Insurance service recommendation method and device, server and storage medium

Abstract

Embodiments of the present application relate to the field of financial technology, and disclose an insurance service recommendation method and device, a server and a storage medium, wherein the method comprises: obtaining basic information of a user, extracting feature data corresponding to the user according to the basic information, classifying the user to obtain a category label of the user, and determining a target service strategy suitable for the user according to the category label; sending an insurance service questionnaire to a terminal device according to the target service strategy and obtaining feedback of the questionnaire; determining an insurance service keyword and an insurance service theme according to the feedback of the questionnaire, generating a plurality of theme cover images according to the insurance service keyword by using an image generation model; generating target insurance service content according to the theme cover images and the insurance service theme, and sending the target insurance service content to the terminal device. The insurance service recommendation method realizes insurance service pushing based on the category of the user, improves the customer service quality of the insurance service, and further improves the transaction rate of the insurance service.

Description

Technical Field

[0001] This application relates to the field of financial technology, and in particular to a method, apparatus, server, and storage medium for recommending insurance services based on user categories. Background Technology

[0002] With the rapid development and widespread use of information technology, users' demand for insurance services is increasing, and at the same time, the number of financial institutions providing insurance services is also growing. Therefore, focusing on customer service quality is key for insurance service companies to attract and enhance customer loyalty. Given the diverse needs of different users within the vast insurance system, how to provide the most needed, high-quality services is a challenge faced by every insurance service company.

[0003] In existing insurance services, sales personnel typically recommend insurance services to different users using a uniform recommendation model, which prevents users from experiencing good customer service quality and results in a low insurance service conversion rate.

[0004] Therefore, improving the quality of customer service in insurance services, and thus increasing the insurance service conversion rate, is of great importance to those skilled in the art. Summary of the Invention

[0005] The main objective of this application is to provide a method, apparatus, server, and storage medium for recommending insurance services based on user categories, aiming to achieve personalized recommendations based on users, thereby improving the customer service quality of insurance services and increasing the insurance service conversion rate.

[0006] In a first aspect, embodiments of this application provide a method for recommending insurance services based on user categories, the method comprising:

[0007] When receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, the system responds to the insurance service request to obtain the user's basic information and extracts the user's corresponding feature data based on the basic information; the system classifies the user using a preset classification model and the feature data to obtain the user's category label, and determines the target service strategy suitable for the user based on the category label.

[0008] According to the target service strategy, an insurance service questionnaire is sent to the terminal device, and the terminal device's response to the insurance service questionnaire is obtained.

[0009] Based on the questionnaire feedback, insurance service keywords and themes were determined, and multiple theme cover images were generated using an image generation model based on the insurance service keywords.

[0010] The target insurance service content is generated based on the theme cover image and the insurance service theme, and then sent to the terminal device so that the terminal device can display the target insurance service content.

[0011] Secondly, embodiments of this application also provide an insurance service recommendation device based on user categories, comprising:

[0012] The information acquisition module is used to, when receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, respond to the insurance service request to obtain the user's basic information, and extract the user's corresponding feature data based on the basic information;

[0013] The strategy configuration module is used to classify the user using a preset classification model and the feature data, obtain the user's category label, and determine the target service strategy suitable for the user based on the category label;

[0014] The insurance questionnaire module is used to send an insurance service questionnaire to the terminal device according to the target service strategy, and to obtain the questionnaire feedback sent by the terminal device in response to the insurance service questionnaire.

[0015] The cover generation module is used to determine insurance service keywords and insurance service themes based on the questionnaire feedback, and to generate multiple themed cover images based on the insurance service keywords using an image generation model;

[0016] The service push module is used to generate target insurance service content based on the theme cover image and the insurance service theme, and send the target insurance service content to the terminal device so that the terminal device can display the target insurance service content.

[0017] Thirdly, embodiments of this application also provide a server, the server including a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for implementing communication between the processor and the memory, wherein when the computer program is executed by the processor, it implements the steps of the insurance service recommendation method based on user category as provided in any embodiment of this application specification.

[0018] Fourthly, embodiments of this application also provide a storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the user category-based insurance service recommendation method provided in any embodiment of this application specification.

[0019] This application provides a method, apparatus, server, and storage medium for recommending insurance services based on user categories. The method involves obtaining basic user information in response to an insurance service request and extracting corresponding feature data based on the basic information; classifying the user using a preset classification model and the feature data to obtain category tags; determining a target service strategy suitable for the user based on the category tags; sending an insurance service questionnaire to a terminal device according to the target service strategy and obtaining questionnaire feedback from the terminal device; determining insurance service keywords and themes based on the questionnaire feedback; generating multiple theme cover images based on the insurance service keywords using an image generation model; generating target insurance service content based on the theme cover images and the insurance service themes; and sending the target insurance service content to the terminal device so that the terminal device can display the target insurance service content.

[0020] By classifying users, different service strategies can be adopted for different user categories. This allows for more targeted insurance service questionnaires to provide personalized information collection services, accurately obtaining user feedback. Based on this feedback, targeted insurance service content can be provided to improve the quality of insurance service and ultimately increase the insurance transaction rate. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating the steps of an insurance service recommendation method based on user category, provided for embodiments of this application;

[0023] Figure 2 A schematic diagram illustrating an application scenario of an insurance service recommendation method based on user categories provided in this application embodiment;

[0024] Figure 3 This is a schematic diagram illustrating the scenario where the terminal device displays insurance service content on the application's program page after the server sends the insurance service content to the terminal device based on the user category insurance service recommendation method.

