Enterprise Recommended Methods, Apparatus, Equipment and Storage Media

Abstract

This specification discloses a method, apparatus, device, and storage medium for enterprise recommendation. The method includes: obtaining an enterprise recommendation request from an enterprise, the enterprise recommendation request carrying enterprise information of the enterprise, the enterprise information representing the enterprise's products and its role in the industry chain, and the enterprise recommendation request requesting recommendations of partner enterprises to the enterprise; structuring the enterprise information to obtain product entities and enterprise role entities in the enterprise information; querying an industry graph based on the product entities to obtain multiple candidate partner enterprises, the industry graph including multiple enterprises belonging to different industry chains, the multiple candidate partner enterprises belonging to the same industry chain as the enterprise; and determining a target partner enterprise to recommend to the enterprise from the multiple candidate partner enterprises based on the enterprise role entities.

Description

Technical Field

[0001] This specification relates to the field of computer technology, and in particular to a business recommendation method, apparatus, device, and storage medium. Background Technology

[0002] With the development of computer technology, more and more companies are using computer technology for business activities. For example, companies use computer technology for procurement and sales activities.

[0003] In related technologies, businesses often need to use search engines to find partners for their operations, such as purchasing and sales companies. However, using search engines to find partners is inefficient. Summary of the Invention

[0004] This specification provides an embodiment of a method, apparatus, device, and storage medium for recommending companies, which can improve the efficiency of identifying cooperative companies. The technical solution is as follows:

[0005] On the one hand, a business recommendation method is provided, the method comprising:

[0006] Obtain a company recommendation request, the company recommendation request carries the company information, the company information is used to represent the company's products and its role in the industry chain, and the company recommendation request is used to request recommendations of cooperative companies to the company;

[0007] The enterprise information is structured to obtain product entities and enterprise role entities within the enterprise information;

[0008] Based on the product entity, a query is performed in the industry map to obtain multiple candidate cooperative enterprises. The industry map includes multiple enterprises belonging to different industry chains, and the multiple candidate cooperative enterprises belong to the same industry chain as the enterprise.

[0009] Based on the enterprise role entity, a target cooperative enterprise is determined from the plurality of candidate cooperative enterprises and recommended to the enterprise.

[0010] On the one hand, an enterprise recommendation device is provided, the device comprising:

[0011] The request acquisition module is used to acquire enterprise recommendation requests from enterprises. The enterprise recommendation requests carry enterprise information, which represents the enterprise's products and its role in the industry chain. The enterprise recommendation requests are used to request recommendations of cooperative enterprises to the enterprise.

[0012] The structuring module is used to structure the enterprise information to obtain product entities and enterprise role entities in the enterprise information;

[0013] The query module is used to query the industry map based on the product entity to obtain multiple candidate cooperative enterprises. The industry map includes multiple enterprises belonging to different industry chains, and the multiple candidate cooperative enterprises belong to the same industry chain as the enterprise.

[0014] The enterprise determination module is used to determine, based on the enterprise role entity, a target cooperative enterprise to be recommended to the enterprise from among the multiple candidate cooperative enterprises.

[0015] In one possible implementation, the structured module is used to perform named entity recognition on the enterprise information to obtain the product entity and the enterprise role entity in the enterprise information.

[0016] In one possible implementation, the structured module is used to input the enterprise information into an entity recognition model, encode multiple characters in the enterprise information through the entity recognition model to obtain the encoded features of the enterprise information; decode the encoded features through the entity recognition model to obtain a label for each character in the enterprise information, the label being used to represent the character type of the corresponding character; and combine the multiple characters based on their character types to obtain the product entity and the enterprise role entity.

[0017] In one possible implementation, the query module is used to determine at least one industry chain name entity corresponding to the product entity; and to query the industry map based on the at least one industry chain name entity to obtain the multiple candidate cooperative enterprises.

[0018] In one possible implementation, the query module is used to standardize the product entity to obtain at least one standard product entity; and to perform a higher-level mapping on the at least one standard product entity to obtain the at least one industry chain name entity.

