 2 can be recommended to him.

[0147] Specifically, the preset collaborative filtering algorithm can be used to find vehicle insurance products with the same preferences of the same customers as candidate recommended vehicle insurance products. The obtained candidate recommended vehicle insurance products are scored. The scoring method can be to score according to the scores of the customers on the vehicle insurance products.

[0148] The combined recommendation unit 6044 is used to obtain the final recommendation scores corresponding to the first candidate recommended vehicle insurance product, the second candidate recommended vehicle insurance product, and the third candidate recommended vehicle insurance product by weighted combination, and select the top n candidate recommended vehicle insurance products based on the final recommendation scores, and recommend them to the refueling customer group as a recommendation list.

[0149] Specifically, the three algorithms recommend candidate vehicle insurance products and their scores respectively, and then obtain the final scores through weighted combination, thereby obtaining the final vehicle insurance product recommendation list. The weights are equal at the beginning, and the weights are continuously adjusted through multiple combined recommendations and actual customer purchases to meet the needs of most customers.

[0150] The intersection of the products to be recommended obtained by the three algorithms is taken to obtain the corresponding candidate products and the scores given by the corresponding recommendation algorithms. Then all the scores are weighted and combined to obtain the final score of each product to be recommended. The top n scores are selected from them and recommended to the refueling customers as a recommendation list.

[0151] In the embodiment of the present invention, the results obtained by three recommendation algorithms, namely, a preset association rule mining algorithm, a preset tag recommendation algorithm, and a preset collaborative filtering algorithm, can be combined to achieve personalized recommendation of vehicle insurance products.

[0152] In the embodiment of the present invention, personalized recommendations for vehicle insurance products are made by combining multiple data sources including refueling data, claims data, and historical vehicle insurance product sales data of refueling customers, as well as multiple algorithm combinations including preset association rule mining algorithms, preset tag recommendation algorithms, and preset collaborative filtering algorithms. This ensures the diversity of recommendation data sources and recommendation algorithms, and is able to fully utilize the characteristics of the insurance product itself, the customer's refueling data, claims data, and other data to recommend products. The continuous updating of customer tags makes the recommendations more real-time and reliable, further improving the recommendation effect of vehicle insurance products.

[0153] To solve the above technical problems, the present application also provides a computer device. Fig.11 , Fig.11 This is a basic structural block diagram of the computer device in this embodiment.

[0154] The computer device 7 includes a memory 71, a processor 72, and a network interface 73 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 7 with components 71-73, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0155] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0156] The memory 71 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 71 can be an internal storage unit of the computer device 7, such as a hard disk or memory of the computer device 7. In other embodiments, the memory 71 can also be an external storage device of the computer device 7, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 7. Of course, the memory 71 can also include both the internal storage unit of the computer device 7 and its external storage device. In this embodiment, the memory 71 is generally used to store the operating system and various application software installed on the computer device 7, such as computer-readable instructions of the recommendation method of vehicle insurance products, etc. In addition, the memory 71 can also be used to temporarily store various types of data that have been output or are to be output.

[0157] The processor 72 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 72 is generally used to control the overall operation of the computer device 7. In this embodiment, the processor 72 is used to run computer-readable instructions or process data stored in the memory 71, such as computer-readable instructions for running the vehicle insurance product recommendation method.

[0158] The network interface 73 may include a wireless network interface or a wired network interface. The network interface 73 is generally used to establish a communication connection between the computer device 7 and other electronic devices.

[0159] The electronic device provided in the embodiment of the present application can implement each process of the vehicle insurance product recommendation method in the above method embodiment, and can achieve the same beneficial effect. To avoid repetition, it will not be described here.

[0160] The present application also provides another implementation, that is, a computer-readable storage medium is provided, the computer-readable storage medium stores computer-readable instructions, the computer-readable instructions can be executed by at least one processor, so that the at least one processor performs the steps of the vehicle insurance product recommendation method as described above. And the same technical effect can be achieved, and to avoid repetition, it will not be repeated here.

[0161] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0162] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for recommending a vehicle insurance product, characterized in that: The steps include: Obtaining refueling data, claims data, and historical vehicle insurance product sales data of the refueling customer group, wherein the claims data includes claims time characteristics and claims consumption characteristics, and the historical vehicle insurance product sales data includes historical vehicle insurance product time characteristics and historical vehicle insurance product consumption characteristics; Performing data fusion on the refueling data, and obtaining individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm; Comprehensively analyzing the claims data and the historical vehicle insurance product sales data with the individual behavior pattern characteristics, and labeling the refueling customer groups according to the analysis results, specifically including: comparatively analyzing the claims time characteristics, the historical vehicle insurance product time characteristics, and the refueling time characteristics to obtain time characteristic analysis results, comprehensively analyzing the claims consumption characteristics, the historical vehicle insurance product consumption characteristics, and the refueling consumption characteristics to obtain consumption characteristic analysis results, and labeling different refueling customer groups in combination with the time characteristic analysis results and the consumption characteristic analysis results; The vehicle insurance products are recommended to the refueling customer group based on a preset association rule mining algorithm, a preset label recommendation algorithm based on the label of the refueling customer group, and a preset collaborative filtering algorithm, specifically including: extracting the association relationship between the historical vehicle insurance products and vehicle insurance products of the refueling customer group based on the preset association rule mining algorithm, taking the vehicle insurance product that satisfies the association relationship as the first candidate recommended vehicle insurance product and performing a recommendation score, taking the vehicle insurance product corresponding to the label of the refueling customer group as the second candidate recommended vehicle insurance product and performing a recommendation score based on the preset label recommendation algorithm, taking the vehicle insurance product with the same historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics as the third candidate recommended vehicle insurance product and performing a recommendation score based on the preset collaborative filtering algorithm, obtaining the final recommendation scores corresponding to the first candidate recommended vehicle insurance product, the second candidate recommended vehicle insurance product, and the third candidate recommended vehicle insurance product by weighted combination, and selecting the top n candidate recommended vehicle insurance products based on the final recommendation score, and recommending them to the refueling customer group as a recommendation list.

