Method and device for determining target object, storage medium and electronic equipment

Abstract

The embodiment of the invention provides a method and device for determining a target object, a storage medium and electronic equipment, and the method comprises the steps: marking a first entity in a constructed knowledge graph, and recognizing an entity having an incidence relation with the first entity from the knowledge graph, and obtaining a first group structure; expanding the first group structure according to the N relation types to obtain a second group structure; and determining a target object from the second group structure. According to the method and the device, the problem that the potential object cannot be accurately determined in the prior art is solved, and the effect of accurately determining the potential object is achieved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computers, and more specifically, to a method and device for determining a target object, a storage medium, and an electronic device. Background Art

[0002] In related technologies, the methods for determining objects mostly rely on customer historical behavior data and use traditional machine learning or deep learning algorithms to build recommendation models. For example, some companies use collaborative filtering technology to determine the objects to be recommended. Or they classify objects through models such as linear regression and decision trees. However, these technologies can only rely on historical data and are difficult to mine complex potential object groups, especially in multi-level relationship mining and customer development. Summary of the invention

[0003] The embodiments of the present application provide a method and device for determining a target object, a storage medium, and an electronic device, so as to at least solve the problem that potential objects cannot be accurately determined in the related art.

[0004] According to one embodiment of the present application, a method for determining a target object is provided, comprising: marking a first entity in a constructed knowledge graph, and identifying an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, wherein the knowledge graph includes multiple entities corresponding to multiple objects for a target task, and the relationship between the multiple entities, and the first object corresponding to the first entity is an entity that meets a preset condition; expanding the first group structure according to N relationship types to obtain a second group structure, wherein the entities included in the second group structure all have an association relationship with the first entity; determining the target object from the second group structure, wherein the target object is an object whose association degree with the first object is greater than a preset threshold.

[0005] In an exemplary embodiment, before marking a first entity in a constructed knowledge graph and identifying an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, the method further includes: obtaining task information of the target task, wherein the task information is unstructured text information; inputting the task information into a target network to obtain a plurality of entities output by the target network and connection relationships between the plurality of entities, wherein the target network is a network obtained by iteratively training an initial network using a plurality of sets of training texts, wherein the initial network includes a feature extraction subnetwork, a recognition subnetwork, a first classification subnetwork, and a second classification subnetwork, and for each set of the training texts, the feature extraction subnetwork is used to identify the entity that has an association relationship with the first entity. The extraction subnetwork is used to extract information of sample tasks and information of sample objects that perform the sample tasks from the training text; the recognition subnetwork is used to identify multiple sample entities and execution data of each sample entity performing the sample task from the information of the sample tasks and the information of the sample objects; the first classification subnetwork is used to classify the multiple sample entities based on the execution data of the multiple sample entities, and classify the relationships between the multiple sample entities; the second classification subnetwork is used to classify the connection relationships between the multiple sample entities based on the categories of the multiple sample entities and the relationships between the multiple sample entities; and the knowledge graph is constructed using the multiple entities and the connection relationships between the multiple entities.

[0006] In an exemplary embodiment, a first entity is marked in a constructed knowledge graph, and an entity that has an association relationship with the first entity is identified from the knowledge graph to obtain a first group structure, including: obtaining business analysis data of a plurality of the objects, wherein the business analysis data includes at least one of the following: business demand data, historical business processing data, and market data; searching for the first entity from the knowledge graph according to the business analysis data; matching the first entity with a node in the knowledge graph to obtain a first node; setting label information of the first node to mark the first entity in the knowledge graph; generating a query instruction according to entity features of the first entity; searching the knowledge graph according to the query instruction for an entity that has an association relationship with the first entity, wherein the association relationship includes at least one of the following: a supply chain relationship, a financial relationship, a customer relationship, a technical relationship, an organizational structure relationship, a human resource relationship, and an industry relationship for the target task; constructing a group structure between the first entity and entities that have an association relationship with the first entity to obtain the first group structure.

[0007] In an exemplary embodiment, the first group structure is expanded according to N relationship types to obtain a second group structure, including: obtaining N relationship types according to the processing type of the target task processed by the first object; identifying a first other entity directly associated with the entity in the first group structure through the N relationship types, and searching for a second other entity indirectly associated with the entity in the first group structure; scoring the first other entity and the second other entity respectively to obtain multiple scores, wherein the scores are used to indicate the importance of the first other entity or the second other entity to the first group structure, and the scores are determined based on at least one of the following information between the first other entity or the second other entity and the first entity: strength of relationship, frequency of interaction, and time of interaction; filtering out a target entity based on the multiple scores, and adding the target entity to the first group structure to obtain the second group structure.

[0008] In an exemplary embodiment, determining the target object from the second group structure includes: determining entity characteristic information of a second entity in the second group structure, wherein the second entity is any entity in the second group structure, and the entity characteristic information includes at least one of the following: supply characteristics, financial characteristics, technical characteristics, and market performance characteristics; determining, based on the entity characteristic information, time association strength information with a third other entity in the second group structure, wherein the association strength information is used to indicate the degree of association between the second entity and the third other entity when processing the target task; and when the association strength information matches preset association strength information, determining the object corresponding to the second entity as the target object.

[0009] In an exemplary embodiment, after determining the target object from the second group structure, the method further includes: locating the category of the target object in the process of processing the target task according to the object characteristics, geographic location and business processing information of the target object to obtain the target category; generating target recommendation information according to the target category, wherein the target recommendation information includes at least one of the following: recommendation time, recommendation reason, recommended product, recommended process, and recommendation scope; and using the target recommendation information to recommend a target product to the target object.

[0010] In an exemplary embodiment, after using the target recommendation information to recommend a target product to the target object, the method further includes: obtaining a response of the target object to the recommendation instruction to obtain a response result, wherein the recommendation instruction is an instruction generated based on the target recommendation information; and using the response result to adjust the association relationship between the target entity corresponding to the target object and the first entity in the second group structure.

[0011] According to another embodiment of the present application, a device for determining a target object is provided, including: a first marking module, used to mark a first entity in a constructed knowledge graph, and identify an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, wherein the knowledge graph includes multiple entities corresponding to multiple objects for a target task, and the relationship between the multiple entities, and the first object corresponding to the first entity is an entity that meets a preset condition; a first expansion module, used to expand the first group structure according to N relationship types to obtain a second group structure, wherein the entities included in the second group structure all have an association relationship with the first entity; a first determination module, used to determine the target object from the second group structure, wherein the target object is an object whose association degree with the first object is greater than a preset threshold.

[0012] According to another embodiment of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0013] According to another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when run.