[0025] Figure 4 A schematic diagram of the module structure of an insurance service recommendation device based on user category provided in an embodiment of this application;

[0026] Figure 5 This is a schematic block diagram of a server structure provided in an embodiment of this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the described order. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0029] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0030] This application provides a method, apparatus, server, and storage medium for recommending insurance services based on user categories. The method for recommending content information is applied to a server, which can be a standalone server or a server cluster; no limitation is made here.

[0031] The method involves the following steps: Upon receiving an insurance service information request from a terminal device logged into by a user's account, the method responds to the request by obtaining the user's basic information and extracting corresponding feature data based on the basic information; classifying the user using a preset classification model and the feature data to obtain the user's category label, and determining a target service strategy suitable for the user based on the category label; sending an insurance service questionnaire to the terminal device according to the target service strategy, and obtaining the questionnaire feedback from the terminal device in response to the questionnaire; determining insurance service keywords and insurance service themes based on the questionnaire feedback, and generating multiple theme cover images based on the insurance service keywords using an image generation model; generating target insurance service content based on the theme cover images and the insurance service themes, and sending the target insurance service content to the terminal device so that the terminal device can display the target insurance service content.

[0032] By classifying users, different service strategies can be adopted for different user categories. This allows for more targeted insurance service questionnaires to provide personalized information collection services, accurately obtaining user feedback. Based on this feedback, targeted insurance service content can be provided to improve the quality of insurance service and ultimately increase the insurance transaction rate.

[0033] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0034] Please refer to Figure 1 , Figure 1 A flowchart illustrating the steps of an insurance service recommendation method based on user category, provided in an embodiment of this application.

[0035] like Figure 1 As shown, the method for recommending this content information includes steps S1 to S5.

[0036] Step S1: When receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, respond to the insurance service request, obtain the user's basic information, and extract the user's corresponding feature data based on the basic information;

[0037] For example, a user can register a unique user account on an insurance service platform using their identity information. The terminal device has the insurance service platform's application installed. The user logs into the insurance service platform using their user account and sends an insurance service information request to the server 20 by triggering a preset area on the insurance service platform. After receiving the insurance service information request, the server 20 responds to the insurance service information request and obtains the user's basic information.

[0038] Basic user information includes, but is not limited to, user identity data, user occupation information, user behavior data, and user interest tags. User identity data includes, but is not limited to, age, gender, and contact information. User behavior data includes, but is not limited to, data generated by user actions such as clicking, sharing, saving, purchasing, and actively searching. For example, data collected by an application after a user logs into a pre-defined application with an account linked to their identity information and performs actions such as clicking, sharing, saving, purchasing, and actively searching within that application. User interest tags can be generated based on the user's selection of preferred services on the platform or automatically categorized based on historical user data.

[0039] It is understandable that the collection and use of user information comply with relevant laws and regulations and have been authorized by the relevant parties.

[0040] like Figure 2 As shown, server 20 responds to insurance service requests to obtain basic user information, and by performing cluster analysis on the basic user information, it can obtain user preference data, user mining value, and other data. Preference data is used to characterize the degree of user interest in a certain type of product, information, or transaction, while user mining value is used to help characterize the user's importance level.

[0041] For example, insurance services typically include various types such as auto insurance, corporate property insurance, household property insurance, engineering insurance, liability insurance, credit insurance, guarantee insurance, marine and cargo insurance, and agricultural insurance. Each type of insurance service has one or more insurance items, and different types of users have different requirements for insurance services; that is, different users may purchase different insurance services or have different insurance items within those services. Therefore, upon receiving a user's insurance service request, the system obtains and processes the user's basic information to extract user characteristic data. This data is then used for cluster analysis to accurately classify users, facilitating the provision of targeted services later.

[0042] Step S2: Classify the user using a preset classification model and the feature data to obtain the user's category label, and determine the target service strategy suitable for the user based on the category label.

[0043] Optionally, the classification model includes, but is not limited to, at least one of the RFM model, AIPL model, and clustering algorithm model.

[0044] For example, by using a trained classification model and feature data to categorize users into different types, user category labels are obtained, which can then be used to identify user types. This allows for the development of different targeted service strategies for different types of users, thereby providing higher-quality customer service to the corresponding users.

[0045] Different user types have different needs for insurance services. Therefore, different service strategies need to be implemented according to different types of users in order to provide users with better services, thereby gaining user trust and increasing user loyalty to the company.

[0046] For example, by classifying users and assigning them corresponding category tags, service strategies can be matched to users based on these tags. Category tags can include A-type, B-type, C-type, and D-type user tags. Different user category tags indicate varying levels of interest in insurance services and different potential value for potential clients. For instance, an A-type user tag indicates a key retention customer with a strong demand for insurance services and who has already purchased them; a B-type user tag indicates a key development customer with strong growth potential; a C-type user tag indicates a key retention customer, suggesting efforts to retain them; and a D-type user tag indicates ambiguity and resistance towards insurance services.

