An insurance product matching method, device and electronic equipment

Abstract

This disclosure relates to an insurance product matching method, apparatus, and electronic device, comprising: acquiring occupational description information corresponding to a target object and a preset node relationship graph corresponding to multiple preset insurance products; extracting features from the occupational description information to obtain occupational feature information corresponding to the target object; performing matching processing on each edge in the preset node relationship graph based on the occupational feature information to obtain at least one target matching edge corresponding to the target object; and determining at least one target insurance product corresponding to the target object based on the target product node corresponding to each of the at least one target matching edge. Using embodiments of this disclosure, accurate matching of insurance products based on the target object's occupation can be achieved, avoiding the problem of inaccurate matching caused by different occupational semantics when occupational information is the same but their industry information is different, thereby improving the user experience.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to an insurance product matching method, apparatus and electronic device. Background Technology

[0002] In existing insurance businesses, insurance products typically impose strict restrictions on eligible occupations. In actual business scenarios, users often describe their occupations to operations personnel via instant messaging tools, and this occupational description is usually in natural language text.

[0003] Currently, operations staff need to rely on their experience to understand and judge the occupations described by users, manually determine the corresponding standard occupational names and categories, and search for eligible occupational configurations among multiple insurance products to complete the insurance product matching. However, this manual processing method relies heavily on the professional knowledge of operations staff, and judgments based on experience are prone to bias, resulting in insufficient matching accuracy. Furthermore, the manual matching process is time-consuming and inefficient, making it difficult to meet the needs of high-concurrency or large-scale business. Therefore, the above method suffers from low matching accuracy and low matching efficiency. Summary of the Invention

[0004] In view of the aforementioned technical problems, this disclosure proposes an insurance product matching method, apparatus, and electronic device.

[0005] According to one aspect of the embodiments of this disclosure, an insurance product matching method is provided, the method comprising: Obtain the target object's occupational description information and the preset node relationship graph corresponding to multiple preset insurance products; the preset node relationship graph is constructed with the multiple preset insurance products and multiple preset occupational information as nodes, and the association relationship between any preset insurance product and any preset occupational information as edges; each edge in the preset node relationship graph includes preset occupational feature information, and the preset occupational feature information of each edge represents the preset industry information to which the preset occupational information corresponding to the product node connected to each edge belongs in the preset occupational information of the occupational node connected to each edge and the industry information corresponding to the occupational node; Feature extraction is performed on the object's occupational description information to obtain the object's occupational feature information corresponding to the target object; the object's occupational feature information represents the target object's occupational information and the industry information in which the object's occupational information is located. Based on the occupational characteristic information of the object, each edge in the preset node relationship graph is matched to obtain at least one target matching edge corresponding to the target object; Based on the target product node corresponding to each of the at least one target matching edge, at least one target insurance product corresponding to the target object is determined.

[0006] According to another aspect of the present disclosure, an insurance product matching device is provided, the device comprising: The first information acquisition module is used to acquire the object occupation description information corresponding to the target object and the preset node relationship diagram corresponding to multiple preset insurance products; the preset node relationship diagram is constructed with the multiple preset insurance products and multiple preset occupation information as nodes, and the association relationship between any preset insurance product and any preset occupation information as edges; each edge in the preset node relationship diagram includes preset occupation feature information, and the preset occupation feature information of each edge represents the preset industry information to which the preset occupation information corresponding to the product node connected to each edge belongs in the preset occupation information of the occupation node connected to each edge and the industry information corresponding to the occupation node; The first feature extraction module is used to extract features from the object's occupational description information to obtain the object's occupational feature information corresponding to the target object; the object's occupational feature information represents the object's occupational information and the industry information in which the object's occupational information is located. The matching processing module is used to perform matching processing on each edge in the preset node relationship graph based on the occupational feature information of the object, so as to obtain at least one target matching edge corresponding to the target object; The target product determination module is used to determine at least one target insurance product corresponding to the target object based on the target product node corresponding to each of the at least one target matching edge.

[0007] Optionally, the preset occupational feature information of each edge includes a preset occupational feature vector and preset occupational feature words, and the object occupational feature information includes an object occupational feature vector and object occupational feature words; the matching processing module includes: A vector matching unit is used to perform vector matching processing on the object's occupational feature vector and the preset occupational feature vector of each edge to obtain a vector matching result; the vector matching result is used to indicate the degree of matching between the preset occupational feature vector of each edge and the object's occupational feature vector; The word segmentation matching unit is used to perform word segmentation matching processing on the word segmentation of the object's occupational features and the preset word segmentation of each edge to obtain the word segmentation matching result; the word segmentation matching result is used to indicate the degree of matching between the preset word segmentation of each edge and the word segmentation of the object's occupational features; The matching edge determination unit is used to select at least one target matching edge from multiple edges of the preset node relationship graph based on the vector matching result and the word segmentation matching result.

[0008] Optionally, the vector matching result includes vector matching data corresponding to each edge, and the word segmentation matching result includes word segmentation matching data corresponding to each edge; the device further includes: The second information acquisition module is used to acquire the preset vector matching weights corresponding to the vector matching dimension and the preset word segmentation matching weights corresponding to the word segmentation matching dimension. The matching result filtering unit includes: The first sorting unit is used to sort the multiple edges according to the vector matching results in descending order of the corresponding vector matching data to obtain a sequence of vector candidate edges; The second sorting unit is used to sort the multiple edges according to the word segmentation matching results in descending order of the corresponding word segmentation matching data to obtain a word segmentation candidate edge sequence. The weighted filtering unit is used to determine at least one target matching edge from the vector candidate edge sequence and the word candidate edge sequence based on the preset vector matching weight and the preset word segmentation matching weight.

[0009] Optionally, the weighted filtering unit includes: The matching dimension determination unit is used to determine a first matching dimension and a second matching dimension from the vector matching dimension and the word segmentation matching dimension based on the preset vector matching weight and the preset word segmentation matching weight; the first matching dimension is the matching dimension corresponding to the largest matching weight among the preset vector matching weight and the preset word segmentation matching weight; the second matching dimension is the matching dimension other than the first matching dimension among the vector matching dimension and the word segmentation matching dimension. The edge number determination unit is used to determine the number of first matching edges corresponding to the first matching dimension and the number of second matching edges corresponding to the second matching dimension based on the preset vector matching weight and the preset word segmentation matching weight. A sequence determination unit is used to determine, from the vector candidate edge sequence and the word segmentation candidate edge sequence, a first candidate edge sequence corresponding to the first matching dimension and a second candidate edge sequence corresponding to the second matching dimension; The first edge determination unit is used to take the number of first matching edges that are the first edge in the first candidate edge sequence as the first matching edge corresponding to the first matching dimension. The second side determination unit is used to take the number of the first second matching edges in the second candidate edge sequence as the second matching edges corresponding to the second matching dimension. The third side determination unit is used to determine the first matching edge and the second matching edge as the at least one target matching edge.

[0010] Optionally, the device further includes: The edge deduplication module is used to perform deduplication processing on the second candidate edge sequence based on the first matching edge when there are duplicate edges in the first matching edge and the second matching edge, so as to obtain the deduplicated second candidate edge sequence. The second side determination unit includes: The fourth edge determination unit is used to select the number of edges that are the first of the second matching edges in the deduplicated second candidate edge sequence as the second matching edges.

[0011] Optionally, the first feature extraction module includes: The occupational analysis unit is used to input the occupational description information of the object into a preset occupational analysis model to perform occupational analysis, and obtain at least one target object occupational information and at least one target object industry information corresponding to the target object. The feature extraction unit is used to extract features from the occupational information and industry information of the at least one target object to obtain the occupational feature information of the object.

[0012] Optionally, the object's occupational feature information includes an object's occupational feature vector and object's occupational feature word segmentation; the feature extraction unit includes: A splicing unit is used to splice the occupational information and industry information of the at least one target object to obtain spliced ​​occupational information; The word segmentation processing unit is used to segment the spliced ​​occupational information to obtain the word segmentation of the object's occupational features; The vectorization processing unit is used to perform vectorization processing on the spliced ​​occupational information to obtain the occupational feature vector of the object.

[0013] Optionally, the device further includes: The third information acquisition module is used to acquire multiple preset product occupation information corresponding to the multiple preset insurance products and preset product industry information corresponding to each preset product occupation information; any preset product occupation information is used to indicate the occupation information applicable to the corresponding preset insurance product, and the preset product industry information corresponding to any preset product occupation information is the industry information to which the preset product occupation information belongs; The occupation deduplication module is used to deduplicatize the multiple preset product occupation information to obtain the multiple preset occupation information. The relationship graph construction module is used to construct the preset node relationship graph with the multiple preset insurance products as product nodes, the multiple preset occupational information as occupational nodes, and the association between any preset insurance product and any preset occupational information as edges. The product occupation determination module is used to determine the target product occupation information corresponding to each edge based on the occupation nodes connected to each edge; The product industry determination module is used to determine the target product industry information corresponding to each edge based on the target product occupation information corresponding to each edge and the preset product industry information; The second feature extraction module is used to extract features from the target product occupation information and the target product industry information corresponding to each edge, so as to obtain the preset occupation feature information corresponding to each edge. The first edge attribute configuration module is used to write the preset occupational characteristic information corresponding to each edge as the edge attribute of each edge into each edge.