[0014] According to another embodiment of the present application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0015] Through this application, by marking the first entity in the constructed knowledge graph and identifying the entity with an associated relationship with the first entity from the knowledge graph, a first group structure is obtained, and the first group structure is expanded according to N relationship types to obtain a second group structure; the target object is determined from the second group structure, wherein the target object is an object whose degree of association with the first object is greater than a preset threshold. A dynamic group expansion mechanism is provided, which can continuously adjust and expand potential objects according to changes and needs. Therefore, the problem of not being able to accurately determine potential objects can be solved, and the effect of accurately determining potential objects can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a hardware structure block diagram of a server device of a method for determining a target object according to an embodiment of the present application;

[0017] Figure 2 is a flowchart of a method for determining a target object according to an embodiment of the present application;

[0018] Figure 3 is a flow chart of identifying potential customers according to an embodiment of the present application;

[0019] Figure 4 It is a structural block diagram of a device for determining a target object according to an embodiment of the present application. DETAILED DESCRIPTION

[0020] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0021] It should be noted that the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0022] The method embodiments provided in the embodiments of the present application can be executed in a server device or a similar computing device. Taking running on a server device as an example, Figure 1 1 is a hardware structure block diagram of a server device of a method for determining a target object in an embodiment of the present application. Figure 1 As shown, the server device may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the server device may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above server device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0023] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for determining the target object in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the server device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0024] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the server device. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0025] In this embodiment, a method for determining a target object is provided. Figure 2 is a flow chart of a method for determining a target object according to an embodiment of the present application. Figure 2 As shown, the process includes the following steps:

[0026] Step S202, marking a first entity in the constructed knowledge graph, and identifying an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, wherein the knowledge graph includes multiple entities corresponding to multiple objects for a target task, and relationships between the multiple entities, and the first object corresponding to the first entity is an entity that meets a preset condition;

[0027] Step S204: Expand the first group structure according to N relationship types to obtain a second group structure, wherein all entities included in the second group structure have an association relationship with the first entity;

[0028] Step S206: determining a target object from the second group structure, wherein the target object is an object whose correlation with the first object is greater than a preset threshold.

[0029] Optionally, the knowledge graph in this embodiment is a data structure that graphically represents entities (e.g., customers, companies, products, etc.) and the relationships between them, and can intuitively display the correlation between entities and the complexity of the network structure.

[0030] In the knowledge graph, the first entity is the known or marked entity with specific importance or that meets the preset conditions. It is the starting point for mining potential customers or expanding the community.

[0031] By analyzing the relationship between the first entity and other entities in the knowledge graph, a set of entities directly or indirectly associated with the first entity is screened out to form an initial group structure, namely, the first group structure.

[0032] The N relationship types are used to expand the association relationships of multiple dimensions of the group structure, such as customer relationships, financial relationships, supply chain relationships, technical cooperation, etc. These relationship types enable a more comprehensive exploration of the connections between entities.

[0033] The second group structure is a group structure containing more related entities obtained by further expanding the entity set through query and analysis of N relationship types on the basis of the first group structure.

[0034] The target object is an entity selected from the second group structure, whose correlation with the first object (entity) reaches or exceeds the preset threshold. It is regarded as a potential high-value target and is the focus of subsequent business development or customer follow-up.

[0035] For example, in the scenario of expanding the high-value customer community, suppose a financial institution wants to identify and expand its high-value customer base to optimize its service strategy for high-end customers. The following steps are included:

[0036] Step S302, collect customer information from multiple sources such as internal CRM systems, public financial reports, social media, etc., use NLP technology to identify entities (such as customers, products, services) and relationships (such as purchases, investments, recommendations), and build a knowledge graph. Based on historical transaction records and customer ratings, mark all high-value customers as the first entity. These customers usually have high transaction volumes, stable investment records, and good credit scores.

[0037] Step S304, N relationship type definitions: Define N relationship types, such as customer referral, joint investment, industry connection, etc. Perform queries based on these relationship types to identify other customers directly or indirectly associated with the first entity (high-value customer). For example, find entities connected to the first entity through customer referral relationships, which may also have high potential value.

[0038] Step S306, scoring the entities extended by the N relationship types, and screening out entities with a high degree of association with the first entity based on the scoring results and a preset association degree threshold. These entities are target objects, and may include customers with similar financial conditions, and potential high-value customers indirectly associated through joint investments or industry connections.

[0039] Financial institutions can develop personalized service plans for target customers, such as exclusive financial services, invitations to high-value customer events, etc., to further attract and develop these potential high-value customers.

[0040] For another example, in the scenario of building a technology cooperation network, a technology company hopes to expand its technology cooperation partner network to promote technological innovation and business expansion.

[0041] Step S402, mark the company's existing important technical partners as the first entity. These partners may have close cooperation with the company in the fields of core product development, patent sharing, etc. Collect data such as technical documents, patent databases, industry reports, etc., identify entities (such as companies, technologies, products) and relationships (such as technology licensing, patent cross-licensing, joint R&D projects), and build a technical cooperation knowledge map.

[0042] Step S404, define N types of relationship types related to technical cooperation, such as patent sharing, joint R&D projects, technology licensing, etc., perform queries on these relationship types, identify other technology companies or research institutions directly or indirectly related to the first entity, and form a second group structure.

[0043] Step S406: Analyze the technical capabilities, market influence, cooperation willingness and other factors of the entities in the second group structure, and select entities with a degree of association with the first entity that is higher than a preset threshold as target objects. These entities may have complementary technologies or have outstanding performance in specific fields and are potential technical partners. Technology companies can contact these target objects to explore the possibility of technical cooperation, such as jointly developing new products, cross-patent licensing, and formulating technical standards, so as to enhance their technological innovation capabilities and market position in the industry.

[0044] In the above embodiments, it can be seen that the method of this embodiment can flexibly mine potential targets associated with specific entities from the knowledge graph according to different business needs and goals, and accurately determine the target objects through dynamic community expansion and threshold screening, thereby providing strong data support for business decision-making and expansion.

[0045] The execution subject of the above steps may be a terminal, a server, a specific processor provided in the terminal or the server, or a processor or processing device provided relatively independently from the terminal or the server, but is not limited thereto.

[0046] Through the above steps, since the first entity is marked in the constructed knowledge graph, and the entity with an associated relationship with the first entity is identified from the knowledge graph, a first group structure is obtained, and the first group structure is expanded according to N relationship types to obtain a second group structure; the target object is determined from the second group structure, wherein the target object is an object whose degree of association with the first object is greater than a preset threshold. A dynamic group expansion mechanism is provided, which can continuously adjust and expand potential objects according to changes and needs. Therefore, the problem of being unable to accurately determine potential objects can be solved, and the effect of accurately determining potential objects can be achieved.