2. The vehicle insurance product recommendation method according to claim 1, characterized in that: The refueling data includes static refueling data and dynamic refueling data, the individual behavior pattern characteristics include refueling consumption characteristics and refueling time characteristics, and the step of fusing the refueling data and obtaining the individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm includes: fusing the static refueling data and the dynamic refueling data to obtain fused data; Performing cluster analysis on the fused data based on the preset ant colony clustering algorithm to obtain a clustering result; The refueling consumption characteristics and refueling time characteristics of the refueling customer group are obtained according to the clustering result.

3. The vehicle insurance product recommendation method according to claim 1, characterized in that: The step of labeling different refueling customer groups by combining the time feature analysis result and the consumption feature analysis result specifically includes: Classify the refueling customer group by combining the time feature analysis result and the consumption feature analysis result to obtain a classification result; Different labels are given to different refueling customer groups according to the classification results.

4. A vehicle insurance product recommendation device, characterized in that: include: A data acquisition module, used to acquire refueling data, claim data, and historical vehicle insurance product sales data of a refueling customer group, wherein the claim data includes claim time characteristics and claim consumption characteristics, and the historical vehicle insurance product sales data includes historical vehicle insurance product time characteristics and historical vehicle insurance product consumption characteristics; A data fusion module, used to perform data fusion on the refueling data and obtain individual behavior pattern characteristics of the refueling customer group based on a preset ant colony clustering algorithm; An analysis module, for comprehensively analyzing the claims data and the historical vehicle insurance product sales data with the individual behavior pattern characteristics, and labeling the refueling customer group according to the analysis results; Wherein, the analysis module includes a first analysis unit, a second analysis unit and a label unit; The first analysis unit is used to compare and analyze the claim settlement time characteristics, the historical vehicle insurance product time characteristics and the refueling time characteristics to obtain a time characteristic analysis result; The second analysis unit is used to comprehensively analyze the claim consumption characteristics, historical vehicle insurance product consumption characteristics, and refueling consumption characteristics to obtain consumption characteristic analysis results; The labeling unit is used to label different refueling customer groups in combination with the time feature analysis result and the consumption feature analysis result; A recommendation module, for recommending vehicle insurance products to the refueling customer group based on a preset association rule mining algorithm, a preset tag recommendation algorithm based on tags of the refueling customer group, and a preset collaborative filtering algorithm; Wherein, the recommendation module includes a first recommendation unit, a second recommendation unit, a third recommendation unit and a combination recommendation unit; The first recommendation unit is used to extract the association relationship between the historical vehicle insurance products of the refueling customer group and the vehicle insurance products based on a preset association rule mining algorithm, select the vehicle insurance product that meets the association relationship as the first candidate recommended vehicle insurance product, and perform a recommendation score; The second recommendation unit is used to select the vehicle insurance product corresponding to the tag of the refueling customer group as the second candidate recommended vehicle insurance product based on a preset tag recommendation algorithm and perform a recommendation score; The third recommendation unit is used to select the vehicle insurance product having the same historical vehicle insurance product consumption characteristics and historical vehicle insurance product consumption time characteristics as the third candidate recommended vehicle insurance product based on a preset collaborative filtering algorithm, and perform a recommendation score; The combination recommendation unit is used to obtain the final recommendation scores corresponding to the first candidate recommended vehicle insurance product, the second candidate recommended vehicle insurance product, and the third candidate recommended vehicle insurance product by weighted combination, and select the top n candidate recommended vehicle insurance products based on the final recommendation scores, and recommend them to the refueling customer group as a recommendation list.

5. The vehicle insurance product recommendation device according to claim 4, characterized in that: The refueling data includes static refueling data and dynamic refueling data, the individual behavior pattern characteristics include refueling consumption characteristics and refueling time characteristics, and the data fusion module includes: A data fusion unit, used for fusing the static refueling data and the dynamic refueling data to obtain fused data; A clustering analysis unit, configured to perform clustering analysis on the fused data based on the preset ant colony clustering algorithm to obtain a clustering result; A feature acquisition unit is used to acquire the refueling consumption characteristics and refueling time characteristics of the refueling customer group according to the clustering result.

6. A computer device comprising a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the vehicle insurance product recommendation method according to any one of claims 1 to 3 when executing the computer-readable instructions.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the vehicle insurance product recommendation method according to any one of claims 1 to 3.

Images