[0014] Optionally, the device further includes: The fourth information acquisition module is used to acquire at least one updated product occupation information corresponding to the updated insurance product and updated product industry information corresponding to each updated product occupation information when receiving an update instruction for the product node of the preset node relationship diagram. The product node generation module is used to create the updated product node corresponding to the updated insurance product in the preset node relationship diagram. The associated node determination module is used to determine at least one associated occupational node corresponding to the updated insurance product from multiple occupational nodes in the preset node relationship graph based on the at least one updated product occupational information. The edge generation module is used to generate update edges between each associated occupation node and the updated product node in the preset node relationship graph; The third feature extraction module is used to extract features from the preset occupational information and the updated product industry information corresponding to each associated occupational node to obtain the updated occupational feature information corresponding to each associated occupational node. The second side attribute configuration module is used to write the updated occupation feature information corresponding to each associated occupation node as the side attribute of the updated edge corresponding to each associated occupation node into the updated edge corresponding to each associated occupation node.

[0015] Optionally, the device further includes: The subgraph determination module is used to determine the target relationship subgraph corresponding to the target object based on the at least one target matching edge and the preset node relationship graph; the target relationship subgraph includes the at least one target matching edge and the nodes connected to the at least one target matching edge; The inductive summary module is used to input the target relationship subgraph into a preset natural language model for inductive summary processing to obtain product matching description information corresponding to the target object.

[0016] According to another aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the above-described insurance product matching method.

[0017] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the above-described insurance product matching method.

[0018] According to another aspect of the present disclosure, a computer program product containing instructions is provided that, when run on a computer, causes the computer to perform the above-described insurance product matching method.

[0019] The technical solutions provided by the embodiments of this disclosure bring at least the following beneficial effects: This process involves obtaining the target object's occupational description information and a pre-defined node relationship graph corresponding to multiple pre-defined insurance products. The pre-defined node relationship graph is constructed using multiple pre-defined insurance products and multiple pre-defined occupational information as nodes, and the relationships between any pre-defined insurance product and any pre-defined occupational information as edges. Each edge in the pre-defined node relationship graph includes pre-defined occupational feature information. The pre-defined occupational feature information of each edge represents the pre-defined occupational information corresponding to the occupational node connected to each edge, and the pre-defined industry information to which the pre-defined occupational information corresponding to each product node connected to each edge belongs. This allows for the acquisition of the target object's occupational description and the pre-defined node relationship graph. Then, feature extraction is performed on the object's occupational description information to obtain the target object's corresponding occupational feature information. The object's occupational feature information represents the target... By analyzing the object's occupational information and the industry information of that occupation, accurate extraction of the target object's occupational characteristics can be achieved. Then, combining this occupational characteristic information, each edge in the preset node relationship graph is matched to obtain at least one target matching edge corresponding to the target object. This ensures accurate edge matching in the preset node relationship graph, avoiding inaccurate matching caused by semantic differences in occupational information when occupational information is the same but industry information differs. Next, by combining the target product nodes corresponding to each of the at least one target matching edge, at least one target insurance product corresponding to the target object is determined. This enables precise matching of insurance products based on the target object's occupation, avoiding inaccurate matching caused by semantic differences in occupational information when occupational information is the same but industry information differs, thereby improving the user experience.

[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0022] Figure 1 This is a schematic diagram illustrating an application system according to an exemplary embodiment; Figure 2 This is a flowchart illustrating an insurance product matching method according to an exemplary embodiment; Figure 3 This is a flowchart illustrating an insurance product matching method according to an exemplary embodiment; Figure 4 This is a schematic diagram illustrating a preset node relationship diagram according to an exemplary embodiment; Figure 5 This is a block diagram illustrating an insurance product matching device according to an exemplary embodiment; Figure 6 This is a block diagram illustrating an electronic device for matching target insurance products to a target object, according to an exemplary embodiment. Figure 7 This is a block diagram illustrating another electronic device for matching target insurance products to a target object, according to an exemplary embodiment. Detailed Implementation

[0023] Various exemplary embodiments, features, and aspects of this application will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0024] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0025] Furthermore, to better illustrate this application, numerous specific details are provided in the following detailed embodiments. Those skilled in the art should understand that this application can be implemented without certain specific details. In some instances, methods, means, components, and circuits well-known to those skilled in the art have not been described in detail in order to highlight the main points of this application.

[0026] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating an application system according to an exemplary embodiment. The application system can be used in the insurance product matching method of this application. The application system may include at least a server 01 and a terminal 02.

[0027] In this embodiment, server 01 can be used to match target insurance products for a target object based on the object's occupational description information. Specifically, server 01 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0028] In this embodiment, terminal 02 can be used to generate job description information corresponding to the target object. Terminal 02 may include physical devices such as smartphones, desktop computers, tablets, laptops, smart speakers, in-vehicle terminals, digital assistants, augmented reality (AR) / virtual reality (VR) devices, and smart wearable devices, and may also include software running on the physical device, such as applications. The operating system running on terminal 02 in this embodiment may include, but is not limited to, Android, GNU / Linux, and Windows systems.

[0029] In addition, it should be noted that, Figure 1 The example shown is merely one application environment provided by this disclosure. In practical applications, other application environments may also be included. For example, the process of matching target insurance products to target targets by combining the target's occupational description information can also be implemented on the terminal.

[0030] In the embodiments described in this specification, the terminal 02 and the server 01 can be directly or indirectly connected through wired or wireless communication, and this application does not limit this connection.

[0031] It should be noted that this specification provides method operation steps as shown in the embodiments or flowcharts, but based on conventional or non-inventive labor, more or fewer operation steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many steps and does not represent the only execution order.

[0032] Specifically, Figure 2 This is a flowchart illustrating an insurance product matching method according to an exemplary embodiment. Figure 2 As shown, this insurance product matching method can be used in electronic devices such as terminals or servers, and may specifically include the following steps: S201: Obtain the job description information of the target object and the preset node relationship diagram corresponding to multiple preset insurance products.

[0033] In one specific embodiment, the target object can refer to the object that needs to be matched with insurance products. Specifically, the target object may include user accounts, etc.

[0034] In one specific embodiment, the object's occupational description information can be used to describe the target object's occupation. Specifically, the object's occupational description information may include multimodal description information such as occupational description text information, occupational description audio information, or occupational description image information. Further, the object's occupational description information may include at least one of occupational identity description information, industry description information, work environment description information, or job content description information.

[0035] In one specific embodiment, the target object can perform a job description input operation on the first terminal, so that the first terminal can obtain the job description information input by the target object; correspondingly, the first terminal can send the job description information to the server. The first terminal can be the user terminal corresponding to the target object.

[0036] In one specific embodiment, the multiple preset insurance products can be multiple insurance products to be matched. Specifically, each preset insurance product can be applicable to at least one occupation. It is understood that when any preset insurance product is applicable to multiple occupations, the preset occupation information corresponding to any preset insurance product can be different from each other.

[0037] In one specific embodiment, the preset node relationship graph can be constructed using multiple preset insurance products and multiple preset occupational information as nodes, and the association between any preset insurance product and any preset occupational information as edges. Specifically, the preset node relationship graph may include multiple occupational nodes, multiple product nodes, and multiple edges.

[0038] In one specific embodiment, any occupation node may store preset occupation information corresponding to that occupation node. Any product node may store multiple preset occupation information adapted to a preset insurance product corresponding to that product node, as well as industry information for each preset occupation and insurance product attribute information for the preset insurance product corresponding to that product node. Furthermore, the insurance product attribute information for the preset insurance product corresponding to any product node may include insurance product identification information for the preset insurance product corresponding to that product node.

[0039] In one specific embodiment, each edge in the preset node relationship graph may include preset occupational characteristic information. Specifically, the preset occupational characteristic information of each edge can characterize the preset industry information to which the preset occupational information corresponding to the product node connected to each edge belongs, within the preset occupational information corresponding to the occupational node connected to each edge and the industry information corresponding to the occupational node. Further, any edge may also include occupational type indication information to which the preset occupational information of the occupational node connected to the product node connected to the aforementioned edge belongs, among multiple preset occupational information. The occupational type indication information can be used to indicate the corresponding occupational type.

[0040] For example, suppose the preset node relationship graph includes occupation node A and product node B, and also includes edge AB connecting occupation node A and product node B. The preset occupation information of occupation node A is "occupation information 1". The multiple preset occupation information that product node B can adapt to include "occupation information 1", "occupation information 2" and "occupation information 3". The industry information corresponding to "occupation information 1" is "industry information 1", the industry information corresponding to "occupation information 2" is "industry information 2", and the industry information corresponding to "occupation information 3" is "industry information 3". Then the preset occupation feature information in the above-mentioned edge AB can be obtained by feature extraction of "occupation information 1" and "industry information 1". It can be understood that edge AB can express that the preset insurance product corresponding to product node B can be applied to occupation "occupation information 1", and that the occupation is specifically limited to industry "industry information 1". Thus, the preset occupation feature information of edge AB can accurately express occupation semantics, avoiding the problem of inaccurate matching caused by different occupation semantics when occupation information is the same but their industry information is different, and improving the accuracy of subsequent occupation matching.