[0047] In an exemplary embodiment, before marking a first entity in a constructed knowledge graph and identifying an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, the method further includes: obtaining task information of the target task, wherein the task information is unstructured text information; inputting the task information into a target network to obtain a plurality of entities output by the target network and connection relationships between the plurality of entities, wherein the target network is a network obtained by iteratively training an initial network using a plurality of sets of training texts, wherein the initial network includes a feature extraction subnetwork, a recognition subnetwork, a first classification subnetwork, and a second classification subnetwork, and for each set of the training texts, the feature extraction subnetwork is used to identify the entity that has an association relationship with the first entity. The extraction subnetwork is used to extract information of sample tasks and information of sample objects that perform the sample tasks from the training text; the recognition subnetwork is used to identify multiple sample entities and execution data of each sample entity performing the sample task from the information of the sample tasks and the information of the sample objects; the first classification subnetwork is used to classify the multiple sample entities based on the execution data of the multiple sample entities, and classify the relationships between the multiple sample entities; the second classification subnetwork is used to classify the connection relationships between the multiple sample entities based on the categories of the multiple sample entities and the relationships between the multiple sample entities; and the knowledge graph is constructed using the multiple entities and the connection relationships between the multiple entities.

[0048] Optionally, the target task in this embodiment refers to a specific business goal or research task, such as potential customer identification, technical cooperation network construction, etc. The target task information in this embodiment is unstructured text information and needs to be parsed and understood through a deep learning model.

[0049] Task information refers to unstructured text information related to the target task, which may include business requirements, market trends, partner information, etc. It is a description or background information of the target task.

[0050] Optionally, the target network is a trained deep learning model for identifying entities, relationships between entities, and entity information for performing specific tasks from unstructured text information. The network consists of a feature extraction subnetwork, a recognition subnetwork, a first classification subnetwork, and a second classification subnetwork, and is capable of processing complex natural language descriptions, extracting key information, and constructing a knowledge graph.

[0051] The feature extraction subnetwork is used to extract features from the input text, including the descriptive features of the task and the entity features of the task execution, providing a basis for subsequent entity recognition and relationship classification.

[0052] The recognition subnetwork is used to identify the entities mentioned in the text and the detailed data of the tasks performed by these entities, such as the time, place, and method of execution.

[0053] The first classification sub-network is used for classification based on the execution data of the entity, and is used to determine the type of the entity and the relationship category between the entities, such as customers, suppliers, partners, etc.

[0054] The second classification subnetwork is used to classify the connection relationships between entities based on the classification information of the entities and the relationship categories between the entities, that is, how to connect these entities in the knowledge graph to form a network structure between the entities.

[0055] For example, in the scenario of potential customer identification, the task background is that an e-commerce platform wants to identify potential high-value customers in order to provide more personalized services and product recommendations. The specific steps include:

[0056] Step 1: Obtain task information: Collect unstructured text information about high-value customer identification, including historical sales records, customer feedback, market reports, etc., which may contain descriptions of high-value customers, their usual purchasing behaviors and preferences, etc.

[0057] Step 2: Build and train the target network: Initial network design: Design the initial network, including the feature extraction subnetwork, recognition subnetwork, first classification subnetwork and second classification subnetwork. Training text preparation: Prepare multiple sets of training texts, each of which contains a sample task (such as identifying high-value customers) and a set of sample objects (such as historical customer records). Network training: Input the training text into the feature extraction subnetwork to extract the features of the task and the object; the recognition subnetwork identifies the entities (such as customers, products) and execution data (such as purchase history) in the text; the first classification subnetwork classifies the execution data of the entity to determine the entity type (such as high-value customers) and the relationship type between entities (such as purchase); the second classification subnetwork classifies the connection relationship between entities according to the entity classification and relationship type, and constructs an entity network.

[0058] Step 3: Identify entities and build a knowledge graph: Entity identification: Input the target task information into the trained target network, and use the feature extraction subnetwork and the recognition subnetwork to identify entities and their execution data from the text. Entity classification and relationship identification: The first classification subnetwork classifies the identified entities and determines their roles (such as high-value customers, ordinary customers, suppliers, etc.); the second classification subnetwork identifies the relationship types between entities (such as purchase, recommendation, attention, etc.). Build a knowledge graph: Based on the classified entities and the relationship types between entities, build a complete knowledge graph containing high-value customers, ordinary customers, products, suppliers and other entities and their relationships.

[0059] Step 4: Mark the first entity and potential customer mining: Mark the first entity: In the knowledge graph, based on historical sales records or customer value scores, mark known high-value customers as the first entity. Potential customer identification: Identify other entities that have an association relationship with the first entity from the knowledge graph to form a first group structure; through community expansion algorithms, such as N relationship type queries, further expand the group to obtain a second group structure; finally, screen out the target object, that is, potential high-value customers, based on the degree of association. Subsequent actions: Based on the identified potential customers, the platform can formulate personalized service strategies, such as providing exclusive discounts, customized product recommendations, etc., to further attract and convert these potential customers.

[0060] Through the above examples, we can clearly see how to start from unstructured task information, identify entities, understand tasks, build knowledge graphs through deep learning models, and finally achieve target entity recognition and community expansion based on knowledge graphs. This process fully utilizes the advantages of natural language processing and graph technology, and provides a strong technical foundation for business intelligence and decision support.

[0061] In an exemplary embodiment, a first entity is marked in a constructed knowledge graph, and an entity that has an association relationship with the first entity is identified from the knowledge graph to obtain a first group structure, including: obtaining business analysis data of a plurality of the objects, wherein the business analysis data includes at least one of the following: business demand data, historical business processing data, and market data; searching for the first entity from the knowledge graph according to the business analysis data; matching the first entity with a node in the knowledge graph to obtain a first node; setting label information of the first node to mark the first entity in the knowledge graph; generating a query instruction according to entity features of the first entity; searching the knowledge graph according to the query instruction for an entity that has an association relationship with the first entity, wherein the association relationship includes at least one of the following: a supply chain relationship, a financial relationship, a customer relationship, a technical relationship, an organizational structure relationship, a human resource relationship, and an industry relationship for the target task; constructing a group structure between the first entity and entities that have an association relationship with the first entity to obtain the first group structure.

[0062] Optionally, the business analysis data in this embodiment refers to data related to the operation, market, finance and other businesses of an enterprise or organization, which is used to analyze and understand business needs, historical business processing and market dynamics, including business demand data (such as market research reports, business development plans), historical business processing data (such as transaction records, customer feedback), market data (such as industry reports, competitor analysis), etc.