[0047] By using big data analytics to categorize customers, more resources can be allocated to users or potential users who have a high intention to purchase insurance services and can create value for the company. This allows for targeted execution of corresponding service strategies for customers, ultimately providing them with high-quality services.

[0048] For example, for users tagged as Category A, we can first identify the insurance products they are interested in, then determine corresponding preferential policies for different insurance products, thereby providing better customer service. For Category B users, we can assess their current actual needs for insurance services, and once we confirm their need, we can plan corresponding insurance purchase projects for them, thus increasing user trust from their perspective. Category C users may have purchased insurance products from competitors; we can analyze the products they have already purchased to develop comparative information on the advantages and disadvantages of different insurance services, and provide them with more optimized policies to retain them as much as possible. For Category D users, we can improve user trust through real-world case analysis.

[0049] In some implementations, the classification model includes a vector transformation network, a first classification network, and a second classification network. The step of classifying the user using the preset classification model and the feature data to obtain the user's category label includes:

[0050] The feature data is converted into corresponding feature vectors using the vector transformation network, and the feature vectors are then input into the first classification network and the second classification network, respectively.

[0051] Obtain the first classification prediction value of the user output by the first classification network, and obtain the second classification prediction value of the user output by the second classification network;

[0052] The user's category label is determined based on the first category prediction value and the second category prediction value.

[0053] For example, the classification model uses two different classification networks to classify the data separately, and then takes a weighted average of the two classification results to make the classification results more reliable and accurate, which means that the category labels of the corresponding users output by the classification model are more accurate.

[0054] For example, by inputting feature data into the vector transformation network of a classification model, the feature data is converted into corresponding feature vectors. These feature vectors are then input into a first classification network and a second classification network, respectively, yielding the outputs of the first and second classification networks. For instance, the first classification network outputs a probability of 0.85 for a user being class A, 0.65 for class B, 0.32 for class C, and 0.72 for class D. The second classification network outputs a probability of 0.91 for a user being class A, 0.31 for class B, 0.3 for class C, and 0.8 for class D.

[0055] Therefore, according to the classification model, the probability of a user being a class A user is (0.85 + 0.91) / 2 = 0.88, the probability of a user being a class B user is (0.65 + 0.31) / 2 = 0.48, the probability of a user being a class C user is (0.32 + 0.30) / 2 = 0.31, and the probability of a user being a class D user is (0.72 + 0.8) / 2 = 0.76. The label type corresponding to the highest probability value is selected as the user's category label based on the probability ranking. For example, the classification model outputs the category label corresponding to a user being a class A user.

[0056] Step S3: Send an insurance service questionnaire to the terminal device according to the target service strategy, and obtain the questionnaire feedback sent by the terminal device in response to the insurance service questionnaire.

[0057] like Figure 2 As shown, for example, different insurance service questionnaires are provided for different types of users. The questionnaire content for each type of insurance service questionnaire is different, thereby providing different questionnaire questions through different questionnaire content. User questionnaire feedback is then collected through the questionnaire questions. By using a preset analysis model to analyze the questionnaire feedback, the specific needs information of users can be obtained, thereby accurately providing corresponding services to users.

[0058] For example, server 20 sets up a relationship between target service strategy and insurance service questionnaire. After determining the target service strategy based on the user's user type, it determines the insurance service questionnaire based on the target service strategy. Then, it can collect user questionnaire feedback in a targeted manner based on the insurance service questionnaire. By analyzing the questionnaire feedback, it can determine the user's insurance service needs and push insurance services in a targeted manner based on the user's insurance service needs.

[0059] In some implementations, sending an insurance service questionnaire to the terminal device according to the target service policy includes:

[0060] Based on the target service strategy, determine the insurance service questionnaire for information collection and download the questionnaire resources corresponding to the insurance service questionnaire from the resource server;

[0061] During the questionnaire resource download process, the questionnaire resource is transmitted to the terminal device so that the terminal device can display the corresponding insurance service questionnaire based on the questionnaire resource.

[0062] For example, different target service strategies correspond to different insurance service questionnaires, each with different content, including but not limited to multiple question-and-answer options. After determining the insurance service questionnaire, the corresponding questionnaire resource is downloaded from a preset resource server. During the download process, the questionnaire resource is transmitted to the terminal device 10, allowing the terminal device 10 to display the corresponding insurance service questionnaire based on the resource. Since the questionnaire resource is transmitted to the terminal device 10 during the download process from the resource server, there is no need to wait for the resource download to complete before transmission, thus effectively saving time.

[0063] Step S4: Determine insurance service keywords and insurance service themes based on the questionnaire feedback, and use an image generation model to generate multiple theme cover images based on the insurance service keywords;

[0064] For example, in the process of pushing insurance services, if only text information is used, the visual experience for users is poor, increasing the likelihood that users will ignore the insurance service push. Therefore, after obtaining users' insurance service needs based on questionnaire feedback, this application aims to enhance users' interest in viewing the information, improve their information viewing experience, and help increase the conversion rate of insurance services. The application determines insurance service keywords and themes through questionnaire feedback, inputs these keywords into a corresponding image generation model, and uses the image generation model to output a cover image matching the insurance service to be pushed. By combining images with text, users can quickly obtain necessary information visually upon seeing the service information, thereby increasing their interest in viewing the information, providing them with high-quality service, and ultimately improving the conversion rate of insurance services.