[0019] In one possible implementation, the query module is used to perform semantic recognition on the product entity to obtain the semantic features of the product entity; and to perform semantic matching in a set of standard product entities based on the semantic features of the product entity to obtain at least one standard product entity corresponding to the product entity, wherein the set of standard product entities includes multiple standard product entities.

[0020] In one possible implementation, the enterprise determination module is used to determine at least one target enterprise role entity corresponding to the enterprise role entity, wherein the at least one target enterprise role entity and the enterprise role entity constitute an upstream and downstream relationship; and based on the at least one target enterprise role entity, to determine the enterprise's target cooperative enterprise from the plurality of candidate cooperative enterprises.

[0021] In one possible implementation, the product entity includes a product name entity and a product specification entity. The product name entity is used to obtain the plurality of candidate partner companies. The company determination module is used to match the plurality of candidate partner companies based on the at least one target company role entity to obtain a plurality of first reference partner companies. Based on the product specification entity and the company profiles of the plurality of first reference partner companies, the target partner company of the company is determined from the plurality of first reference partner companies.

[0022] In one possible implementation, the enterprise determination module is used to match the plurality of first reference cooperative enterprises based on the product specification entity to obtain a plurality of second reference cooperative enterprises; and to determine the target cooperative enterprise of the enterprise from the plurality of second reference cooperative enterprises based on the enterprise profiles of the plurality of second reference cooperative enterprises.

[0023] In one possible implementation, the enterprise determination module is configured to determine the enterprise score of each of the plurality of second reference partner enterprises based on the enterprise profiles of the plurality of second reference partner enterprises; sort the plurality of second reference partner enterprises according to the enterprise scores of the plurality of second reference partner enterprises; and determine the top target number of sorted second reference partner enterprises as the target partner enterprises of the enterprise.

[0024] In one possible implementation, the enterprise determination module is further configured to perform at least one of the following:

[0025] The company's historical partners are identified as the company's target partners.

[0026] At least one enterprise of the same type as the historical partner enterprise is identified as the target partner enterprise of the enterprise.

[0027] On one hand, a computer device is provided, the computer device including one or more processors and one or more memories, the one or more memories storing at least one computer program, the computer program being loaded and executed by the one or more processors to implement the enterprise recommendation method.

[0028] On one hand, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer program, which is loaded and executed by a processor to implement the enterprise recommendation method.

[0029] On one hand, a computer program product or computer program is provided, which includes program code stored in a computer-readable storage medium, wherein a processor of a computer device reads the program code from the computer-readable storage medium and executes the program code, causing the computer device to perform the method recommended by Company X described above.

[0030] The technical solution provided in the embodiments of this specification obtains a company recommendation request, which carries the company's information. The company information is structured to obtain product entities and company role entities. Based on the product entities, a query is performed in the industry map to obtain multiple candidate partner companies belonging to the same industry chain as the company. These multiple candidate partner companies are companies that may cooperate with the company. Based on the company role entities, a target partner company is determined from these multiple candidate partner companies; this target partner company is also the partner company recommended to the company. The above-described company recommendation method is highly efficient. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram of the implementation environment of an enterprise recommendation method provided in the embodiments of this specification;

[0033] Figure 2 This is a flowchart of an enterprise recommendation method provided in the embodiments of this specification;

[0034] Figure 3 This is a flowchart of another enterprise recommendation method provided in the embodiments of this specification;

[0035] Figure 4 This is a schematic diagram of an enterprise recommendation interface provided in an embodiment of this specification;

[0036] Figure 5 This is a flowchart of yet another enterprise recommendation method provided in the embodiments of this specification;

[0037] Figure 6This is a schematic diagram of the structure of an enterprise recommendation device provided in the embodiments of this specification;

[0038] Figure 7 This is a schematic diagram of the structure of a terminal provided in an embodiment of this specification;

[0039] Figure 8 This is a schematic diagram of the structure of a server provided in the embodiments of this specification. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this specification clearer, the embodiments of this specification will be further described in detail below with reference to the accompanying drawings.