[0063] Optionally, in the knowledge graph, the first entity that is marked or selected as the starting point of the analysis according to business needs or preset conditions is usually a high-value customer, key partner, industry leader, etc., with important business or strategic value.

[0064] Optionally, in the knowledge graph, the graph node corresponding to the first entity is found by matching the features of the first entity with the features of the node in the knowledge graph, and a label is set for subsequent community discovery and expansion.

[0065] Optionally, entity features are attributes or parameters used to describe the properties and behaviors of an entity, such as a company's financial status, a customer's purchase history, a product's technical characteristics, etc. Entity features are the key basis for building and querying knowledge graphs, and can help identify relationships between entities.

[0066] Optionally, association relationships refer to various connections between entities in the knowledge graph, including supply chain relationships (connections between suppliers and purchasers), financial relationships (capital flow, investment, loans, etc.), customer relationships (purchase, use, recommendation, etc.), technical relationships (patent cross-licensing, technical cooperation), organizational structure relationships (departments, subsidiaries, etc.), human resources relationships (employees, management, consultants, etc.), industry relationships (competitors, industry standard setters, etc.). These relationships are used to build network structures between entities and support community discovery and extended analysis.

[0067] For example, in the scenario of supply chain optimization and supplier identification, a large manufacturing company hopes to optimize its supply chain network and identify potential high-quality suppliers to reduce costs and improve product quality. The specific steps include:

[0068] Step 1: Obtain business analysis data: Collect business analysis data such as the company's supply chain management reports, historical procurement records, supplier evaluation data, and industry supply chain analysis reports to analyze the pain points of the current supply chain and identify potential optimization opportunities.

[0069] Step 2: Mark the first entity: Based on the business demand data, identify the key products or components in the company's current supply chain, find the existing suppliers of these products or components, and mark them as the first entity.

[0070] Step 3: Construct the first group structure: Match the first entity with the nodes in the knowledge graph: Search for the corresponding first node in the knowledge graph through entity features (such as supplier name, product description), and set label information, such as "key supplier". Generate query instructions: Generate query instructions based on the business needs of the first entity, such as product type, technical standards, procurement volume, etc., to search for potential suppliers in the knowledge graph. Search for related entities: Execute query instructions to identify other entities that have supply chain relationships, technical relationships or industry relationships with the first entity (key supplier). Construct a group structure: Construct the first entity and the searched related entities into a group structure to form a first group structure, which may include direct suppliers, indirect suppliers, and entities that have technical cooperation with current suppliers.

[0071] Step 4: Expansion and optimization: Based on the first group structure, further explore entities that have financial dealings, technical cooperation, etc. with existing suppliers, and form a second group structure through the community expansion algorithm. From the second group structure, the company can screen out potential suppliers with high association with key suppliers, good evaluation and meeting product requirements as target objects for supply chain optimization. Subsequent actions: The company can negotiate with these target objects, evaluate their supply capabilities and product quality, and finally choose to cooperate to achieve supply chain optimization and cost reduction.

[0072] Through the above embodiments, we can see the application of this technical solution in the supply chain optimization scenario. Through the construction of knowledge graphs, entity labeling, association relationship query and group structure construction, it achieves the goal of accurately mining potential suppliers from existing suppliers, and provides strong technical support for manufacturing companies to optimize supply chains and improve efficiency.

[0073] In an exemplary embodiment, the first group structure is expanded according to N relationship types to obtain a second group structure, including: obtaining N relationship types according to the processing type of the target task processed by the first object; identifying a first other entity directly associated with the entity in the first group structure through the N relationship types, and searching for a second other entity indirectly associated with the entity in the first group structure; scoring the first other entity and the second other entity respectively to obtain multiple scores, wherein the scores are used to indicate the importance of the first other entity or the second other entity to the first group structure, and the scores are determined based on at least one of the following information between the first other entity or the second other entity and the first entity: strength of relationship, frequency of interaction, and time of interaction; filtering out a target entity based on the multiple scores, and adding the target entity to the first group structure to obtain the second group structure.

[0074] Optionally, in this embodiment, the relationship type refers to various types of connections or interactions of different natures that may exist between entities, including but not limited to supply chain relationships, financial relationships, customer relationships, technical relationships, organizational structure relationships, human resources relationships, and industry relationships, etc. Each relationship type represents a specific way of association between entities, for example, a supply chain relationship may refer to the exchange of goods and services between a supplier and a purchaser.

[0075] Optionally, in this embodiment, the first other entity refers to an entity connected to the first entity (i.e., the marked entity) through a direct association in the first group structure. These entities are directly related to the first entity and may share business needs, supply chain connections, or technical cooperation, etc.

[0076] Optionally, in this embodiment, the second other entity refers to an entity indirectly associated with the first entity, which may establish a connection with the first entity through an intermediate entity. For example, in a supply chain relationship, the supplier of the first entity (i.e., the supplier's upstream supplier) belongs to the second other entity.

[0077] Optionally, in this embodiment, the score is used to quantify the importance of the first other entity or the second other entity to the first group structure, and the score reflects factors such as the strength of the relationship between the entities, the frequency and time of interaction, etc. The scoring mechanism is the key to evaluating the contribution of an entity to the group structure, and is crucial for community expansion and identification of target objects.

[0078] For example, in the scenario of potential customer mining in B2B customer relationship management, a B2B enterprise hopes to identify potential high-value customers by optimizing customer relationship management to promote business growth and market expansion. The specific steps include:

[0079] Step 1: Determine N relationship types: Assume that the target task is B2B sales. Based on the processing type of the target task for the first object (i.e., the known high-value customer), N relationship types are determined, including customer recommendation relationship, supply chain relationship, and industry contact relationship. These relationship types will be used to expand the first group structure and find potential customers directly or indirectly related to high-value customers.

[0080] Step 2: Identify the first other entity and the second other entity: Identify directly related first other entities: Perform queries for N relationship types to identify entities directly related to the first entity (high-value customers), such as customers directly recommended through customer recommendation relationships. Search for indirectly related second other entities: Use community expansion algorithms, such as queries through supply chain relationships and industry connection relationships to identify entities that are indirectly related to the first entity through intermediate entities (such as common suppliers or industry organizations).

[0081] Step 3: Scoring and target entity screening: Scoring system design: Set scoring rules for each relationship type (customer referral, supply chain, industry connection). For example, for customer referral relationships, scoring can be based on the recommender's reputation, recommendation frequency and time; for supply chain relationships, scoring may be based on transaction volume, transaction frequency and transaction time. Entity scoring: Score the identified first other entity and second other entity to evaluate their importance to the first group structure. Target entity screening: Set a preset threshold based on the scoring results, and screen out entities with scores above the threshold as target entities. These entities are considered potential high-value customers that are closely associated with the first entity (high-value customers).