[0065] In some implementations, determining insurance service keywords and insurance service themes based on the questionnaire feedback includes:

[0066] Based on the questionnaire feedback, determine the insurance service items that the user is interested in;

[0067] Based on the insurance service items, determine the appropriate insurance service keywords and themes for the user.

[0068] For example, the questionnaire feedback includes multiple questions. By comprehensively analyzing the information collected from each question, the insurance service items that the user is interested in can be identified. Then, the corresponding insurance service keywords and themes can be determined based on these insurance service items.

[0069] For example, when it is confirmed that a user of type B is interested in sub-category X insurance in car insurance, server 20 automatically matches and retrieves insurance service keywords and insurance service topics that are suitable for the insurance service items that the user is interested in from the database. For example, when it is confirmed that a user of type B is interested in sub-category X insurance in car insurance, the obtained insurance service keywords are "the specific handling process of sub-category X insurance, the claims process of sub-category X insurance, and the precautions for handling sub-category X insurance", and the obtained insurance service topic is "sub-category X insurance that you have overlooked".

[0070] In some implementations, the image generation model includes a text parsing network, an image generation network, and an image enhancement network. The step of generating multiple themed cover images based on the insurance service keywords using the image generation model includes:

[0071] The text parsing network is used to perform semantic parsing on the insurance service keywords to obtain semantic parsing feature vectors;

[0072] The image generation network outputs multiple initial cover images based on the semantic parsing feature vector, and the image enhancement network performs image enhancement processing on the multiple initial cover images to obtain multiple themed cover images.

[0073] For example, a text parsing network performs semantic parsing on insurance service keywords to obtain corresponding semantic parsing feature vectors. An image generation network then outputs multiple initial cover images matching the insurance service keywords based on these feature vectors. An image enhancement network performs image enhancement processing on these initial cover images, such as pixel enhancement and contrast enhancement, to obtain multiple themed cover images. Based on these multiple themed cover images and the resulting image enhancement processing, the display effect is improved, thus providing a better visual experience for users when viewing the themed cover images.

[0074] Step S5: Generate target insurance service content based on the theme cover image and the insurance service theme, and send the target insurance service content to the terminal device so that the terminal device can display the target insurance service content.

[0075] For example, after determining the theme cover image and the insurance service theme, the target insurance service content is generated based on the theme cover image and the insurance service theme, and the target insurance service content is sent to the terminal device, so that the terminal device displays the target insurance service content on the preset program page of the target application.

[0076] For example, if the target application is an insurance application platform, users can quickly obtain key information about the insurance services they are interested in by viewing the insurance service themes and cover images displayed on the corresponding program pages of the insurance application platform, thereby increasing users' interest in viewing insurance service information.

[0077] like Figure 3 As shown, the terminal device 10 has an application program of the insurance application platform installed. After receiving the target insurance service content sent by the server 20, the terminal device 10 displays the target insurance service content in the first display area 102 of the preset page 101 of the application.

[0078] Based on this, the insurance service content includes the insurance service theme and the theme cover image. Users can extract some key information from the text part of the insurance service theme and another part of key information from the image part of the theme cover image. Furthermore, the combination of text and images can effectively improve the user's visual experience and enhance the user's interest in reading the insurance service information.

[0079] In some implementations, the method further includes, before generating multiple themed cover images based on the insurance service keywords using an image generation model:

[0080] The initial image to be trained is used to generate a model that outputs the first image set;

[0081] Obtain the image similarity of each first image in the first image set;

[0082] The loss value of the loss function of the initial image generation model is calculated based on the image similarity.

[0083] The model parameters of the initial image generation model are updated based on the loss value until the image generation model is obtained.

[0084] For example, the initial image generation model is a text-image model, that is, generating an image that matches the text by taking text as input.

[0085] Since different users may have different interests in insurance services, and different content requires different recommended cover images, higher demands are placed on the diversity of images generated by image generation models.

[0086] Image generation models have seen significant development in recent years, but they tend to generate highly homogeneous images. This problem is generally caused by overfitting due to excessive training epochs, or by insufficient training samples and high complexity. Image generation models do not discriminate against the diversity of generated images during training, making it difficult to address the issue of image homogeneity. This deficiency directly results in the difficulty of obtaining satisfactory recommended cover images using existing image generation models.

[0087] To improve the diversity of images output by the image generation model, this embodiment utilizes an initial image generation model to output a first image set. Then, it extracts the image similarity of each first image in the first image set from multiple dimensions. The image similarity includes at least grayscale similarity, contour similarity of target objects in the images, and spatial position similarity of target objects in the images. After obtaining the image similarity of each first image in the first image set, the loss value of the loss function of the initial image generation model is calculated using the image similarity. The model parameters of the initial image generation model are then adjusted and updated based on the loss value until a target image generation model is obtained. This results in lower similarity between images in the image set output by the target image generation model, thus achieving diversity in the output images.