[0041] In this manual, the terms "first," "second," etc., are used to distinguish identical or similar items with essentially the same function. It should be understood that there is no logical or temporal dependency between "first," "second," and "nth," nor are there any restrictions on quantity or execution order.

[0042] Knowledge Graph: In the library and information science field, a knowledge graph is also known as a knowledge domain visualization or knowledge domain mapping map. It is a series of different graphics that show the development process and structural relationships of knowledge. It uses visualization technology to describe knowledge resources and their carriers, and to mine, analyze, construct, draw and display knowledge and the interrelationships between them.

[0043] Industry map: A knowledge representation that connects enterprises, products, industries, and events.

[0044] Industrial chain: An industrial chain refers to the chain-like linkages objectively formed between various industrial sectors based on certain technological and economic connections and spatiotemporal layout relationships. It can usually be examined from four dimensions: value chain, enterprise chain, supply and demand chain, and spatial chain. The industrial chain covers the entire process of product production or service provision, including power supply, raw material production, technology research and development, intermediate product manufacturing, end product manufacturing, and even distribution and consumption. It is a unity of industrial organization, production process, and value realization.

[0045] Intent recognition: Used to identify the themes and intents contained in text.

[0046] Entity: An entity is an object that exists in the real world and can be distinguished from other objects. We use a series of attributes to describe this entity and highlight its differences.

[0047] Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence. Its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning by demonstration.

[0048] Embedded coding, mathematically speaking, represents a correspondence, that is, mapping data in space X to space Y using a function F. This function F is injective, and the mapping result preserves the structure. An injective function means that the mapped data uniquely corresponds to the original data, and preserving the structure means that the order of the original data remains the same. For example, if there are data X1 and X2 before mapping, after mapping we get Y1 corresponding to X1 and Y2 corresponding to X2. If the original data X1 > X2, then correspondingly, the mapped data Y1 > Y2. For words, this means mapping words to another space to facilitate subsequent machine learning and processing.

[0049] It should be noted that all information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in this manual are authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the enterprise information and enterprise profiles involved in this manual were obtained with full authorization.

[0050] Figure 1 This is a schematic diagram illustrating the implementation environment of an enterprise recommendation method provided in the embodiments of this specification. See also... Figure 1 The implementation environment may include terminal 110 and server 140.

[0051] Terminal 110 connects to server 140 via a wireless or wired network. Optionally, terminal 110 can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. Terminal 110 has applications installed and running that support partner company recommendations.

[0052] Server 140 is a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing 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, Content Delivery Network (CDN), and big data and artificial intelligence platforms. Server 140 provides background services for applications running on terminal 110.

[0053] Those skilled in the art will understand that the number of terminals described above can be more or less. For example, there may be only one terminal, or there may be dozens or hundreds of terminals, or even more, in which case other terminals may also be included in the above implementation environment. This specification does not limit the number of terminals or the type of devices in the embodiments.

[0054] After introducing the implementation environment of the embodiments of this specification, the application scenarios of the embodiments of this specification will be described below in conjunction with the above implementation environment. In the following description, the terminal is also the terminal 110 in the above implementation environment, and the server is also the server 140 in the above implementation environment. The technical solutions provided by the embodiments of this specification can be applied to scenarios of recommending cooperative enterprises to enterprises with different roles. For example, it can be applied to scenarios of recommending raw material enterprises and sales enterprises to manufacturing enterprises, or to recommend manufacturing enterprises to raw material enterprises, or to recommend manufacturing enterprises to sales enterprises. The embodiments of this specification do not limit this.

[0055] After adopting the technical solution provided in the embodiments of this specification, when a company wants to recommend partner companies, it inputs its company information on the terminal. This company information describes the company's situation. In response to the operation on the terminal, the terminal sends a company recommendation request to the server. This company recommendation request carries the company information and is used to request recommendations of partner companies to the company. The server receives the company recommendation request and obtains the company information from it. The server structures the company information to obtain the product entity and the company role entity. This structuring is done to extract information for company recommendation. Based on the product entity, the server queries the industry map to obtain multiple candidate partner companies. These multiple candidate partner companies are companies related to the product entity. Since the industry map includes multiple companies belonging to different industry chains, these multiple candidate partner companies can reflect relevant information of the industry chain. Based on the company role entity, the target partner company can be determined from the multiple candidate partner companies. This target partner company is the partner company recommended to the company.