[0082] Step 4: Construction of the second group structure: Add the selected target entity to the first group structure, and form the second group structure together with the first entity and the first other entity. This group structure will include all directly and indirectly related entities, which may all be potential high-value customers. Based on the target entities in the second group structure, enterprises can formulate personalized strategies or customer development plans, such as targeted marketing, product customization, cooperation proposals, etc., to attract these potential high-value customers and promote business growth.

[0083] Through this embodiment, the rich information of the knowledge graph and the intelligent processing capabilities of deep learning are fully utilized to provide B2B companies with an efficient and accurate method for mining potential customers.

[0084] In an exemplary embodiment, determining the target object from the second group structure includes: determining entity characteristic information of a second entity in the second group structure, wherein the second entity is any entity in the second group structure, and the entity characteristic information includes at least one of the following: supply characteristics, financial characteristics, technical characteristics, and market performance characteristics; determining, based on the entity characteristic information, time association strength information with a third other entity in the second group structure, wherein the association strength information is used to indicate the degree of association between the second entity and the third other entity when processing the target task; and when the association strength information matches preset association strength information, determining the object corresponding to the second entity as the target object.

[0085] Optionally, in this embodiment, the second entity is an entity that is directly or indirectly associated with the first entity through N relationship types and added to the second group structure during the community expansion process. The second entity may represent a potential supplier, partner, customer or competitor, etc., depending on the nature of the target task and the community expansion strategy.

[0086] Optionally, in this embodiment, the entity characteristic information is used to describe a data set of specific attributes of the second entity, including supply characteristics (such as supply capacity, delivery time), financial characteristics (such as asset size, credit rating), technical characteristics (such as number of patents, R&D investment), market performance characteristics (such as market share, customer evaluation), etc. Entity characteristic information is used to evaluate the business capabilities and market position of the entity, and is an important basis for community expansion and target object identification.

[0087] Optionally, in this embodiment, the third other entity is other entity in the second group structure except the first entity and the second entity. They may be directly or indirectly associated with the second entity through different relationship types and are part of the community structure.

[0088] The strength of association information is used to quantify the degree of connection between the second entity and the third entity when processing the target task, reflecting their mutual dependence in the business process. The strength of association information may be generated based on indicators such as transaction volume, cooperation frequency, and strength of recommendation relationship between entities, and is used to evaluate the potential and importance of cooperation between entities in the community.

[0089] The preset association strength information is the association strength threshold used as a screening criterion in the process of community expansion and target object determination. It is used to distinguish whether the strength of the relationship between entities in the community meets the specific standards related to the target task, and helps determine which entities are more likely to become valuable target objects.

[0090] For example, in the scenario of supply chain optimization and partner identification, a manufacturing company is seeking to optimize its supply chain to improve product quality and reduce costs while looking for new business opportunities. Through community expansion, the company builds a second group structure that includes existing suppliers, potential suppliers, technical partners, etc. The specific steps include:

[0091] Step 1: Determine the entity characteristic information of the second entity: Analyze the supply capacity and stability of the second entity, such as delivery time, quality control, production capacity, etc., to obtain supply characteristics. Evaluate the financial status of the second entity, including asset size, cash flow, credit rating, etc., to ensure its financial health and cooperation potential, and obtain financial characteristics. Study the technical level and innovation ability of the second entity, such as the number of patents, R&D investment, technical team size, etc., to obtain technical characteristics. Examine the performance of the second entity in the market, including market share, customer satisfaction, brand influence, etc., to determine its industry position and market acceptance, and obtain market performance characteristics.

[0092] Step 2: Determine the strength of association information: Based on the entity characteristic information of the second entity and the degree of association between the second entity and the third other entity (such as existing suppliers, other potential suppliers) when processing the target task (such as supply chain optimization, product development), the enterprise calculates the strength of association information between entities. For example, for supply chain optimization, the strength of association may be quantified based on data such as the supply entity's delivery record, product quality feedback, and financial cooperation stability.

[0093] Step 3: Determine the target object: Set the preset association strength information as the threshold, and determine the objects corresponding to the second entity that has a strong association strength with the third other entity in the second group structure as the target object. The target objects may include suppliers or partners with high-quality supply chains, good financial conditions, high technical strength or excellent market performance, who are more likely to become key players in supply chain optimization or new product development. Based on the determination of the target objects, enterprises can further evaluate and approach these entities, formulate supply chain optimization strategies or new product development plans, introduce more high-quality partners, and improve overall business efficiency and market competitiveness.

[0094] Through this embodiment, by quantitatively analyzing the entity feature information and association strength information of the second entity, it helps enterprises to accurately identify high-value objects that are closely related to the target task, and provides powerful decision-making support for supply chain management and business cooperation.

[0095] In an exemplary embodiment, after determining the target object from the second group structure, the method further includes: locating the category of the target object in the process of processing the target task according to the object characteristics, geographic location and business processing information of the target object to obtain the target category; generating target recommendation information according to the target category, wherein the target recommendation information includes at least one of the following: recommendation time, recommendation reason, recommended product, recommended process, and recommendation scope; and using the target recommendation information to recommend a target product to the target object.

[0096] Optionally, in this embodiment, the target object is a group of entities in the second group structure that are determined to have potential business value or are highly relevant to the target task through a series of analysis and screening processes. These entities may be customers, suppliers, partners or other key participants related to the enterprise's target tasks.

[0097] Object characteristics refer to specific attributes or indicators of the target object, such as enterprise size, industry type, technical capabilities, market performance, etc., which are used to fully understand the business capabilities and market position of the target object. Object characteristics are an important basis for classification and recommendation strategy formulation.

[0098] Geographic location is the physical location information of the target object, which is particularly important for business scenarios involving logistics, sales areas, market coverage, etc. Geographic location information can help companies understand the market potential and supply chain location advantages of the target object, and is one of the considerations when formulating business processing strategies.

[0099] Business processing information is data related to the target object’s business activities, including but not limited to transaction records, business needs, cooperation history, etc., which is used to evaluate the target object’s business activity and cooperation tendency. Business processing information is the key to determining the target object’s role and value in the business process.

[0100] The target category is based on the target object's object characteristics, geographical location and business processing information to locate the role or type of the target object in the process of processing the target task, such as "high-value potential customer", "high-quality supplier", "technical innovation partner", etc. The target category will guide the subsequent formulation of recommendation strategies.