[0088] It can be understood that the initial image generation model is completed when the training of the initial image generation model is completed, which can be indicated by the loss value of the loss function of the initial image generation model converging to a preset value, or by the initial image generation model reaching a preset number of iterations.

[0089] As can be understood, in this embodiment, the loss function is the total similarity of the first image set. That is, the total similarity of the first image set is used as the loss function of the initial image generation model, and the loss value of the loss function is positively correlated with the total similarity of the first image set; the higher the total similarity, the higher the loss value; the lower the total similarity, the lower the loss value. In other words, the lower the similarity between images, the better the diversity, the more the requirements are met, and the lower the loss value.

[0090] In some implementations, the image similarity includes at least grayscale similarity, contour similarity, and spatial location similarity, and the first image set includes multiple sub-images. Obtaining the image similarity of each first image in the first image set includes:

[0091] A grayscale comparison operation is performed on the image control group in the first image set to compare the grayscale of the sub-images within the image control group and obtain the grayscale similarity of the image control group, wherein the image control group is composed of any two sub-images in the first image set;

[0092] A contour extraction operation is performed on the image control group to compare the contours of the sub-images within the image control group and obtain the contour similarity of the image control group.

[0093] A spatial position comparison operation is performed on the image control group to compare the spatial positions of target objects in the sub-images within the image control group, thereby obtaining the spatial position similarity of the image control group.

[0094] The grayscale similarity, contour similarity, and spatial location similarity are obtained multiple times, and the image similarity corresponding to the first image set is determined based on the multiple obtained grayscale similarity, contour similarity, and spatial location similarity.

[0095] For example, any two sub-images are obtained from the first image set to form an image control group, and then the image similarity between the two sub-images in the image control group is calculated. The image similarity includes at least grayscale similarity, contour similarity, and spatial position similarity.

[0096] A grayscale comparison operation is performed on the image control group in the first image set to obtain a grayscale difference that can clearly represent the grayscale difference between the two sub-images in the image control group, thereby obtaining the grayscale detail difference between the two sub-images in the image control group, and finally determining the similarity of the grayscale details between the two sub-images in the image control group.

[0097] Contour extraction is performed on the image control group in the first image set to obtain the contour information of the target object in the sub-image of the image control group. Based on the contour information of the target object in the sub-image, the difference in the contour size of the target object in different images is known, and then the contour similarity between the two images in the image control group is obtained.

[0098] A spatial position comparison operation is performed on the image control group in the first image set to compare the spatial positions of target objects in the sub-images within the image control group, thereby obtaining the differences in the spatial positions of target objects in the sub-images within the image control group, and thus obtaining the spatial position similarity of the image control group.

[0099] After obtaining the grayscale similarity, contour similarity, and spatial location similarity of the image control group, the average value of the grayscale similarity, contour similarity, and spatial location similarity is calculated to obtain the image similarity of the image control group.

[0100] After performing grayscale comparison, contour extraction, and spatial position comparison operations on the image control group composed of any two images in the first image set, the image similarity corresponding to the image control group is obtained, and then the image similarity corresponding to each first image in the first image set is obtained.

[0101] In some implementations, performing a grayscale comparison operation on the image control group of the first image set to compare the grayscale of sub-images within the image control group and obtain the grayscale similarity of the image control group includes:

[0102] The two sub-images in the image control group are transformed by grayscale to obtain the first grayscale image and the second grayscale image;

[0103] Obtain the third pixel vector of the first grayscale image and the fourth pixel vector of the second grayscale image;

[0104] The grayscale similarity between the two sub-images in the image control group is obtained by calculating the cosine similarity based on the third pixel vector and the fourth pixel vector.

[0105] For example, to eliminate the interference of image color on image detail extraction, the sub-images in the image control group are first subjected to grayscale transformation to obtain the first grayscale image and the second grayscale image corresponding to the two sub-images. The grayscale transformation can use binarization technology to convert the color image into a grayscale image, and the image value contains only two data types. The threshold for binarization transformation can be set by the user.

[0106] When a grayscale image contains only 0 and 1, the location information corresponding to 0 represents the background, and the location information corresponding to 1 represents the target object, such as a vehicle. Therefore, after obtaining the third pixel vector of the first grayscale image and the fourth pixel vector of the second grayscale image, the larger the calculation result of the cosine similarity, the greater the grayscale similarity.

[0107] In some implementations, performing contour extraction on the image control group to compare the contours of sub-images within the image control group and obtain the contour similarity of the image control group includes:

[0108] The first grayscale image is blurred by adding noise data, and the second grayscale image is blurred by adding noise data, to obtain the second noise image.

[0109] Contour extraction is performed on the first noisy image using binarization technology to obtain a first contour vector, and contour extraction is performed on the second noisy image using binarization to obtain a second contour vector;

[0110] Based on the first contour vector and the second contour vector, cosine similarity is calculated to obtain the contour similarity between two sub-images in the image control group.

[0111] For example, to reduce the interference of color on image contour similarity, the images in the image control group are processed into grayscale to obtain a first grayscale image and a second grayscale image. Then, noise data is added to both the first and second grayscale images for blurring, where the noise data can be white noise or random noise, resulting in a first noise image corresponding to the first grayscale image and a second noise image corresponding to the second grayscale image. Based on this, the first and second noise images are binarized to obtain the image contour information. Cosine similarity is calculated based on the first and second contour vectors to obtain the contour similarity between the two sub-images in the image control group.