[0041] For example, taking the occupational information "bookbinder" as an example, this occupational information can correspond to different industry information. For instance, it could include industry information such as "publishing and advertising," or "manufacturing / printing," or "culture, sports, and entertainment / news and publishing," or "news and advertising / news and magazines," etc. It is understandable that when the industry information corresponding to the occupational information is different, the occupational semantics determined based on that industry information will also be different.

[0042] In a specific embodiment, the preset node relationship diagram may be obtained in the following ways: Obtain occupational information for multiple preset insurance products and industry information for each preset product's occupational information; Multiple preset product occupation information is deduplicated to obtain multiple preset occupation information; A pre-defined node relationship graph is constructed using multiple pre-defined insurance products as product nodes, multiple pre-defined occupational information as occupational nodes, and the relationship between any pre-defined insurance product and any pre-defined occupational information as edges. Based on the occupational nodes connected to each edge, determine the target product occupational information corresponding to each edge; Based on the target product occupation information and preset product industry information corresponding to each edge, determine the target product industry information corresponding to each edge; Feature extraction is performed on the target product occupation information and target product industry information corresponding to each edge to obtain the preset occupation feature information corresponding to each edge; The preset occupational characteristic information corresponding to each edge is written as the edge attribute of each edge.

[0043] In one specific embodiment, any preset product occupational information can be used to indicate the occupational information applicable to the corresponding preset insurance product. Specifically, any preset product occupational information may include an occupational name and the corresponding product identification information of the preset insurance product.

[0044] In one specific embodiment, the preset product industry information corresponding to any preset product occupational information can be the industry information to which any preset product occupational information belongs. Specifically, any preset product industry information may include the industry name and the product identification information of the corresponding preset insurance product.

[0045] In one specific embodiment, the service provider can perform a product information input operation for each preset insurance product, enabling the second terminal to obtain multiple preset product occupational information and preset product industry information corresponding to each preset product occupational information. Accordingly, the second terminal can send the multiple preset product occupational information and preset product industry information corresponding to each preset insurance product to the server. Here, the second terminal can refer to the terminal corresponding to the service provider.

[0046] In a specific embodiment, the occupational information of multiple preset insurance products corresponding to multiple preset insurance products can be deduplicated to obtain the aforementioned multiple preset occupational information.

[0047] In one specific embodiment, the target product occupation information corresponding to either side can be used to indicate the preset occupation information corresponding to the occupation node connected to either side.

[0048] In a specific embodiment, the target product industry information corresponding to any side can be used to indicate the industry information to which the preset occupation information of the occupation node connected to any side belongs among a plurality of preset product industry information applicable to the product node connected to the aforementioned side.

[0049] In a specific embodiment, multiple preset product industry information corresponding to the preset insurance products corresponding to the product nodes connected to any one of the above sides can be determined based on the preset insurance products. The preset product industry information to which the preset product occupation information corresponding to the occupation node connected to any one of the above sides belongs can be used as the target product industry information corresponding to any one of the above sides.

[0050] In a specific embodiment, the preset occupational feature information corresponding to any side can be used to indicate the preset occupational information of the occupational node connected to the aforementioned side, and the industry information to which the preset occupational information of the occupational node connected to the aforementioned side belongs among multiple preset product industry information applicable to the product node connected to the aforementioned side. Specifically, the preset occupational feature information corresponding to any side may include a preset occupational feature vector and a preset occupational feature word segmentation corresponding to any side. The preset occupational feature vector of any side can represent the occupational semantics corresponding to any side in vector form. The preset occupational feature word segmentation of any side can represent the occupational semantics of any side in word segmentation form. The occupational semantics corresponding to any side can refer to the semantics of the preset occupational information corresponding to any side determined based on the preset industry information corresponding to any side.

[0051] In a specific embodiment, the target product occupation information and the target product industry information corresponding to either side can be concatenated to obtain the concatenated product occupation information corresponding to either side; the concatenated product occupation information corresponding to either side can be segmented to obtain the preset occupation feature word segmentation corresponding to either side; the concatenated product occupation information corresponding to either side can be vectorized to obtain the preset occupation feature vector corresponding to either side.

[0052] In the above embodiments, by acquiring multiple preset product occupation information corresponding to multiple preset insurance products and preset product industry information corresponding to each preset product occupation information, and deduplicating the multiple preset product occupation information, multiple preset occupation information is obtained. Using multiple preset insurance products as product nodes, multiple preset occupation information as occupation nodes, and the relationship between any preset insurance product and any preset occupation information as edges, a preset node relationship graph is constructed. Based on the occupation nodes connected to each edge, the target product occupation information corresponding to each edge is determined. Based on the target product occupation information and preset product industry information corresponding to each edge, the target product industry information corresponding to each edge is determined. Feature extraction is performed on the target product occupation information and target product industry information corresponding to each edge to obtain the preset occupation feature information corresponding to each edge. The preset occupation feature information corresponding to each edge is used as an edge attribute and written to each edge. This enables the construction of the preset node relationship graph, making the relationship graph representing the relationship between occupations and products more concise, facilitating node updates and data updates and maintenance. Furthermore, by using the preset occupation feature information corresponding to each edge as an edge attribute, the preset occupation semantics can be accurately expressed, improving the accuracy of occupation matching.

[0053] In one specific embodiment, the above method may further include: Upon receiving a product node update instruction for a preset node relationship diagram, obtain at least one updated product occupation information corresponding to the updated insurance product and updated product industry information corresponding to each updated product occupation information. Create an updated product node corresponding to the updated insurance product in the preset node relationship graph; Based on at least one updated product occupation information, determine at least one associated occupation node corresponding to the updated insurance product from multiple occupation nodes in a preset node relationship graph; Generate update edges between each associated occupation node and the updated product node in the preset node relationship graph; Feature extraction is performed on the preset occupational information and the updated product industry information corresponding to each associated occupational node to obtain the updated occupational feature information corresponding to each associated occupational node. The updated occupational feature information corresponding to each associated occupational node is used as the edge attribute of the updated edge corresponding to each associated occupational node, and written into the updated edge corresponding to each associated occupational node.

[0054] In one specific embodiment, a product node update instruction can be used to instruct the addition of product nodes corresponding to at least one insurance product to be updated to a preset node relationship graph. The product node update instruction may include product attribute information corresponding to each of the at least one insurance product to be updated, at least one preset product occupation information corresponding to each of the at least one insurance product to be updated, and preset product industry information corresponding to each preset product occupation information.

[0055] In one specific embodiment, the insurance product to be updated can be any one of the at least one insurance product to be updated mentioned above. Specifically, any one of the at least one insurance product to be updated corresponding to the product node update instruction can be used as the insurance product to be updated.

[0056] In one specific embodiment, at least one associated occupational node can refer to an occupational node in a preset node relationship graph that is associated with the updated insurance product. Specifically, occupational nodes in the preset node relationship graph whose preset occupational information is the same as the occupational information of each updated product can be used as at least one associated occupational node corresponding to the aforementioned updated insurance product.

[0057] In a specific embodiment, if any of the updated product occupation information is different from the preset occupation information corresponding to the occupation node in the preset node relationship graph, a corresponding updated occupation node can be created in the preset node relationship graph based on any of the updated product occupation information, and the updated product occupation information can be written into the updated occupation node as a node attribute of the updated occupation node.

[0058] In one specific embodiment, the update edge corresponding to each associated occupation node can be used to indicate the association between each associated occupation node and the updated product node. Specifically, the update edge corresponding to each associated occupation node can store the updated occupation feature information corresponding to each associated occupation node.

[0059] In one specific embodiment, the updated occupational feature information corresponding to any associated occupational node can be used to indicate the industry information to which the preset occupational information corresponding to the updated product node belongs in the occupational information corresponding to any updated product occupational node and the industry information corresponding to any associated occupational node. Specifically, the feature extraction process for the preset occupational information corresponding to each associated occupational node and the updated product industry information corresponding to each associated occupational node can refer to the aforementioned specific feature extraction process for the target product occupational information and the target product industry information, which will not be repeated here.

[0060] In the above embodiments, upon receiving an update instruction for a product node in a preset node relationship graph, at least one updated product occupational information and updated product industry information corresponding to each updated product occupational information are obtained. An updated product node corresponding to the updated insurance product is created in the preset node relationship graph. Based on at least one updated product occupational information, at least one associated occupational node corresponding to the updated insurance product is determined from multiple occupational nodes in the preset node relationship graph. An update edge is generated between each associated occupational node and the updated product node in the preset node relationship graph. Feature extraction is performed on the preset occupational information and updated product industry information corresponding to each associated occupational node to obtain updated occupational feature information corresponding to each associated occupational node. This updated occupational feature information is used as the edge attribute of the update edge corresponding to each associated occupational node and written into the update edge corresponding to each associated occupational node. This enables accurate updating of the preset node relationship graph when an insurance product is updated, improving the update dimensional efficiency of the preset node relationship graph.