[0101] Target recommendation information is a targeted recommendation based on the category of the target object, such as recommendation time (when to make a recommendation), recommendation reason (based on what business context to recommend), recommended product (specifically which products or services to recommend), recommendation process (recommended execution steps), and recommendation scope (recommended coverage). Target recommendation information is designed to optimize the business interaction between the enterprise and the target object, and improve the success rate of recommendation and customer satisfaction.

[0102] For example, in the scenario of customized product recommendation based on knowledge graph, a multinational e-commerce platform hopes to recommend customized products to potential high-value customers by analyzing global user communities to increase sales and customer satisfaction. The specific steps include:

[0103] Step 1: Determine the target audience: Through the community expansion algorithm, the platform has built a second group structure that includes global users, products, suppliers, etc. After analyzing the user's historical purchase behavior, search history and social network influence, a group of users with high consumption power, high activity and high social influence are identified as the target audience.

[0104] Step 2: Identify the target category of the target object: Investigate the target object's purchase history, average spending amount and purchase frequency to determine that it belongs to the category of "high spending power users". Combine the target object's geographical location information to understand the market trends and consumer preferences in the area where it is located, and it may be classified as a "high-value user in a specific market". By analyzing the target object's shopping cart saved products, browsing history and shopping preferences, determine its target category of "high interest potential customers".

[0105] Step 3: Generate target recommendation information: Based on the target object's "highly interested potential customers" category, the platform generates the following target recommendation information:

[0106] Recommendation time: Analyze the shopping habits and history of the target audience and make recommendations during new product launches or specific holiday promotions to increase conversion rates.

[0107] Reason for recommendation: Based on the target’s geographic location and local market trends, recommend products or services that meet the preferences of consumers in the target area, such as holiday-exclusive gifts and regional specialty products.

[0108] Recommended products: Based on the target object’s shopping preferences and historical purchase records, recommend highly rated and highly matched specific products, such as high-end electronic products, customized clothing, etc.

[0109] Recommendation process: Design a personalized recommendation process, including pushing customized content, providing exclusive customer service, and tracking feedback for subsequent optimization.

[0110] Recommendation scope: Considering the social influence of the target object, the recommendation scope can be expanded to the contacts in their social network to promote word-of-mouth communication and increase the exposure of the platform.

[0111] Step 4: Use the target recommendation information to recommend target products to target objects:

[0112] According to the generated target recommendation information, the platform pushes customized product recommendations to the target object. For example, during the launch of new products, high-end electronic products are recommended to high-value users in a specific market, with exclusive discounts and customer service support. In addition, through the target object's social network, the platform also recommends to its contacts, using the target object's influence to improve the recommendation effect.

[0113] Follow-up actions: Based on recommendation feedback, the platform continuously optimizes the knowledge graph and recommendation algorithm and adjusts the recommendation strategy to improve the success rate of customized product recommendations and customer satisfaction.

[0114] Through this embodiment, by analyzing the characteristics, geographic location and business information of the target object, locating the target category and generating customized recommendation information, accurate product recommendations are finally achieved, which promotes business growth and deepening of customer relationships.

[0115] In an exemplary embodiment, after using the target recommendation information to recommend a target product to the target object, the method further includes: obtaining a response of the target object to the recommendation instruction to obtain a response result, wherein the recommendation instruction is an instruction generated based on the target recommendation information; and using the response result to adjust the association relationship between the target entity corresponding to the target object and the first entity in the second group structure.

[0116] Optionally, in this embodiment, the response result refers to the target object's feedback or behavior result on the recommendation instruction, including but not limited to acceptance, rejection, feedback, purchase behavior, etc. The response result is an important basis for evaluating the recommendation effect and optimizing the community structure.

[0117] The recommendation instruction is a specific operation guide generated based on the target recommendation information, instructing the target object on how to receive and respond to the recommendation, such as email notification, phone invitation, product demonstration invitation, etc. The recommendation instruction is intended to promote business interaction between the target object and the first entity.

[0118] Adjusting the association relationship means that in the process of community discovery and expansion, according to the response results of the target object, the system will dynamically adjust the association strength or level between the target entity and the first entity in the knowledge graph. This may include increasing or decreasing the connection weight, updating entity feature information, etc., to more accurately reflect the actual business connection between entities.

[0119] Through this embodiment, by collecting and analyzing the response results of the target objects and dynamically adjusting the association relationships between entities in the community structure, refined management and optimization of potential customer relationships are achieved.

[0120] The present application is described in detail below with reference to specific embodiments:

[0121] This specific embodiment takes the identification of more potential high-end customers in a financial scenario as an example for explanation. Figure 3 As shown, the following steps are included:

[0122] S301, build knowledge graph:

[0123] Integrate internal customer information of financial institutions (such as transaction records, asset status, investment preferences), public industry data (such as wealth lists, financial news), social media data (such as customer online activities, comments), etc., to build a knowledge graph that includes entities such as customers, products, services, industry trends and their relationships.

[0124] S302, marking a first entity and identifying a first group structure:

[0125] According to the needs of wealth management business, existing high-end customers are marked as the first entity, and a first group structure including these customers is identified through a community discovery algorithm (such as the Louvain algorithm), which may include customers with similar financial status and investment preferences.

[0126] S303, expanding the second group structure:

[0127] Identify N relationship types: for example, financial relationship (common financial advisor), industry relationship (attending the same industry conference), customer relationship (common investment consulting), etc.

[0128] Identify the first other entity and the second other entity: By analyzing the high-end customers within the first group structure, identify other potential high-end customers directly or indirectly related to them, such as individuals or companies with similar investment records and asset sizes.

[0129] Scoring and screening: Score the identified first other entity and second other entity, and screen out target entities with a high degree of correlation with the first entity based on financial characteristics, investment behavior, and market performance characteristics.

[0130] S304, determine target objects and customized recommendations:

[0131] Positioning target categories: Analyze the target entity’s financial status, investment preferences, and geographic location information to determine its target category in wealth management business, such as “high-asset potential customers.”

[0132] Generate target recommendation information: Based on the target category, design customized financial service recommendation information, including recommendation time (such as market volatility period), recommended products (such as high-end financial products, private banking services), and recommendation process (such as making an appointment with a private financial advisor).

[0133] Recommendations and response results: Push customized recommendations to target entities and collect their response results, such as whether to make an appointment with a financial advisor, whether they show interest, etc.

[0134] S305, optimize the community structure: according to the response result of the target entity, dynamically adjust its association with the first entity in the second group structure. The target entity that responds positively increases its weight in the community and is regarded as a potential high-value customer; the entity that responds negatively reduces its weight and makes strategic adjustments to more accurately locate customer needs.