[0056] After introducing the implementation environment and application scenarios of the embodiments in this specification, the technical solutions provided in the implementation of this specification will be introduced below.

[0057] Figure 2 This is a flowchart of an enterprise recommendation method provided in the embodiments of this specification. See also... Figure 2 Taking the server as the executing entity as an example, the method includes the following steps.

[0058] 202. The server obtains a company recommendation request. The company recommendation request carries the company's information, which represents the company's products and its role in the industry chain. The company recommendation request is used to request recommendations of partner companies to the company.

[0059] In this context, "enterprise" refers to the company to which a partner company is to be recommended. The enterprise recommendation request is used to request recommendations of potential partners, i.e., to request recommendations of potential partners. In some embodiments, recommending a partner company is also referred to as recommending a business opportunity. The enterprise information represents the enterprise's products and its role in the industry chain. The enterprise's products can be those it manufactures or those it sells; this specification does not limit this. The enterprise's role in the industry chain refers to its position within the chain. Enterprises with different roles play different parts of the industry chain, including raw material supply, product manufacturing, and product sales.

[0060] 204. The server structures the enterprise information to obtain the product entities and enterprise role entities in the enterprise information.

[0061] The purpose of structuring enterprise information is to extract key information, specifically product entities and enterprise role entities. Product entities represent the enterprise's products, while enterprise role entities represent the enterprise's role in the industry chain. After obtaining these entity types, subsequent enterprise recommendations can be made based on these information.

[0062] 206. The server queries the industry map based on the product entity and obtains multiple candidate partner companies. The industry map includes multiple companies belonging to different industry chains.

[0063] This industry map includes multiple companies across different industry chains. Industry chains reflect the roles of companies and the relationships between them, and can be seen as a highly abstract representation of the relationships between companies. By understanding industry chains, one can quickly grasp the situation of companies within those chains. The industry map can be viewed as a collection of knowledge providing information from multiple industry chains, recording the industry chain to which a company belongs and its role within that chain. Since a product entity reflects the industry chain in which a company operates, searching the industry map based on that product entity can yield multiple candidate partners within the same industry chain. These candidate partners are all potential collaborators for that company.

[0064] 208. Based on the enterprise role entity, the server determines the target cooperative enterprise to recommend to the enterprise from among the multiple candidate cooperative enterprises.

[0065] The enterprise role entity reflects the enterprise's role in the industry chain. The target partner is determined from multiple candidate partners based on the enterprise role entity. In other words, the multiple candidate partners are screened based on the enterprise role entity, and the target partner is finally obtained. The target partner is the enterprise that matches the enterprise role.

[0066] The technical solution provided in the embodiments of this specification obtains a company recommendation request, which carries the company's information. The company information is structured to obtain product entities and company role entities. Based on the product entities, a query is performed in the industry map to obtain multiple candidate partner companies belonging to the same industry chain as the company. These multiple candidate partner companies are companies that may cooperate with the company. Based on the company role entities, a target partner company is determined from these multiple candidate partner companies; this target partner company is also the partner company recommended to the company. The above-described company recommendation method is highly efficient.

[0067] Steps 202-208 above are a brief introduction to the technical solutions provided in the embodiments of this specification. The technical solutions provided in the embodiments of this specification will be explained more clearly below with reference to some examples. See also... Figure 3 Taking the server as the executing entity as an example, the method includes the following steps.

[0068] 302. The server obtains a company recommendation request. The company recommendation request carries the company's information, which represents the company's products and its role in the industry chain. The company recommendation request is used to request recommendations of partner companies to the company.