[0061] S203: Extract features from the occupational description information of the target object to obtain the occupational feature information corresponding to the target object.

[0062] In a specific embodiment, object occupational feature information can represent the target object's occupational information and the industry information in which that occupational information belongs. Specifically, object occupational feature information can include object occupational feature vectors and object occupational feature word segmentation. The object occupational feature vector can represent the occupational semantics corresponding to the target object in vector form. The object occupational feature word segmentation can represent the occupational semantics corresponding to the target object in word segmentation form. The occupational semantics corresponding to the target object can refer to the semantics of the target object's occupational information determined based on the industry information of the target object's occupation. It can be understood that the occupational semantics corresponding to the target object carries both the semantic information of its occupation and the semantic information of the industry in which its occupation belongs, making the occupational semantics of the target object represented by the above-mentioned object occupational feature information more accurate.

[0063] In a specific embodiment, the above-mentioned feature extraction of the object's occupational description information to obtain the object's occupational feature information corresponding to the target object may include: Input the object's occupational description information into a preset occupational analysis model to perform occupational analysis, and obtain at least one target object occupational information and at least one target object industry information corresponding to the target object. Feature extraction is performed on at least one target object's occupational information and at least one target object's industry information to obtain the object's occupational feature information.

[0064] In one specific embodiment, a pre-defined occupational analysis model can be used to analyze the occupation and industry of a target object based on its occupational description information. Specifically, the pre-defined occupational analysis model can be a pre-trained natural language model. Optionally, the pre-defined natural language model can include ChatGPT (Chat Generative Pre-trained Transformer) or BERT (Bidirectional Encoder Representation from Transformers) models, etc.

[0065] In one specific embodiment, the occupational information of any target object can be used to indicate an occupation predicted by a preset occupational analysis model that conforms to the above-mentioned occupational description information of the target object. The occupational information of any target object may include an occupational name.

[0066] In one specific embodiment, the target object industry information can be used to indicate the industries predicted by a preset occupational analysis model that match the aforementioned occupational description information for the target object. The target object industry information may include the industry name.

[0067] In one specific embodiment, the object's occupational description information and preset occupational analysis instructions can be input into a preset occupational analysis model to perform occupational analysis, thereby obtaining at least one target object occupational information and at least one target object industry information corresponding to the target object.

[0068] In one specific embodiment, a preset occupational analysis instruction can be used to instruct a preset occupational analysis model to combine object occupational description information to output at least one occupation and at least one industry that conforms to the description.

[0069] In a specific embodiment, when the object occupation feature information includes an object occupation feature vector and object occupation feature word segmentation, the above-mentioned feature extraction of at least one target object occupation information and at least one target object industry information to obtain the object occupation feature information may include: The occupational information and industry information of at least one target object are concatenated to obtain concatenated occupational information. The concatenated occupational information is segmented to obtain the occupational feature words of the object. The spliced ​​occupational information is vectorized to obtain the object's occupational feature vector.

[0070] In one specific embodiment, the splicing of occupational information can be obtained by splicing at least one target object occupational information and at least one target object industry information. For example, assuming that at least one target object occupational information is "Occupational Information 1" and "Occupational Information 2", and at least one target object industry information is "Industry Information 1" and "Industry Information 2", the spliced ​​occupational information can be determined to be "Occupational Information 1, Occupational Information 2, Industry Information 1, Industry Information 2".

[0071] In a specific embodiment, a preset whitelist word segmentation set can be set, wherein the preset whitelist word segmentation set may include multiple preset whitelist word segments; by matching the concatenated occupational information with the preset whitelist word segmentation set, if a match is found, the matched word is used as the object occupational feature word segmentation; if no match is found in the preset whitelist word segmentation set, the whole can be used directly as the object occupational feature word segmentation without splitting.

[0072] In a specific embodiment, multiple segmented words can be obtained by segmenting the spliced ​​occupational information. Correspondingly, these multiple segmented words can be used as the occupational feature words of the object.

[0073] In one specific embodiment, the concatenated occupational information can be input into a preset encoding model for vectorization processing to obtain the object's occupational feature vector. Specifically, the preset encoding model may include a word vector model, a bidirectional encoder representation model, or a semantic sentence vector model, etc.

[0074] In the above embodiments, by inputting the object's occupational description information into a preset occupational analysis model for occupational analysis, at least one target object occupational information and at least one target object industry information corresponding to the target object are obtained. Users can describe their occupational information from multiple dimensions, which can reduce the threshold of users' terminology and also enable the preset occupational analysis model to better understand the target object's occupational information. The model can also better capture the dynamics and complexity of occupations. Then, feature extraction is performed on at least one target object occupational information and at least one target object industry information to obtain object occupational feature information, which can achieve accurate extraction of the target object's occupational features, thereby improving the accurate matching of occupations and their corresponding insurance products.

[0075] S205: Based on the object's occupational characteristic information, perform matching processing on each edge in the preset node relationship graph to obtain at least one target matching edge corresponding to the target object.

[0076] In a specific embodiment, at least one target matching edge can refer to an edge in the preset node relationship graph that matches the preset occupational feature information with the object's occupational feature information.

[0077] In related technologies, a sequential processing method of "matching industries first and then occupations" or "screening industries first and then calculating occupational similarity" can be adopted. Industry information is only used as a basis for judging whether the conditions are met, but it is not used as a continuous and measurable semantic feature to participate in the overall matching score. It is difficult to achieve a refined quantification of occupational similarity, which leads to low accuracy in matching occupations and insurance products.

[0078] In one specific embodiment, the preset occupational feature information for each edge may include a preset occupational feature vector and preset occupational feature word segments. The object's occupational feature information may include an object's occupational feature vector and object's occupational feature word segments.

[0079] In a specific embodiment, the above-mentioned matching process for each edge in the preset node relationship graph based on the object's occupational characteristic information to obtain at least one target matching edge corresponding to the target object may include: Vector matching is performed on the object's occupational feature vector and the preset occupational feature vector of each edge to obtain the vector matching result; The word segmentation matching process is performed on the word segmentation of the object's occupational features and the word segmentation of the preset occupational features of each edge to obtain the word segmentation matching result; Based on vector matching results and word segmentation matching results, at least one target matching edge is selected from multiple edges of the preset node relationship graph.

[0080] In one specific embodiment, the vector matching result can be used to indicate the degree of matching between the preset occupational feature vector of each edge and the object's occupational feature vector. The vector matching result can include the vector matching data corresponding to each edge. The vector matching data corresponding to any edge can characterize the degree of matching between the preset occupational feature vector of any edge and the object's occupational feature vector.

[0081] In a specific embodiment, the vector similarity can be calculated between the preset occupational feature vector of each edge and the object's occupational feature vector to obtain the vector matching data corresponding to each edge.

[0082] In one specific embodiment, the word segmentation matching result can be used to indicate the degree of matching between the preset occupational feature word segmentation and the object occupational feature word segmentation of each edge. The word segmentation matching result can include the word segmentation matching data corresponding to each edge. The word segmentation matching data corresponding to any edge can characterize the degree of matching between the preset occupational feature word segmentation and the object occupational feature word segmentation of any edge. Furthermore, the vector matching result can also include the edge identification information of each edge among the matched edges.

[0083] In a specific embodiment, semantic features can be extracted from the word segmentation of the object's occupational features to obtain the first word segmentation semantic information corresponding to the word segmentation of the object's occupational features; semantic features can be extracted from the preset occupational feature word segmentation of each edge to obtain the second word segmentation semantic information corresponding to each edge; semantic similarity can be calculated between the first word segmentation semantic information and the second word segmentation semantic information corresponding to each edge to obtain the word segmentation matching data corresponding to each edge. Furthermore, the word segmentation matching result may also include the edge identifier information of each edge among the matched edges.

[0084] In a specific embodiment, the vector matching results and word segmentation matching results can be initially screened to obtain initially screened vector matching results and initially screened word segmentation matching results. Correspondingly, based on the initially screened vector matching results and initially screened word segmentation matching results, at least one target matching edge can be selected from multiple edges of a preset node relationship graph. Specifically, edges in the vector matching results whose corresponding vector matching data is less than a first preset matching data can be filtered to obtain initially screened vector matching results; edges in the word segmentation matching results whose corresponding word segmentation matching data is less than a second preset matching data can be filtered to obtain initially screened word segmentation matching results. Specifically, the aforementioned first preset matching data and second preset matching data can be set according to actual application needs, and this disclosure does not limit them. Optionally, the value range of the aforementioned first preset matching data and second preset matching data can be [0.6-0.8]; for example, the aforementioned first preset matching data and second preset matching data can be 0.7.