[0135] In the scenario of game player community management and in-game activity optimization, a game development company plans to optimize the player community structure to improve the participation in in-game activities and player satisfaction, while attracting more potential players, expanding the game's user base and increasing game revenue. The specific steps include:

[0136] Step 1: Build a player knowledge graph: Collect in-game player data (such as game time, consumption records, game achievements), player social media data (such as interactions and comments between players), and market research data (such as player preferences and game trends), and build a knowledge graph that includes entities such as players, games, and social relationships.

[0137] Step 2: Mark the first entity and identify the first group structure: According to the needs of in-game activity optimization, mark players with high activity and contribution value as the first entity, and use the community discovery algorithm to identify the initial group structure containing these players. This group may include players who often team up in the game and participate in activities together.

[0138] Step 3: Expand the second group structure: Determine N types of relationships: for example, shared game time, team frequency, in-game transactions, social network connections, etc. Identify the first other entity and the second other entity: By analyzing the behavior of active players in the community, identify other players who are directly or indirectly related to them, such as players who often team up together, players who follow the same game content, etc. Rating and screening: Score the identified entities and screen out the target players who are most likely to be interested in in-game activities based on in-game performance, activity participation, and social activity.

[0139] Step 4: Determine the target audience and activity recommendations: Identify the target category: Analyze the target player's game preferences, activity participation history, and social behavior, and determine their target category in in-game activity participation, such as "socially active players." Generate target recommendation information: Based on the target category, design customized activity recommendation information, including recommended time (such as large weekend events), recommended reasons (such as player birthdays), and recommended activities (such as team dungeon challenges, social gameplay). Recommendation and response results: Recommend customized activities to target players and collect their response results, such as whether to participate in the activity, activity feedback, etc.

[0140] Step 5: Optimize the player community structure: According to the response results of the target players, adjust their association in the second group structure. Players who actively participate in the activities will have their community weight increased and will be regarded as core players; players with low responses will have their weight reduced, and the reasons for their non-participation will be analyzed to optimize the activity design or recommendation strategy to attract a wider range of players to participate.

[0141] Through these two embodiments, the technical solution of this specific embodiment is not only applicable to high-end customer management in the financial field, but can also be applied to player community optimization in the gaming industry. Through intelligent community discovery, expansion and precise recommendation strategies, personalized business adjustment and efficiency improvement are achieved.

[0142] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0143] In the present embodiment, a device for determining a target object is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions thereof will not be repeated. As used below, the term "module" may implement a combination of software and / or hardware of a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and contemplated.

[0144] Figure 4 is a structural block diagram of a device for determining a target object according to an embodiment of the present application, such as Figure 4 As shown, the device comprises:

[0145] A first marking module 42 is used to mark a first entity in the constructed knowledge graph, and identify an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, wherein the knowledge graph includes a plurality of entities corresponding to a plurality of objects for a target task, and a relationship between the plurality of entities, and a first object corresponding to the first entity is an entity that meets a preset condition;

[0146] A first expansion module 44 is configured to expand the first group structure according to N relationship types to obtain a second group structure, wherein all entities included in the second group structure have an association relationship with the first entity;

[0147] The first determining module 46 is configured to determine a target object from the second group structure, wherein the target object is an object whose association degree with the first object is greater than a preset threshold.

[0148] In an exemplary embodiment, the above-mentioned device also includes: a first acquisition module, which is used to mark the first entity in the constructed knowledge graph, and identify the entity that has an association relationship with the first entity from the knowledge graph, and before obtaining the first group structure, obtain the task information of the target task, wherein the task information is unstructured text information; a first input module, which is used to input the task information into the target network, and obtain the connection relationship between the multiple entities output by the target network and the multiple entities, wherein the target network is a network obtained by iteratively training the initial network using multiple groups of training texts, and the initial network includes a feature extraction subnetwork, a recognition subnetwork, a first classification subnetwork and a second classification subnetwork, and for each group of the training texts, the The feature extraction subnetwork is used to extract information about sample tasks and information about sample objects that perform the sample tasks from the training text; the recognition subnetwork is used to identify multiple sample entities and execution data of each sample entity performing the sample task from the information about the sample tasks and the information about the sample objects; the first classification subnetwork is used to classify multiple sample entities based on the execution data of the multiple sample entities, and classify the relationships between the multiple sample entities; the second classification subnetwork is used to classify the connection relationships between the multiple sample entities based on the categories of the multiple sample entities and the relationships between the multiple sample entities; the first construction module is used to construct the knowledge graph using the multiple entities and the connection relationships between the multiple entities.

[0149] In an exemplary embodiment, the above-mentioned first marking module includes: a first acquisition unit, used to acquire business analysis data of multiple objects, wherein the business analysis data includes at least one of the following: business demand data, historical business processing data, and market data; a first search unit, used to search for the first entity from the knowledge graph according to the business analysis data; a first matching unit, used to match the first entity with a node in the knowledge graph to obtain a first node; a first setting unit, used to set label information of the first node to mark the first entity in the knowledge graph; a first generation unit, used to generate a query instruction according to the entity characteristics of the first entity; a first search unit, used to search in the knowledge graph according to the query instruction for an entity that has an associated relationship with the first entity, wherein the associated relationship includes at least one of the following: a supply chain relationship, a financial relationship, a customer relationship, a technical relationship, an organizational structure relationship, a human resources relationship, and an industry relationship for the target task; a first construction unit, used to construct a group structure between the first entity and an entity that has an associated relationship with the first entity to obtain the first group structure.

[0150] In an exemplary embodiment, the above-mentioned first expansion module includes: a first determination unit, used to obtain N relationship types according to the processing type of the target task processed by the first object; a first identification unit, used to identify the first other entity directly associated with the entity in the first group structure through the N relationship types, and to find the second other entity indirectly associated with the entity in the first group structure; a first scoring unit, used to score the first other entity and the second other entity respectively to obtain multiple scores, wherein the score is used to indicate the importance of the first other entity or the second other entity to the first group structure, and the score is determined based on at least one of the following information between the first other entity or the second other entity and the first entity: the strength of the relationship, the frequency of interaction, and the time of interaction; a first screening unit, used to screen out the target entity based on the multiple scores, and add the target entity to the first group structure to obtain the second group structure.

[0151] In an exemplary embodiment, the above-mentioned first determination module includes: a second determination unit, used to determine the entity characteristic information of the second entity in the second group structure, wherein the second entity is any entity in the second group structure, and the entity characteristic information includes at least one of the following: supply characteristics, financial characteristics, technical characteristics, and market performance characteristics; a third determination unit, used to determine the association strength information with the third other entity in the second group structure based on the entity characteristic information, wherein the association strength information is used to indicate the degree of association between the second entity and the third other entity when processing the target task; a fourth determination unit, used to determine the object corresponding to the second entity as the target object when the association strength information matches the preset association strength information.