[0069] In this document, the enterprise is the company to be recommended for cooperation. The enterprise recommendation request is used to request recommendations of potential partners to the enterprise. In some embodiments, recommending a partner is also referred to as recommending a business opportunity. The enterprise information is used to represent the enterprise's products and its role in the industry chain. The enterprise's products can be products manufactured by the enterprise or products sold by the enterprise; this specification does not limit this. In some embodiments, the enterprise information is also referred to as the enterprise's product demand data. This enterprise information is set by the enterprise's staff based on the enterprise's actual situation. For example, the enterprise information may be "producing plastic figurine molds," where "producing" indicates that the enterprise's role in the industry chain is product manufacturing, and "plastic figurine molds" indicates that the type of the enterprise's product is molds. In some embodiments, in addition to carrying the type of the enterprise's product, the enterprise information also carries the specifications of the enterprise's product. Accordingly, the enterprise information may be "producing plastic figurine molds, length x width Y height Z," where "length x width Y height Z" is the specification of the enterprise's product.

[0070] The enterprise's role in the industrial chain refers to its position within the chain. Enterprises with different roles play different functions in the industrial chain. These roles include raw material supply, product manufacturing, and product sales. In some embodiments, if the enterprise's role is product manufacturing, then enterprises in the same industrial chain that are raw material suppliers and product sellers may be enterprises of interest to the enterprise, that is, enterprises that may cooperate with the enterprise.

[0071] In one possible implementation, the server obtains a company recommendation request from the terminal, which carries the company's information. The terminal is associated with the company, such as a terminal used by the company's employees; however, this specification does not limit this specific instance.

[0072] In this implementation, the server can obtain the enterprise recommendation request from the terminal, and subsequently recommend partner companies to the enterprise based on the enterprise recommendation request.

[0073] For example, the terminal displays a company recommendation interface used to obtain company information. In response to an operation on this interface, the terminal sends a company recommendation request to the server, carrying the company information obtained from the interface. The server then retrieves the company recommendation request.

[0074] For example, see Figure 4The terminal displays a company recommendation interface 400, which includes a company information acquisition area 401 and a request sending control 402. The company information acquisition area 401 is used to input company information, and the request sending control 402 is used to send a company recommendation request. In response to an information input operation in the company information acquisition area 401, the corresponding company information is displayed in the company information acquisition area 401. In response to a click operation on the request sending control 402, the terminal sends a company recommendation request to the server, which carries the company information displayed in the company information acquisition area 401.

[0075] Based on the above implementation, optionally, before the server obtains the enterprise recommendation request from the terminal, the server can also authenticate the enterprise by performing the following steps. If the authentication is successful, the server obtains the enterprise recommendation request and makes an enterprise recommendation based on the enterprise recommendation request; if the authentication fails, the server does not obtain the enterprise recommendation request sent by the terminal or does not respond to the enterprise recommendation request sent by the terminal.

[0076] In some embodiments, the server obtains authentication information from the terminal, which carries the enterprise's identifier. The server then authenticates the enterprise based on this authentication information. For example, the server obtains the authentication information from the terminal and extracts the enterprise's identifier from it. The server then authenticates the enterprise based on the enterprise's identifier. If the enterprise is successfully authenticated based on the authentication information, the server sends authentication success information to the terminal, indicating that the enterprise's authentication has been successful; if the enterprise fails to be authenticated based on the authentication information, the server sends authentication failure information to the terminal, indicating that the enterprise's authentication has failed.

[0077] 304. The server performs named entity recognition on the enterprise information to obtain the product entity and enterprise role entity in the enterprise information.

[0078] The process of performing named entity recognition (NER) on the enterprise information is essentially structuring that information. This structuring aims to extract key information, specifically product entities and enterprise role entities. Product entities represent the enterprise's products, while enterprise role entities represent its role in the industry chain. After obtaining these entities, subsequent enterprise recommendations can be made based on them. Named Entity Recognition (NER), also known as proper noun recognition, identifies entities with specific meanings in text, primarily including personal names, place names, organization names, and proper nouns. In this embodiment, NER is used to identify product entities and enterprise role entities, which are among the aforementioned proper nouns.