[0085] In one specific embodiment, the above method may further include: Get the preset vector matching weights corresponding to the vector matching dimension and the preset word segmentation matching weights corresponding to the word segmentation matching dimension; Accordingly, the process of selecting at least one target matching edge from multiple edges of a preset node relationship graph based on vector matching results and word segmentation matching results can include: Based on the vector matching results, the multiple edges are sorted in descending order of the corresponding vector matching data to obtain a sequence of vector candidate edges; Based on the word segmentation matching results, the multiple edges are sorted in descending order of the corresponding word segmentation matching data to obtain the word segmentation candidate edge sequence; Based on preset vector matching weights and preset word segmentation matching weights, at least one target matching edge is determined from the sequence of candidate vector edges and the sequence of candidate word segments.

[0086] In one specific embodiment, a preset vector matching weight can be used to indicate the importance of the vector matching dimension. Specifically, the preset vector matching weight can be set according to the actual application needs, and this disclosure does not limit it. For example, the preset vector matching weight can be 0.8.

[0087] In one specific embodiment, a preset word segmentation matching weight can be used to indicate the importance of the word segmentation matching dimension. Specifically, the preset word segmentation matching weight can be set according to actual application needs, and this disclosure does not limit it. For example, the preset word segmentation matching weight can be 0.2.

[0088] In one specific embodiment, the vector candidate edge sequence can refer to a sequence of multiple edges sorted in descending order of their corresponding vector matching data. Specifically, the vector candidate edge sequence may include multiple edges. Furthermore, if the vector matching results have undergone initial screening, the aforementioned vector candidate edge sequence may include at least one edge from the initial screening.

[0089] In one specific embodiment, the word segmentation candidate edge sequence can refer to a sequence of multiple edges sorted in descending order of their corresponding word segmentation matching data. Specifically, the word segmentation candidate edge sequence may include multiple edges. Furthermore, if the word segmentation matching results have undergone initial screening, the aforementioned word segmentation candidate edge sequence may include at least one edge from the initial screening.

[0090] In a specific embodiment, when the vector matching result is the vector matching result after initial screening, multiple edges in the vector matching result after initial screening can be sorted in descending order of vector matching data to obtain a sequence of vector edges to be selected.

[0091] In a specific embodiment, when the word segmentation matching result is the word segmentation matching result after initial screening, multiple edges in the word segmentation matching result after initial screening can be sorted in descending order of word segmentation matching data to obtain a sequence of candidate edges for word segmentation.

[0092] In one specific embodiment, edges with the same proportion can be selected from the vector candidate edge sequence and the word candidate edge sequence based on the ratio of preset vector matching weight and preset word segmentation matching weight, and used as at least one target matching edge.

[0093] In a specific embodiment, determining at least one target matching edge from the sequence of candidate vector edges and the sequence of candidate word edges based on preset vector matching weights and preset word segmentation matching weights may include: Based on preset vector matching weights and preset word segmentation matching weights, the first matching dimension and the second matching dimension are determined from the vector matching dimension and the word segmentation matching dimension. Based on preset vector matching weights and preset word segmentation matching weights, determine the number of first matching edges corresponding to the first matching dimension and the number of second matching edges corresponding to the second matching dimension. From the vector candidate edge sequence and the word segmentation candidate edge sequence, determine the first candidate edge sequence corresponding to the first matching dimension and the second candidate edge sequence corresponding to the second matching dimension. The number of first matching edges in the first candidate edge sequence is taken as the first matching edge corresponding to the first matching dimension. The number of the first second matching edges in the second candidate edge sequence is taken as the second matching edge corresponding to the second matching dimension. The first matching edge and the second matching edge are used as at least one target matching edge.

[0094] In a specific embodiment, the first matching dimension can be the matching dimension corresponding to the largest matching weight among the preset vector matching weight and the preset word segmentation matching weight.

[0095] In one specific embodiment, the second matching dimension can be a matching dimension other than the first matching dimension, which is either the vector matching dimension or the word segmentation matching dimension.

[0096] In one specific embodiment, a preset vector matching weight and a preset word segmentation matching weight can be compared to obtain a first matching weight and a second matching weight. The first matching weight can be the largest of the preset vector matching weight and the preset word segmentation matching weight. The second matching weight can be the smallest of the preset vector matching weight and the preset word segmentation matching weight. Correspondingly, the matching dimension corresponding to the first matching weight in the vector matching dimension and the word segmentation matching dimension can be used as the first matching dimension; the matching dimension corresponding to the second matching weight in the vector matching dimension and the word segmentation matching dimension can be used as the second matching dimension.

[0097] In a specific embodiment, the first matching edge count can refer to the number of edges matched based on the first matching dimension among at least one target matching edge. The second matching edge count can refer to the number of edges matched based on the second matching dimension among at least one target matching edge. It is understood that when the first matching dimension is a vector matching dimension, the first matching edge count can refer to the number of edges matched based on the vector matching dimension among at least one target matching edge; when the first matching dimension is a word segmentation matching dimension, the first matching edge count can refer to the number of edges matched based on the word segmentation matching dimension among at least one target matching edge; correspondingly, when the second matching dimension is a vector matching dimension, the second matching edge count can refer to the number of edges matched based on the vector matching dimension among at least one target matching edge; when the second matching dimension is a word segmentation matching dimension, the second matching edge count can refer to the number of edges matched based on the word segmentation matching dimension among at least one target matching edge.

[0098] In a specific embodiment, a preset number of matching edges can be obtained; a target matching edge ratio can be determined based on preset vector matching weights and preset word segmentation matching weights; and a first matching edge number corresponding to a first matching dimension and a second matching edge number corresponding to a second matching dimension can be determined based on the target matching edge ratio and the preset number of matching edges. Here, the preset number of matching edges may refer to a preset expected number of target matching edges. The target matching edge ratio can be used to indicate the ratio of edges matched based on the word segmentation matching dimension to edges matched based on the vector matching dimension among at least one target matching edge. The target matching edge ratio may be the ratio between preset vector matching weights and preset word segmentation matching weights.

[0099] In a specific embodiment, the first candidate edge sequence can refer to the candidate edge sequence corresponding to the first matching dimension in both the vector candidate edge sequence and the word segmentation candidate edge sequence. Specifically, when the first matching dimension is the vector matching dimension, the first candidate edge sequence corresponding to the first matching dimension can be determined as the vector candidate edge sequence; when the first matching dimension is the word segmentation matching dimension, the first candidate edge sequence corresponding to the first matching dimension can be determined as the word segmentation candidate edge sequence.

[0100] In a specific embodiment, the second candidate edge sequence can refer to the candidate edge sequence corresponding to the second matching dimension in both the vector candidate edge sequence and the word segmentation candidate edge sequence. Specifically, when the second matching dimension is the vector matching dimension, the second candidate edge sequence corresponding to the second matching dimension can be determined as the vector candidate edge sequence; when the second matching dimension is the word segmentation matching dimension, the second candidate edge sequence corresponding to the second matching dimension can be determined as the word segmentation candidate edge sequence.

[0101] In one specific embodiment, the first matching edge can refer to an edge matched based on a first matching dimension from at least one target matching edge. The first matching edge may include a number of edges, and the corresponding matching data is among the first edges in the first candidate edge sequence.

[0102] In one specific embodiment, the second matching edge can refer to an edge matched based on a second matching dimension from at least one target matching edge. The second matching edge can include a number of edges, and the corresponding matching data is one of the earlier edges in the second candidate edge sequence.

[0103] In the above embodiments, based on preset vector matching weights and preset word segmentation matching weights, a first matching dimension and a second matching dimension are determined from the vector matching dimension and the word segmentation matching dimension. Based on the preset vector matching weights and preset word segmentation matching weights, the number of first matching edges corresponding to the first matching dimension and the number of second matching edges corresponding to the second matching dimension are determined. From the vector candidate edge sequence and the word candidate edge sequence, the first candidate edge sequence corresponding to the first matching dimension and the second candidate edge sequence corresponding to the second matching dimension are determined. The first number of edges in the first candidate edge sequence are taken as the first matching edges corresponding to the first matching dimension, and the first number of edges in the second candidate edge sequence are taken as the second matching edges corresponding to the second matching dimension. The first matching edges and the second matching edges are taken as at least one target matching edge. This can ensure that the edges matched by the matching dimension with the greater weight mainly occupy the above at least one target matching edge, while improving the diversity of at least one target matching edge through multi-dimensional matching. At the same time, it can avoid the result of one matching dimension occupying all the results and causing matching anomalies when any matching dimension is abnormal, thereby improving the stability of product matching.

[0104] In one specific embodiment, the above method may further include: If there are duplicate edges in the first and second matching edges, the second candidate edge sequence is deduplicated based on the first matching edge to obtain the deduplicated second candidate edge sequence. Accordingly, the above-mentioned method of taking the number of edges that are the first two matching edges in the second candidate edge sequence as the second matching edges corresponding to the second matching dimension can include: The number of edges that are the first two matching edges in the deduplicated second candidate edge sequence are taken as the second matching edges.

[0105] In one specific embodiment, the deduplicated second candidate edge sequence can be the sequence after removing edges that are duplicates of the first matching edge from the second candidate edge sequence. Specifically, the deduplicated second candidate edge sequence may include at least one edge.