[0152] In an exemplary embodiment, the above-mentioned device also includes: a first positioning module, which is used to locate the category of the target object in the process of processing the target task according to the object characteristics, geographical location and business processing information of the target object after determining the target object from the second group structure, so as to obtain the target category; a first generation module, which is used to generate target recommendation information according to the target category, wherein the target recommendation information includes at least one of the following: recommendation time, recommendation reason, recommended product, recommended process, and recommendation scope; and a first recommendation module, which is used to recommend a target product to the target object using the target recommendation information.

[0153] In an exemplary embodiment, the above-mentioned device also includes: a second acquisition module, which is used to obtain the target object's response to the recommendation instruction after recommending the target product to the target object using the target recommendation information, and obtain a response result, wherein the recommendation instruction is an instruction generated based on the target recommendation information; a first adjustment module, which is used to adjust the association relationship between the target entity corresponding to the target object and the first entity in the second group structure using the response result.

[0154] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0155] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0156] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0157] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0158] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0159] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any one of the above method embodiments are implemented.

[0160] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0161] The embodiments of the present application also provide a computer program, which includes computer instructions, which are stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps of any one of the above method embodiments.

[0162] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0163] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0164] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for determining a target object, characterized in that: include: Marking a first entity in the constructed knowledge graph, and identifying an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, wherein the knowledge graph includes multiple entities corresponding to multiple objects for a target task, and relationships between the multiple entities, and a first object corresponding to the first entity is an entity that meets a preset condition; Expanding the first group structure according to N relationship types to obtain a second group structure, wherein all entities included in the second group structure have an association relationship with the first entity; A target object is determined from the second group structure, wherein the target object is an object whose association degree with the first object is greater than a preset threshold.

2. The method according to claim 1, characterized in that Before marking the first entity in the constructed knowledge graph and identifying an entity having an association relationship with the first entity from the knowledge graph to obtain the first group structure, the method further includes: Acquire task information of the target task, wherein the task information is unstructured text information; Input the task information into the target network to obtain the multiple entities output by the target network and the connection relationships between the multiple entities, wherein the target network is a network obtained by iteratively training an initial network using multiple sets of training texts, and the initial network includes a feature extraction subnetwork, a recognition subnetwork, a first classification subnetwork and a second classification subnetwork. For each set of training texts, the feature extraction subnetwork is used to extract information about sample tasks and information about sample objects that perform the sample tasks from the training texts, the recognition subnetwork is used to identify multiple sample entities and execution data of each sample entity performing the sample tasks from the information about the sample tasks and the information about the sample objects, the first classification subnetwork is used to classify the multiple sample entities based on the execution data of the multiple sample entities, and classify the relationships between the multiple sample entities, and the second classification subnetwork is used to classify the connection relationships between the multiple sample entities based on the categories of the multiple sample entities and the relationships between the multiple sample entities; The knowledge graph is constructed using the multiple entities and the connection relationships between the multiple entities.

3. The method according to claim 1, characterized in that Marking a first entity in the constructed knowledge graph, and identifying an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, including: Acquire business analysis data of the plurality of objects, wherein the business analysis data includes at least one of the following: business demand data, historical business processing data, and market data; Searching for the first entity from the knowledge graph according to the business analysis data; Matching the first entity with a node in the knowledge graph to obtain a first node; Setting label information of the first node to mark the first entity in the knowledge graph; generating a query instruction according to the entity feature of the first entity; Searching the knowledge graph according to the query instruction for entities that have an association relationship with the first entity, wherein the association relationship includes at least one of the following: a supply chain relationship, a financial relationship, a customer relationship, a technical relationship, an organizational structure relationship, a human resources relationship, and an industry relationship for the target task; A group structure is constructed between the first entity and entities that have an association relationship with the first entity to obtain the first group structure.

4. The method according to claim 1, characterized in that The first group structure is expanded according to N relationship types to obtain a second group structure, including: Obtaining N relationship types according to a processing type of the target task processed by the first object; Identify a first other entity that is directly associated with the entity in the first group structure through N of the relationship types, and find a second other entity that is indirectly associated with the entity in the first group structure; Scoring the first other entity and the second other entity respectively to obtain a plurality of scores, wherein the scores are used to indicate the importance of the first other entity or the second other entity to the first group structure, and the scores are determined based on at least one of the following information between the first other entity or the second other entity and the first entity: strength of relationship, frequency of interaction, and time of interaction; A target entity is screened out based on the multiple scores, and the target entity is added to the first group structure to obtain the second group structure.

5. The method according to claim 1, characterized in that Determining a target object from the second group structure includes: Determine entity characteristic information of a second entity in the second group structure, wherein the second entity is any entity in the second group structure, and the entity characteristic information includes at least one of the following: supply characteristics, financial characteristics, technical characteristics, and market performance characteristics; Determining, according to the entity feature information, information on the time association strength with a third other entity in the second group structure, wherein the information on the time association strength is used to indicate the degree of association between the second entity and the third other entity when processing the target task; In a case where the association strength information matches the preset association strength information, an object corresponding to the second entity is determined as the target object.

6. The method according to claim 1, characterized in that After determining the target object from the second group structure, the method further includes: Locating the category of the target object in the process of processing the target task according to the object characteristics, geographical location and business processing information of the target object, and obtaining the target category; Generating target recommendation information according to the target category, wherein the target recommendation information includes at least one of the following: recommendation time, recommendation reason, recommended product, recommended process, and recommended scope; The target product is recommended to the target object using the target recommendation information.

7. The method according to claim 6, characterized in that After recommending the target product to the target object using the target recommendation information, the method further includes: Acquire a response of the target object to the recommendation instruction to obtain a response result, wherein the recommendation instruction is an instruction generated based on the target recommendation information; The response result is used to adjust the association relationship between the target entity corresponding to the target object and the first entity in the second group structure.

8. A device for determining a target object, characterized in that: include: A first marking module is used to mark a first entity in the constructed knowledge graph, and identify an entity that has an association relationship with the first entity from the knowledge graph to obtain a first group structure, wherein the knowledge graph includes multiple entities corresponding to multiple objects for a target task, and the relationship between the multiple entities, and the first object corresponding to the first entity is an entity that meets a preset condition; A first expansion module, configured to expand the first group structure according to N relationship types to obtain a second group structure, wherein all entities included in the second group structure have an association relationship with the first entity; A first determining module is configured to determine a target object from the second group structure, wherein the target object is an object whose association degree with the first object is greater than a preset threshold.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the method described in any one of claims 1 to 7 when executed by a processor.

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

Images