[0079] In one possible implementation, the server inputs the enterprise information into an entity recognition model. The model encodes multiple characters within the enterprise information to obtain its encoded features. The server then decodes these encoded features using the entity recognition model to obtain labels for each character in the enterprise information. These labels represent the character type of the corresponding character. Based on the character types, the server combines these characters to obtain the product entity and the enterprise role entity.

[0080] This entity recognition model is used for named entity recognition. The input to the model is the text to be recognized, and the output is the character types of multiple characters in the text. The model is trained through multiple iterations. The company information includes multiple characters; encoding these characters is done to abstractly represent them, which helps the entity recognition model identify the relationships between them, thus achieving named entity recognition.

[0081] In this implementation, the server can perform named entity recognition on enterprise information through an entity recognition model, thereby obtaining product entities and enterprise role entities in the enterprise information. Since the entity recognition model has a certain generalization ability, the efficiency and accuracy of using the entity recognition model to perform named entity recognition on enterprise information are both high.

[0082] To illustrate the above embodiments more clearly, several examples will be used below.

[0083] Example 1: The server inputs the enterprise information into an entity recognition model. The encoder of the entity recognition model encodes multiple characters in the enterprise information based on an attention mechanism, obtaining the encoded features of the enterprise information. The server then inputs these encoded features into the decoder of the entity recognition model. The decoder decodes these features based on an attention mechanism, outputting a tag sequence corresponding to the multiple characters. This tag sequence includes multiple tags, each tag corresponding to one character, and the tags represent the type of the corresponding character. In this embodiment, the tags include three types: non-entity, product entity, and enterprise role entity. Based on this tag sequence, the server combines adjacent characters in the enterprise information corresponding to the same tag to obtain the product entity and the enterprise role entity within the enterprise information.

[0084] For example, the server embeds and encodes multiple characters in the enterprise information to obtain the embedding features of each character. These embedding features include character features and positional features, with the positional features representing the character's location within the enterprise information. The server inputs these embedding features into the encoder of the entity recognition model. The encoder performs a linear transformation on the embedding features of each character to obtain the query matrix, key matrix, and value matrix for each character. For any given character, the server, using the encoder, determines the attention weights of the multiple characters on that character based on the query matrix and the key matrix. These attention weights include the character's own attention weights. The server, using the encoder, weights and fuses the value matrices of the multiple characters based on their attention weights to obtain the character's attention features. Finally, the server fuses these attention features to obtain the encoded features of the enterprise information. The server inputs the encoded features of the enterprise information into the decoder of the entity recognition model. The decoder performs multiple rounds of iterative decoding on the encoded features based on an attention mechanism, and outputs a sequence of labels corresponding to the multiple characters. In the process of multiple rounds of iterative decoding, except for the last round of iterative decoding, each round of iterative decoding outputs one label from the label sequence. The output of each round of iterative decoding is used as the input for the next round of iterative decoding, and the output of the last round of iterative decoding is a terminator used to indicate the cessation of iterative decoding.

[0085] In some embodiments, the product entity includes a product name entity and a product specification entity. Accordingly, the tags in the tag sequence output by the entity recognition model include four types: non-entity, product name entity, product specification, and enterprise role entity. Ultimately, the product name entity, product specification entity, and enterprise role entity in the enterprise information can be obtained.