[0112] Please see Figure 4 , Figure 4 This is a block diagram illustrating the structure of an insurance service recommendation device based on user categories, as provided in an embodiment of this application.

[0113] like Figure 4 As shown, an insurance service recommendation device 200 based on user categories is applied to a server 20 and used to execute the aforementioned insurance service recommendation method based on user categories. The insurance service recommendation device 200 includes an information acquisition module 201, a strategy configuration module 202, an insurance questionnaire module 203, a cover generation module 204, and a service push module 205.

[0114] The information acquisition module 201 is used to, when receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, respond to the insurance service request to obtain the user's basic information, and extract the user's corresponding feature data based on the basic information;

[0115] The strategy configuration module 202 is used to classify the user using a preset classification model and the feature data, obtain the user's category label, and determine the target service strategy suitable for the user based on the category label;

[0116] The insurance questionnaire module 203 is used to send an insurance service questionnaire to the terminal device according to the target service strategy, and to obtain the questionnaire feedback sent by the terminal device in response to the insurance service questionnaire.

[0117] The cover generation module 204 is used to determine insurance service keywords and insurance service themes based on the questionnaire feedback, and to generate multiple themed cover images based on the insurance service keywords using an image generation model.

[0118] The service push module 205 is used to generate target insurance service content based on the theme cover image and the insurance service theme, and send the target insurance service content to the terminal device so that the terminal device can display the target insurance service content.

[0119] In some implementations, the classification model includes a vector transformation network, a first classification network, and a second classification network. The step of classifying the user using the preset classification model and the feature data to obtain the user's category label includes:

[0120] The feature data is converted into corresponding feature vectors using the vector transformation network, and the feature vectors are then input into the first classification network and the second classification network, respectively.

[0121] Obtain the first classification prediction value of the user output by the first classification network, and obtain the second classification prediction value of the user output by the second classification network;

[0122] The user's category label is determined based on the first category prediction value and the second category prediction value.

[0123] In some implementations, sending an insurance service questionnaire to the terminal device according to the target service policy includes:

[0124] Based on the target service strategy, determine the insurance service questionnaire for information collection and download the questionnaire resources corresponding to the insurance service questionnaire from the resource server;

[0125] During the questionnaire resource download process, the questionnaire resource is transmitted to the terminal device so that the terminal device can display the corresponding insurance service questionnaire based on the questionnaire resource.

[0126] In some implementations, determining insurance service keywords and insurance service themes based on the questionnaire feedback includes:

[0127] Based on the questionnaire feedback, determine the insurance service items that the user is interested in;

[0128] Based on the insurance service items, determine the appropriate insurance service keywords and themes for the user.

[0129] In some implementations, the image generation model includes a text parsing network, an image generation network, and an image enhancement network. The step of generating multiple themed cover images based on the insurance service keywords using the image generation model includes:

[0130] The text parsing network is used to perform semantic parsing on the insurance service keywords to obtain semantic parsing feature vectors;

[0131] The image generation network outputs multiple initial cover images based on the semantic parsing feature vector, and the image enhancement network performs image enhancement processing on the multiple initial cover images to obtain multiple themed cover images.

[0132] In some implementations, the insurance service recommendation device 200 based on user categories further includes a model training module. This module is used to output a first image set using an initial image generation model to be trained before generating multiple themed cover images based on the insurance service keywords using an image generation model. It also obtains the image similarity of each first image in the first image set, calculates the loss value of the loss function of the initial image generation model based on the image similarity, and updates the model parameters of the initial image generation model based on the loss value until an image generation model is obtained.

[0133] In some implementations, the image similarity includes at least grayscale similarity, contour similarity, and spatial location similarity, and the first image set includes multiple sub-images. Obtaining the image similarity of each first image in the first image set includes:

[0134] A grayscale comparison operation is performed on the image control group in the first image set to compare the grayscale of the sub-images within the image control group and obtain the grayscale similarity of the image control group, wherein the image control group is composed of any two sub-images in the first image set;

[0135] A contour extraction operation is performed on the image control group to compare the contours of the sub-images within the image control group and obtain the contour similarity of the image control group.

[0136] A spatial position comparison operation is performed on the image control group to compare the spatial positions of target objects in the sub-images within the image control group, thereby obtaining the spatial position similarity of the image control group.

[0137] The grayscale similarity, contour similarity, and spatial location similarity are obtained multiple times, and the image similarity corresponding to the first image set is determined based on the multiple obtained grayscale similarity, contour similarity, and spatial location similarity.

[0138] It should be noted that those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the aforementioned embodiment of the insurance service recommendation method based on user category, and will not be repeated here.

[0139] Please see Figure 5 , Figure 5 This is a schematic block diagram of the server structure provided in an embodiment of this application.

[0140] like Figure 5 As shown, server 20 includes processor 21 and memory 22, which are connected via bus 23, such as I2C (Inter-integrated Circuit) bus.

[0141] Specifically, processor 21 provides computing and control capabilities to support the operation of the entire server. Processor 21 can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0142] Specifically, the memory 22 can be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a portable hard drive, etc.