[0106] In one specific embodiment, edges that are duplicates of the first matching edge in the second candidate edge sequence can be deleted to obtain a deduplicated second candidate edge sequence. Specifically, each edge in the first matching edge sequence can be compared with each edge in the second candidate edge sequence to obtain at least one duplicate edge in the second candidate edge sequence; correspondingly, at least one duplicate edge in the second candidate edge sequence can be deleted to obtain a deduplicated second candidate edge sequence. Here, any duplicate edge among the at least one duplicate edge can refer to an edge that exists in both the first matching edge sequence and the second candidate edge sequence.

[0107] In the above embodiments, when there are duplicate edges in the first matching edge and the second matching edge, the second candidate edge sequence is deduplicated based on the first matching edge to obtain a deduplicated second candidate edge sequence. The number of edges that are the first second matching edges in the deduplicated second candidate edge sequence are taken as the second matching edges. The first matching edge and the second matching edge are taken as at least one target matching edge. This can avoid the occurrence of duplicates in at least one target matching edge, thereby improving the effectiveness and diversity of the matching results of multi-dimensional matching.

[0108] In the above embodiments, by performing vector matching processing on the object's occupational feature vector and the preset occupational feature vector of each edge, the vector matching result can be obtained, which can realize occupational matching in the vector matching dimension. By performing word segmentation matching processing on the object's occupational feature word segmentation and the preset occupational feature word segmentation of each edge, the word segmentation matching result can be obtained, which can realize occupational matching in the word segmentation dimension. Based on the vector matching result and the word segmentation matching result, at least one target matching edge is selected from multiple edges of the preset node relationship graph, which can improve matching accuracy and robustness, overcome the inherent limitations of a single matching mode, and enhance the interpretability and controllability of the system.

[0109] S207: Based on the target product node corresponding to each of the at least one target matching edge, determine at least one target insurance product corresponding to the target object.

[0110] In one specific embodiment, the target product node can refer to a product node that is connected to any target matching edge.

[0111] In a specific embodiment, the preset insurance product indicated by the target product node corresponding to each of the above-mentioned at least one target matching edge can be regarded as at least one target insurance product corresponding to the target object.

[0112] In one specific embodiment, preset product description information corresponding to the target insurance product can be obtained based on the product attribute information in the target product node; the preset product description information corresponding to at least one target insurance product and the matching data corresponding to each target insurance product can be sent to enable the first terminal to display an insurance product matching page, wherein the insurance product matching page includes the preset product description information corresponding to at least one target insurance product and the matching data corresponding to each target insurance product.

[0113] In one specific embodiment, the above method may further include: Based on at least one target matching edge and a preset node relationship graph, determine the target relationship subgraph corresponding to the target object; The target relationship subgraph is input into a preset natural language model for inductive and summarizing processing to obtain product matching description information corresponding to the target object.

[0114] In one specific embodiment, the target relationship subgraph may include at least one target matching edge and nodes connected to the at least one target matching edge. It is understood that the target relationship subgraph may be a subgraph of a predefined node relationship graph.

[0115] In one specific embodiment, a target relationship subgraph can be determined based on nodes connected to at least one target matching edge and the connection relationship indicated by at least one target matching edge; wherein, the occupation nodes in the target relationship subgraph include preset occupation information, the product nodes in the target relationship subgraph include product attribute information corresponding to preset insurance products, and the edges in the target relationship subgraph include corresponding preset occupation feature information.

[0116] In one specific embodiment, the product matching description information can be used to describe the relevant information and matching status of the preset insurance products obtained by matching the occupation of the target object.

[0117] In one specific embodiment, the target relationship subgraph and a preset inductive summary instruction are input into a preset natural language model for inductive summary processing to obtain product matching description information corresponding to the target object. The preset inductive summary instruction can be used to instruct the preset natural language model to understand the target relationship subgraph and perform inductive summary for the target insurance product.

[0118] In one specific embodiment, the aforementioned product matching description information can be sent to enable the first terminal to display an insurance product matching page, which includes the aforementioned product matching description information.

[0119] In the above embodiments, based on at least one target matching edge and a preset node relationship graph, a target relationship subgraph corresponding to the target object is determined. The target relationship subgraph is input into a preset natural language model for inductive and summarizing processing to obtain product matching description information corresponding to the target object. The preset natural language model can utilize its strong understanding of relationship graphs to accurately understand and reason about the target relationship subgraph, thereby better summarizing product matching description information, improving the readability of the output results, and thus enhancing the user experience.

[0120] In the above embodiments, the target object's occupational description information and a preset node relationship graph corresponding to multiple preset insurance products are obtained. The preset node relationship graph is constructed with multiple preset insurance products and multiple preset occupational information as nodes, and the association relationship between any preset insurance product and any preset occupational information as edges. Each edge in the preset node relationship graph includes preset occupational feature information. The preset occupational feature information of each edge represents the preset industry information to which the preset occupational information corresponding to the product node connected to each edge belongs, in the preset occupational information and industry information corresponding to each occupational node. This can realize the acquisition of occupational-related descriptions and preset node relationship graphs for the target object. Then, feature extraction is performed on the object's occupational description information to obtain the target object's corresponding occupational feature information. By representing the target object's occupational information and the industry information of that occupation, accurate extraction of the target object's occupational characteristics can be achieved. Then, combining this occupational characteristic information, each edge in the preset node relationship graph is matched to obtain at least one target matching edge corresponding to the target object. This ensures accurate edge matching in the preset node relationship graph, avoiding inaccurate matching caused by semantic differences in occupational information when occupational information is the same but industry information differs. Finally, by combining the target product nodes corresponding to each of the at least one target matching edge, at least one target insurance product corresponding to the target object is determined. This enables precise matching of insurance products based on the target object's occupation, avoiding inaccurate matching caused by semantic differences in occupational information when occupational information is the same but industry information differs, thereby improving the user experience.

[0121] Figure 3 This is a flowchart illustrating an insurance product matching method according to an exemplary embodiment. Specifically, as shown below... Figure 3 As shown, a preset node relationship diagram can be constructed in advance; correspondingly, the target insurance product corresponding to the target object can be matched by combining the above preset node relationship diagram.

[0122] Furthermore, such as Figure 3As shown, the process of constructing the preset node relationship graph involves obtaining multiple preset product occupation information corresponding to multiple preset insurance products and preset product industry information corresponding to each preset product occupation information; deduplicating the multiple preset product occupation information to obtain multiple preset occupation information; constructing the preset node relationship graph using multiple preset insurance products as product nodes, multiple preset occupation information as occupation nodes, and the association between any preset insurance product and any preset occupation information as edges; determining the target product occupation information corresponding to each edge based on the occupation nodes connected to each edge; determining the target product industry information corresponding to each edge based on the target product occupation information and preset product industry information corresponding to each edge; extracting features from the target product occupation information and target product industry information corresponding to each edge to obtain the preset occupation feature information corresponding to each edge; and writing the preset occupation feature information corresponding to each edge as the edge attribute of each edge.

[0123] Figure 4 This is a schematic diagram illustrating a preset node relationship diagram according to an exemplary embodiment. Specifically, as shown below... Figure 4 As shown, the preset node relationship graph can include 4 occupational nodes (i.e. Figure 4 The nodes W1, W2, W3, and W4 are indicated by dashed circles, along with six product nodes (i.e., Figure 4 The nodes I1, I2, I3, I4, I5, and I6 are shown as solid circles. Each occupational node can be connected to different product nodes through multiple edges, and similarly, each product node can be connected to different occupational nodes through multiple edges. Taking occupational node W3 as an example, occupational node W3 can have an association relationship with product nodes I3, I6, and I5. Taking product node I6 as an example, the occupational information applicable to the preset insurance product corresponding to product node I6 includes the preset occupational information corresponding to occupational node W1 (and its industry information is limited to the corresponding preset industry information), the preset occupational information corresponding to occupational node W2 (and its industry information is limited to the corresponding preset industry information), the preset occupational information corresponding to occupational node W3 (and its industry information is limited to the corresponding preset industry information), and the preset occupational information corresponding to occupational node W4 (and its industry information is limited to the corresponding preset industry information). It is understandable that the preset insurance product corresponding to product node I6 may not be applicable to all industries of the preset occupation information corresponding to occupation node W1. The specific applicable industry information when product node I6 applies the preset occupation information of occupation node can be determined by the preset occupation feature information in the edge connecting product node I6 and occupation node.

[0124] Furthermore, such as Figure 3As shown, the matching process for target insurance products involves: obtaining the occupational description information of the target object; inputting the occupational description information into a preset occupational analysis model for occupational analysis to obtain at least one target object occupational information and at least one target object industry information; extracting features from the at least one target object occupational information and at least one target object industry information to obtain object occupational feature information; performing vector matching processing on the object occupational feature vector and the preset occupational feature vector of each edge to obtain vector matching results; performing word segmentation matching processing on the object occupational feature word segmentation and the preset occupational feature word segmentation of each edge to obtain word segmentation matching results; based on the vector matching results and word segmentation matching results, selecting at least one target matching edge from multiple edges in a preset node relationship graph; and determining at least one target insurance product corresponding to the target object based on the target product node corresponding to each of the at least one target matching edge.