[0086] Example 2: The entity recognition model includes a product entity recognition sub-model and an enterprise role entity recognition sub-model. The server inputs the enterprise information into the product entity recognition sub-model and the enterprise role entity recognition sub-model. The encoder of the product entity recognition sub-model encodes multiple characters in the enterprise information based on an attention mechanism to obtain the encoded features of the enterprise information. The server inputs the encoded features of the enterprise information into the decoder of the product entity recognition sub-model. The decoder of the product entity recognition sub-model decodes the encoded features of the enterprise information based on an attention mechanism, outputting a first label sequence corresponding to the multiple characters. This first label sequence includes multiple labels, one label corresponding to one character, and the label is used to indicate the type of the corresponding character. In the embodiments of this specification, the labels include both non-entity and product entity types. Based on the first label sequence, the server combines adjacent characters in the enterprise information corresponding to the same label to obtain the product entity in the enterprise information. The encoder of the enterprise role entity recognition sub-model encodes multiple characters in the enterprise information based on an attention mechanism to obtain the encoded features of the enterprise information. The server inputs the encoded features of the enterprise information into the decoder of the enterprise role entity recognition sub-model. The decoder, based on an attention mechanism, decodes the encoded features of the enterprise information and outputs a second tag sequence corresponding to the multiple characters. This second tag sequence includes multiple tags, each tag corresponding to one character, and the tags indicate the type of the corresponding character. In the embodiments of this specification, the tags include both non-entity and enterprise role entity tags. Based on this second tag sequence, the server combines adjacent characters in the enterprise information that correspond to the same tag to obtain the enterprise role entity within the enterprise information.

[0087] It should be noted that the method described above for determining the tag sequence of enterprise information through encoders and decoders belongs to the same inventive concept as the description in Example 1 above, and the implementation process will not be repeated here.

[0088] In some embodiments, the product entity includes a product name entity and a product specification entity. Correspondingly, the entity identification model includes a product name entity identification sub-model, a product specification entity identification sub-model, and an enterprise role entity identification sub-model. The tags in the tag sequence include four types: non-entity, product name entity, product specification, and enterprise role entity. Ultimately, the product name entity, product specification entity, and enterprise role entity in the enterprise information can be obtained.

[0089] The above implementation method is illustrated by taking the server as an example to perform named entity recognition on enterprise information through an entity recognition model. The following describes other methods for the server to perform named entity recognition on enterprise information.

[0090] In one possible implementation, the server segments the enterprise information to obtain multiple candidate words. Based on these candidate words, the server matches them against an entity set to obtain product entities and enterprise role entities within the enterprise information. This entity set includes multiple candidate product entities and multiple candidate enterprise role entities.

[0091] The purpose of segmenting the company information is to break it down into multiple candidate words. These candidate words can contain the same number of characters or different numbers of characters; this embodiment does not impose such a limitation. The multiple candidate product entities and multiple candidate company role entities in the entity set are set by technical personnel according to actual conditions and can be updated at any time.

[0092] In this implementation, the server can extract product entities and enterprise role entities from the enterprise information through word segmentation and matching, which can be applied to a wider range of scenarios.

[0093] For example, the server inputs the company information into a word segmentation model, which segments the information to obtain multiple candidate words. This word segmentation model is used to segment the input text. The server compares these candidate words with candidate product entities and multiple candidate enterprise role entities in the entity set, and identifies the candidate product entities and candidate enterprise role entities that match the candidate words as the product entities and enterprise role entities in the company information. For example, this word segmentation model is an N-gram model, where N is a positive integer.

[0094] In some embodiments, the product entity includes a product name entity and a product specification entity. Accordingly, the multiple candidate product entities in the entity set include multiple candidate product name entities and multiple candidate product specification entities. After segmenting the enterprise information, the multiple candidate words obtained are matched with the entity set to ultimately obtain the product name entity, product specification entity, and enterprise role entity in the enterprise information.

[0095] It should be noted that the server can use any of the above methods to perform named entity recognition on the enterprise information, and the embodiments in this specification do not limit this.

[0096] The above describes the method for the server to perform named entity recognition on the enterprise information. In some embodiments, after step 302, the server can either execute the above step 304 or directly execute at least one of the following steps to directly determine the target cooperative enterprise recommended to the enterprise. This specification does not limit this embodiment.

[0097] In one possible implementation, the server identifies the enterprise's historical partners as the enterprise's target partners.

[0098] Among them, the company's historical partners are those companies that have cooperated with the company.

[0099] In this implementation, after receiving a company recommendation request, the server can directly identify the company's historical partners as target partners to recommend to the company without further processing, resulting in high efficiency in company recommendations.