[0143] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the embodiments of this application, and does not constitute a limitation on the server to which the embodiments of this application are applied. A specific server may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0144] The processor 21 is used to run a computer program stored in the memory, and implements the user category-based insurance service recommendation method provided in any embodiment of this application when executing the computer program.

[0145] In some implementations, processor 21 is used to implement the following method steps:

[0146] When receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, the system responds to the insurance service request to obtain the user's basic information and extracts the user's corresponding feature data based on the basic information.

[0147] The user is classified using a preset classification model and the feature data to obtain the user's category label, and the target service strategy adapted to the user is determined based on the category label;

[0148] According to the target service strategy, an insurance service questionnaire is sent to the terminal device, and the terminal device's response to the insurance service questionnaire is obtained.

[0149] Based on the questionnaire feedback, insurance service keywords and themes were determined, and multiple theme cover images were generated using an image generation model based on the insurance service keywords.

[0150] The target insurance service content is generated based on the theme cover image and the insurance service theme, and then sent to the terminal device so that the terminal device can display the target insurance service content.

[0151] In some implementations, the classification model includes a vector transformation network, a first classification network, and a second classification network. The step of classifying the user using the preset classification model and the feature data to obtain the user's category label includes:

[0152] The feature data is converted into corresponding feature vectors using the vector transformation network, and the feature vectors are then input into the first classification network and the second classification network, respectively.

[0153] Obtain the first classification prediction value of the user output by the first classification network, and obtain the second classification prediction value of the user output by the second classification network;

[0154] The user's category label is determined based on the first category prediction value and the second category prediction value.

[0155] In some implementations, sending an insurance service questionnaire to the terminal device according to the target service policy includes:

[0156] Based on the target service strategy, determine the insurance service questionnaire for information collection and download the questionnaire resources corresponding to the insurance service questionnaire from the resource server;

[0157] During the questionnaire resource download process, the questionnaire resource is transmitted to the terminal device so that the terminal device can display the corresponding insurance service questionnaire based on the questionnaire resource.

[0158] In some implementations, determining insurance service keywords and insurance service themes based on the questionnaire feedback includes:

[0159] Based on the questionnaire feedback, determine the insurance service items that the user is interested in;

[0160] Based on the insurance service items, determine the appropriate insurance service keywords and themes for the user.

[0161] In some implementations, the image generation model includes a text parsing network, an image generation network, and an image enhancement network. The step of generating multiple themed cover images based on the insurance service keywords using the image generation model includes:

[0162] The text parsing network is used to perform semantic parsing on the insurance service keywords to obtain semantic parsing feature vectors;

[0163] The image generation network outputs multiple initial cover images based on the semantic parsing feature vector, and the image enhancement network performs image enhancement processing on the multiple initial cover images to obtain multiple themed cover images.

[0164] In some embodiments, the processor 21 is further configured to perform the following method steps before generating multiple themed cover images based on the insurance service keywords using an image generation model:

[0165] The initial image to be trained is used to generate a model that outputs the first image set;

[0166] Obtain the image similarity of each first image in the first image set;

[0167] The loss value of the loss function of the initial image generation model is calculated based on the image similarity.

[0168] The model parameters of the initial image generation model are updated based on the loss value until the image generation model is obtained.

[0169] In some implementations, the image similarity includes at least grayscale similarity, contour similarity, and spatial location similarity, and the first image set includes multiple sub-images. Obtaining the image similarity of each first image in the first image set includes:

[0170] A grayscale comparison operation is performed on the image control group in the first image set to compare the grayscale of the sub-images within the image control group and obtain the grayscale similarity of the image control group, wherein the image control group is composed of any two sub-images in the first image set;

[0171] A contour extraction operation is performed on the image control group to compare the contours of the sub-images within the image control group and obtain the contour similarity of the image control group.

[0172] A spatial position comparison operation is performed on the image control group to compare the spatial positions of target objects in the sub-images within the image control group, thereby obtaining the spatial position similarity of the image control group.

[0173] The grayscale similarity, contour similarity, and spatial location similarity are obtained multiple times, and the image similarity corresponding to the first image set is determined based on the multiple obtained grayscale similarity, contour similarity, and spatial location similarity.

[0174] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the server described above can be referred to the corresponding process in the aforementioned embodiment of the insurance service recommendation method based on user categories, and will not be repeated here.

[0175] This application also provides a storage medium for computer-readable storage, which stores one or more programs that can be executed by one or more processors to implement the steps of any of the user-category-based insurance service recommendation methods provided in the embodiments of this application.

[0176] The storage medium can be the internal storage unit of the server in the aforementioned embodiments, such as server memory. Alternatively, the storage medium can be an external storage device for the server, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card.