[0125] Figure 5 This is a block diagram illustrating an insurance product matching device according to an exemplary embodiment. Specifically, as shown below... Figure 5 As shown, the device may include: The first information acquisition module 510 is used to acquire the object occupation description information corresponding to the target object and the preset node relationship diagram corresponding to multiple preset insurance products; the preset node relationship diagram is constructed with the multiple preset insurance products and multiple preset occupation information as nodes, and the association relationship between any preset insurance product and any preset occupation information as edges; each edge in the preset node relationship diagram includes preset occupation feature information, and the preset occupation feature information of each edge represents the preset industry information to which the preset occupation information corresponding to the product node connected to each edge belongs in the preset occupation information of the occupation node connected to each edge and the industry information corresponding to the occupation node; The first feature extraction module 520 is used to extract features from the object occupation description information to obtain the object occupation feature information corresponding to the target object; the object occupation feature information represents the object occupation information of the target object and the industry information in which the object occupation information is located. Matching processing module 530 is used to perform matching processing on each edge in the preset node relationship graph based on the object's occupational feature information to obtain at least one target matching edge corresponding to the target object; The target product determination module 540 is used to determine at least one target insurance product corresponding to the target object based on the target product node corresponding to each of the at least one target matching edge.

[0126] In an optional embodiment, the preset occupational feature information of each edge includes a preset occupational feature vector and preset occupational feature word segmentation, and the object occupational feature information includes an object occupational feature vector and object occupational feature word segmentation; the matching processing module includes: A vector matching unit is used to perform vector matching processing on the object's occupational feature vector and the preset occupational feature vector of each edge to obtain a vector matching result; the vector matching result is used to indicate the degree of matching between the preset occupational feature vector of each edge and the object's occupational feature vector; The word segmentation matching unit is used to perform word segmentation matching processing on the word segmentation of the object's occupational features and the preset word segmentation of each edge to obtain the word segmentation matching result; the word segmentation matching result is used to indicate the degree of matching between the preset word segmentation of each edge and the word segmentation of the object's occupational features; The matching edge determination unit is used to select at least one target matching edge from multiple edges of the preset node relationship graph based on the vector matching result and the word segmentation matching result.

[0127] In an optional embodiment, the vector matching result includes vector matching data corresponding to each edge, and the word segmentation matching result includes word segmentation matching data corresponding to each edge; the device further includes: The second information acquisition module is used to acquire the preset vector matching weights corresponding to the vector matching dimension and the preset word segmentation matching weights corresponding to the word segmentation matching dimension. The matching result filtering unit includes: The first sorting unit is used to sort the multiple edges according to the vector matching results in descending order of the corresponding vector matching data to obtain a sequence of vector candidate edges; The second sorting unit is used to sort the multiple edges according to the word segmentation matching results in descending order of the corresponding word segmentation matching data to obtain a word segmentation candidate edge sequence. The weighted filtering unit is used to determine at least one target matching edge from the vector candidate edge sequence and the word candidate edge sequence based on the preset vector matching weight and the preset word segmentation matching weight.

[0128] In an optional embodiment, the weighted filtering unit includes: The matching dimension determination unit is used to determine a first matching dimension and a second matching dimension from the vector matching dimension and the word segmentation matching dimension based on the preset vector matching weight and the preset word segmentation matching weight; the first matching dimension is the matching dimension corresponding to the largest matching weight among the preset vector matching weight and the preset word segmentation matching weight; the second matching dimension is the matching dimension other than the first matching dimension among the vector matching dimension and the word segmentation matching dimension. The edge number determination unit is used to determine the number of first matching edges corresponding to the first matching dimension and the number of second matching edges corresponding to the second matching dimension based on the preset vector matching weight and the preset word segmentation matching weight. A sequence determination unit is used to determine, from the vector candidate edge sequence and the word segmentation candidate edge sequence, a first candidate edge sequence corresponding to the first matching dimension and a second candidate edge sequence corresponding to the second matching dimension; The first edge determination unit is used to take the number of first matching edges that are the first edge in the first candidate edge sequence as the first matching edge corresponding to the first matching dimension. The second side determination unit is used to take the number of the first second matching edges in the second candidate edge sequence as the second matching edges corresponding to the second matching dimension. The third side determination unit is used to determine the first matching edge and the second matching edge as the at least one target matching edge.

[0129] In an optional embodiment, the apparatus further includes: The edge deduplication module is used to perform deduplication processing on the second candidate edge sequence based on the first matching edge when there are duplicate edges in the first matching edge and the second matching edge, so as to obtain the deduplicated second candidate edge sequence. The second side determination unit includes: The fourth edge determination unit is used to select the number of edges that are the first of the second matching edges in the deduplicated second candidate edge sequence as the second matching edges.

[0130] In an optional embodiment, the first feature extraction module includes: The occupational analysis unit is used to input the occupational description information of the object into a preset occupational analysis model to perform occupational analysis, and obtain at least one target object occupational information and at least one target object industry information corresponding to the target object. The feature extraction unit is used to extract features from the occupational information and industry information of the at least one target object to obtain the occupational feature information of the object.

[0131] In an optional embodiment, the object occupational feature information includes an object occupational feature vector and object occupational feature word segmentation; the feature extraction unit includes: A splicing unit is used to splice the occupational information and industry information of the at least one target object to obtain spliced ​​occupational information; The word segmentation processing unit is used to segment the spliced ​​occupational information to obtain the word segmentation of the object's occupational features; The vectorization processing unit is used to perform vectorization processing on the spliced ​​occupational information to obtain the occupational feature vector of the object.

[0132] In an optional embodiment, the apparatus further includes: The third information acquisition module is used to acquire multiple preset product occupation information corresponding to the multiple preset insurance products and preset product industry information corresponding to each preset product occupation information; any preset product occupation information is used to indicate the occupation information applicable to the corresponding preset insurance product, and the preset product industry information corresponding to any preset product occupation information is the industry information to which the preset product occupation information belongs; The occupation deduplication module is used to deduplicatize the multiple preset product occupation information to obtain the multiple preset occupation information. The relationship graph construction module is used to construct the preset node relationship graph with the multiple preset insurance products as product nodes, the multiple preset occupational information as occupational nodes, and the association between any preset insurance product and any preset occupational information as edges. The product occupation determination module is used to determine the target product occupation information corresponding to each edge based on the occupation nodes connected to each edge; The product industry determination module is used to determine the target product industry information corresponding to each edge based on the target product occupation information corresponding to each edge and the preset product industry information; The second feature extraction module is used to extract features from the target product occupation information and the target product industry information corresponding to each edge, so as to obtain the preset occupation feature information corresponding to each edge. The first edge attribute configuration module is used to write the preset occupational characteristic information corresponding to each edge as the edge attribute of each edge into each edge.

[0133] In an optional embodiment, the apparatus further includes: The fourth information acquisition module is used to acquire at least one updated product occupation information corresponding to the updated insurance product and updated product industry information corresponding to each updated product occupation information when receiving an update instruction for the product node of the preset node relationship diagram. The product node generation module is used to create the updated product node corresponding to the updated insurance product in the preset node relationship diagram. The associated node determination module is used to determine at least one associated occupational node corresponding to the updated insurance product from multiple occupational nodes in the preset node relationship graph based on the at least one updated product occupational information. The edge generation module is used to generate update edges between each associated occupation node and the updated product node in the preset node relationship graph; The third feature extraction module is used to extract features from the preset occupational information and the updated product industry information corresponding to each associated occupational node to obtain the updated occupational feature information corresponding to each associated occupational node. The second side attribute configuration module is used to write the updated occupation feature information corresponding to each associated occupation node as the side attribute of the updated edge corresponding to each associated occupation node into the updated edge corresponding to each associated occupation node.

[0134] In an optional embodiment, the apparatus further includes: The subgraph determination module is used to determine the target relationship subgraph corresponding to the target object based on the at least one target matching edge and the preset node relationship graph; the target relationship subgraph includes the at least one target matching edge and the nodes connected to the at least one target matching edge; The inductive summary module is used to input the target relationship subgraph into a preset natural language model for inductive summary processing to obtain product matching description information corresponding to the target object.

[0135] Regarding the apparatus in the above embodiments, the specific manner in which each module and unit performs its operations has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0136] Figure 6 This is a block diagram illustrating an electronic device for matching target insurance products to a target object, according to an exemplary embodiment. The electronic device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown, the electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements an insurance product matching method.

[0137] Figure 7 This is a block diagram illustrating another electronic device for matching target insurance products to a target object, according to an exemplary embodiment. The electronic device may be a terminal, and its internal structure diagram may be as follows: Figure 7 As shown, the electronic device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements an insurance product matching method. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0138] Those skilled in the art will understand that Figure 6 or Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the electronic device to which the present disclosure is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0139] In an exemplary embodiment, an electronic device is also provided, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the insurance product matching method as described in the embodiments of this disclosure.

[0140] In an exemplary embodiment, a computer-readable storage medium is also provided, wherein instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the insurance product matching method of the present disclosure embodiments.

[0141] In an exemplary embodiment, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to perform the insurance product matching method of the present disclosure embodiments.