[0177] It will be understood by those skilled in the art that all or some of the steps in the methods disclosed above, and the functional modules / units in the apparatus, can be implemented as software, firmware, hardware, and suitable combinations thereof. In hardware embodiments, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0178] It should be understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. It should be noted that, herein, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0179] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for recommending insurance services based on user categories, characterized in that, The method includes: When receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, the system responds to the insurance service request to obtain the user's basic information and extracts the user's corresponding feature data based on the basic information. The user is classified using a preset classification model and the feature data to obtain the user's category label, and the target service strategy adapted to the user is determined based on the category label; According to the target service strategy, an insurance service questionnaire is sent to the terminal device, and the terminal device's response to the insurance service questionnaire is obtained. Based on the questionnaire feedback, insurance service keywords and themes were determined, and multiple theme cover images were generated using an image generation model based on the insurance service keywords. Based on the theme cover image and the insurance service theme, target insurance service content is generated and sent to the terminal device so that the terminal device can display the target insurance service content. The image generation model includes a text parsing network, an image generation network, and an image enhancement network. The step of generating multiple themed cover images based on the insurance service keywords using the image generation model includes: performing semantic parsing on the insurance service keywords using the text parsing network to obtain a semantic parsing feature vector; outputting multiple initial cover images based on the semantic parsing feature vector using the image generation network; and performing image enhancement processing on the multiple initial cover images using the image enhancement network to obtain multiple themed cover images.

2. The method according to claim 1, characterized in that, The classification model includes a vector transformation network, a first classification network, and a second classification network. The process of classifying the user using the preset classification model and the feature data to obtain the user's category label includes: The feature data is converted into corresponding feature vectors using the vector transformation network, and the feature vectors are then input into the first classification network and the second classification network, respectively. Obtain the first classification prediction value of the user output by the first classification network, and obtain the second classification prediction value of the user output by the second classification network; The user's category label is determined based on the first category prediction value and the second category prediction value.

3. The method according to claim 1, characterized in that, Sending the insurance service questionnaire to the terminal device according to the target service policy includes: Based on the target service strategy, determine the insurance service questionnaire for information collection and download the questionnaire resources corresponding to the insurance service questionnaire from the resource server; During the questionnaire resource download process, the questionnaire resource is transmitted to the terminal device so that the terminal device can display the corresponding insurance service questionnaire based on the questionnaire resource.

4. The method according to claim 1, characterized in that, The process of determining insurance service keywords and themes based on the questionnaire feedback includes: Based on the questionnaire feedback, determine the insurance service items that the user is interested in; Based on the insurance service items, determine the appropriate insurance service keywords and themes for the user.

5. The method according to claim 1, characterized in that, Before generating multiple themed cover images based on the insurance service keywords using an image generation model, the method further includes: The initial image to be trained is used to generate a model that outputs the first image set; Obtain the image similarity of each first image in the first image set; The loss value of the loss function of the initial image generation model is calculated based on the image similarity. The model parameters of the initial image generation model are updated based on the loss value until the image generation model is obtained.

6. The method according to claim 5, characterized in that, The image similarity includes at least grayscale similarity, contour similarity, and spatial location similarity, and the first image set includes multiple sub-images. Obtaining the image similarity of each first image in the first image set includes: A grayscale comparison operation is performed on the image control group in the first image set to compare the grayscale of the sub-images within the image control group and obtain the grayscale similarity of the image control group, wherein the image control group is composed of any two sub-images in the first image set; A contour extraction operation is performed on the image control group to compare the contours of the sub-images within the image control group and obtain the contour similarity of the image control group. A spatial position comparison operation is performed on the image control group to compare the spatial positions of target objects in the sub-images within the image control group, thereby obtaining the spatial position similarity of the image control group. The grayscale similarity, contour similarity, and spatial location similarity are obtained multiple times, and the image similarity corresponding to the first image set is determined based on the multiple obtained grayscale similarity, contour similarity, and spatial location similarity.

7. An insurance service recommendation device based on user categories, characterized in that, include: The information acquisition module is used to, when receiving an insurance service information request sent by the terminal device logged in by the user's corresponding user account, respond to the insurance service request to obtain the user's basic information, and extract the user's corresponding feature data based on the basic information; The strategy configuration module is used to classify the user using a preset classification model and the feature data, obtain the user's category label, and determine the target service strategy suitable for the user based on the category label; The insurance questionnaire module is used to send an insurance service questionnaire to the terminal device according to the target service strategy, and to obtain the questionnaire feedback sent by the terminal device in response to the insurance service questionnaire. The cover generation module is used to determine insurance service keywords and insurance service themes based on the questionnaire feedback, and to generate multiple themed cover images based on the insurance service keywords using an image generation model; The service push module is used to generate target insurance service content based on the theme cover image and the insurance service theme, and send the target insurance service content to the terminal device so that the terminal device can display the target insurance service content; The image generation model includes a text parsing network, an image generation network, and an image enhancement network. The step of generating multiple themed cover images based on the insurance service keywords using the image generation model includes: performing semantic parsing on the insurance service keywords using the text parsing network to obtain a semantic parsing feature vector; outputting multiple initial cover images based on the semantic parsing feature vector using the image generation network; and performing image enhancement processing on the multiple initial cover images using the image enhancement network to obtain multiple themed cover images.

8. A server, characterized in that, The server includes a processor and a memory; The memory is used to store computer programs; The processor is configured to execute the computer program and, in executing the computer program, implement the insurance service recommendation method based on user category as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, When the computer-readable storage medium is executed by one or more processors, the one or more processors perform the steps of the insurance service recommendation method based on user category as described in any one of claims 1 to 6.

Images