[0142] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0143] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0144] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for matching insurance products, characterized in that, The method includes: Obtain the target object's occupational description information and the preset node relationship graph corresponding to multiple preset insurance products; the preset node relationship graph is constructed with the multiple preset insurance products and multiple preset occupational information as nodes, and the association relationship between any preset insurance product and any preset occupational information as edges; each edge in the preset node relationship graph includes preset occupational feature information, and the preset occupational feature information of each edge represents the preset industry information to which the preset occupational information corresponding to the product node connected to each edge belongs in the preset occupational information of the occupational node connected to each edge and the industry information corresponding to the occupational node; Feature extraction is performed on the object's occupational description information to obtain the object's occupational feature information corresponding to the target object; the object's occupational feature information represents the target object's occupational information and the industry information in which the object's occupational information is located. Based on the occupational characteristic information of the object, each edge in the preset node relationship graph is matched to obtain at least one target matching edge corresponding to the target object; Based on the target product node corresponding to each of the at least one target matching edge, at least one target insurance product corresponding to the target object is determined.

2. The method according to claim 1, characterized in that, The preset occupational feature information of each edge includes a preset occupational feature vector and a preset occupational feature word segmentation; the object occupational feature information includes an object occupational feature vector and an object occupational feature word segmentation; the matching process performed on each edge in the preset node relationship graph based on the object occupational feature information to obtain at least one target matching edge corresponding to the target object includes: Vector matching processing is performed on the object's occupational feature vector and the preset occupational feature vector of each edge to obtain a vector matching result; the vector matching result is used to indicate the degree of matching between the preset occupational feature vector of each edge and the object's occupational feature vector; The word segmentation of the object's occupational features and the preset occupational features of each edge are subjected to word segmentation matching processing to obtain word segmentation matching results; the word segmentation matching results are used to indicate the degree of matching between the preset occupational features of each edge and the object's occupational features. Based on the vector matching result and the word segmentation matching result, at least one target matching edge is selected from multiple edges of the preset node relationship graph.

3. The method according to claim 2, characterized in that, The vector matching result includes vector matching data corresponding to each edge, and the word segmentation matching result includes word segmentation matching data corresponding to each edge; the method further includes: Get the preset vector matching weights corresponding to the vector matching dimension and the preset word segmentation matching weights corresponding to the word segmentation matching dimension; The step of selecting at least one target matching edge from multiple edges of the preset node relationship graph based on the vector matching result and the word segmentation matching result includes: Based on the vector matching results, the multiple edges are sorted in descending order of the corresponding vector matching data to obtain a sequence of vector candidate edges; Based on the word segmentation matching results, the multiple edges are sorted in descending order of the corresponding word segmentation matching data to obtain a sequence of candidate edges for word segmentation. Based on the preset vector matching weight and the preset word segmentation matching weight, at least one target matching edge is determined from the vector candidate edge sequence and the word segmentation candidate edge sequence.

4. The method according to claim 3, characterized in that, The step of determining the at least one target matching edge from the sequence of candidate vector edges and the sequence of candidate word edges based on the preset vector matching weight and the preset word segmentation matching weight includes: Based on the preset vector matching weight and the preset word segmentation matching weight, a first matching dimension and a second matching dimension are determined from the vector matching dimension and the word segmentation matching dimension; the first matching dimension is the matching dimension corresponding to the largest matching weight among the preset vector matching weight and the preset word segmentation matching weight; the second matching dimension is the matching dimension other than the first matching dimension among the vector matching dimension and the word segmentation matching dimension. Based on the preset vector matching weight and the preset word segmentation matching weight, determine the number of first matching edges corresponding to the first matching dimension and the number of second matching edges corresponding to the second matching dimension. From the vector candidate edge sequence and the word segmentation candidate edge sequence, determine the first candidate edge sequence corresponding to the first matching dimension and the second candidate edge sequence corresponding to the second matching dimension; The number of first matching edges in the first candidate edge sequence is taken as the first matching edge corresponding to the first matching dimension. The number of the first two matching edges in the second candidate edge sequence are taken as the second matching edges corresponding to the second matching dimension. The first matching edge and the second matching edge are used as the at least one target matching edge.

5. The method according to claim 4, characterized in that, The method further includes: If there are duplicate edges in the first matching edge and the second matching edge, the second candidate edge sequence is deduplicated based on the first matching edge to obtain the deduplicated second candidate edge sequence. The step of taking the number of edges that are the first few of the second matching edges in the second candidate edge sequence as the second matching edges corresponding to the second matching dimension includes: The number of edges that are the first two matching edges in the deduplicated second candidate edge sequence are taken as the second matching edges.

6. The method according to claim 1, characterized in that, The step of extracting features from the occupational description information of the object to obtain the occupational feature information corresponding to the target object includes: The occupational description information of the object is input into a preset occupational analysis model for occupational analysis to obtain at least one occupational information of the target object and at least one industry information of the target object. Feature extraction is performed on the occupational information and industry information of the at least one target object to obtain the occupational feature information of the object.

7. The method according to claim 6, characterized in that, The object occupational feature information includes an object occupational feature vector and object occupational feature word segmentation; the step of extracting features from the at least one target object occupational information and the at least one target object industry information to obtain the object occupational feature information includes: The occupational information and industry information of the at least one target object are spliced ​​together to obtain spliced ​​occupational information; The concatenated occupational information is segmented to obtain the occupational feature words of the object; The spliced ​​occupational information is vectorized to obtain the occupational feature vector of the object.

8. The method according to claim 1, characterized in that, The preset node relationship diagram is obtained in the following ways: Obtain the occupational information of multiple preset insurance products corresponding to the multiple preset product occupational information and the industry information of the preset product corresponding to each preset product occupational information; Any preset product occupation information is used to indicate the occupation information applicable to the corresponding preset insurance product, and the preset product industry information corresponding to any preset product occupation information is the industry information to which the preset product occupation information belongs; The multiple preset product occupation information is deduplicated to obtain the multiple preset occupation information; Using the multiple preset insurance products as product nodes, the multiple preset occupational information as occupational nodes, and the association between any preset insurance product and any preset occupational information as edges, construct the preset node relationship graph; Based on the occupational nodes connected to each edge, determine the target product occupational information corresponding to each edge; Based on the target product occupation information corresponding to each edge and the preset product industry information, the target product industry information corresponding to each edge is determined; Feature extraction is performed on the target product occupation information and the target product industry information corresponding to each edge to obtain the preset occupation feature information corresponding to each edge; The preset occupational characteristic information corresponding to each edge is written into each edge as an edge attribute.

9. The method according to claim 8, characterized in that, The method further includes: Upon receiving a product node update instruction for the preset node relationship diagram, acquire at least one updated product occupation information corresponding to the updated insurance product and updated product industry information corresponding to each updated product occupation information. Create an update product node corresponding to the updated insurance product in the preset node relationship graph; Based on the at least one updated product occupation information, at least one associated occupation node corresponding to the updated insurance product is determined from multiple occupation nodes in the preset node relationship graph; In the preset node relationship graph, an update edge is generated between each associated occupation node and the updated product node; Feature extraction is performed on the preset occupational information corresponding to each associated occupational node and the updated product industry information corresponding to each associated occupational node to obtain the updated occupational feature information corresponding to each associated occupational node. The updated occupational feature information corresponding to each associated occupational node is used as the edge attribute of the updated edge corresponding to each associated occupational node, and written into the updated edge corresponding to each associated occupational node.

10. The method according to any one of claims 1-9, characterized in that, The method further includes: Based on the at least one target matching edge and the preset node relationship graph, a target relationship subgraph corresponding to the target object is determined; the target relationship subgraph includes the at least one target matching edge and the nodes connected to the at least one target matching edge; The target relationship subgraph is input into a preset natural language model for inductive and summarizing processing to obtain product matching description information corresponding to the target object.

11. An insurance product matching device, characterized in that, The device includes: The first information acquisition module is used to acquire the object occupation description information corresponding to the target object and the preset node relationship diagram corresponding to multiple preset insurance products; the preset node relationship diagram is constructed with the multiple preset insurance products and multiple preset occupation information as nodes, and the association relationship between any preset insurance product and any preset occupation information as edges; each edge in the preset node relationship diagram includes preset occupation feature information, and the preset occupation feature information of each edge represents the preset industry information to which the preset occupation information corresponding to the product node connected to each edge belongs in the preset occupation information of the occupation node connected to each edge and the industry information corresponding to the occupation node; The first feature extraction module is used to extract features from the object's occupational description information to obtain the object's occupational feature information corresponding to the target object; the object's occupational feature information represents the object's occupational information and the industry information in which the object's occupational information is located. The matching processing module is used to perform matching processing on each edge in the preset node relationship graph based on the occupational feature information of the object, so as to obtain at least one target matching edge corresponding to the target object; The target product determination module is used to determine at least one target insurance product corresponding to the target object based on the target product node corresponding to each of the at least one target matching edge.

12. An electronic device, characterized in that, include: processor; Memory used to store computer programs; The processor is configured to execute the computer program to implement the insurance product matching method according to any one of claims 1 to 10.

Images