[0100] For example, the company recommendation request also carries the company's identifier. After receiving the recommendation request, the server retrieves the company's identifier from it. Based on this identifier, the server queries the company's partner database to obtain the company's historical partners. This database stores cooperation records between companies, which are uploaded by the companies themselves. The server then identifies these historical partners as the company's target partners.

[0101] In one possible implementation, the server identifies at least one enterprise of the same type as the historical partner enterprise as the target partner enterprise.

[0102] In this implementation, the server can determine the target partner companies of the enterprise through collaborative filtering, resulting in high efficiency in enterprise recommendations.

[0103] For example, the company recommendation request also carries the company's identifier. After receiving the recommendation request, the server retrieves the company's identifier from it. Based on this identifier, the server queries the company cooperation database to obtain the company's historical partners. This database stores cooperation records between companies, which are uploaded by the companies themselves. The server identifies at least one company of the same type as these historical partners and designates this at least one company as the company's target partner.

[0104] In some embodiments, under the two implementation methods described above, if the enterprise does not have any historical partner enterprises, the server can continue to execute step 304.

[0105] 306. The server determines at least one industry chain name entity corresponding to the product entity.

[0106] In this context, the industry chain name entity represents the name of the industry chain. The industry chain name entity corresponding to the product entity also indicates the industry chain to which the product belongs. For example, if a product entity is "plastic toy," and the corresponding industry chain name entity is "toy," then the product belonging to the "plastic toy" industry chain is "toy." When the product entity includes both a product name entity and a product specification entity, step 306 above determines at least one industry chain name entity corresponding to the product name entity. Accordingly, the product entity described below can be the product name entity.

[0107] In one possible implementation, the server standardizes the product entity to obtain at least one standard product entity. The server then performs a higher-level mapping on the at least one standard product entity to obtain the at least one industry chain name entity.

[0108] The standardization of the product entity aims to map it into a standard product system. A standardized product entity can more accurately represent the corresponding product. The standard product system is a product description method jointly developed by the industry. Mapping the standard product entity to a higher level determines the category to which the standard product belongs, that is, the industrial chain to which the standard product entity belongs.

[0109] In this implementation, the server can determine the industry chain name entity through standardization and higher-level mapping, and then query related companies through the industry chain name entity, which is highly efficient.

[0110] To provide a clearer explanation of the above embodiments, the following description will be divided into two parts.

[0111] Part 1: The server standardizes the product entity to obtain at least one standard product entity.

[0112] In one possible implementation, the server performs semantic recognition on the product entity to obtain its semantic features. Based on these features, the server performs semantic matching within a set of standard product entities to obtain at least one standard product entity corresponding to the product entity. This set of standard product entities includes multiple standard product entities.

[0113] The set of standard product entities is the aforementioned standard product system.

[0114] In this implementation, the server can extract the semantic features of the product entity through semantic recognition, and determine the standard product entity based on the semantic features of the product entity. Since the semantic features can reflect the semantics of the product entity, product entities of different forms and expressions can all obtain corresponding standard product entities, and the accuracy of determining the standard product entity is high.

[0115] The above implementation method includes two processes: semantic recognition and semantic matching. These two processes will be described in detail below.

[0116] Process A: The server performs semantic recognition on the product entity to obtain the semantic features of the product entity.

[0117] In one possible implementation, the server inputs the product entity into a semantic recognition model, performs semantic recognition on the product entity through the semantic recognition model, and obtains the semantic features of the product entity.

[0118] The semantic recognition model has the function of semantic recognition and can extract the semantic features of the input text. This semantic recognition model is a model obtained through multiple rounds of iterative training. The semantic features of a product entity are an abstract expression of the semantics of the product entity. Since multiple entities with the same semantics may have different forms, semantic features can unify multiple entities with the same semantics.

[0119] In this implementation, the server can extract the semantic features of product entities through a semantic recognition model, making full use of the generalization ability of the semantic recognition model to achieve efficient and accurate semantic feature extraction.

[0120] For example, the server inputs the product entity into the semantic recognition model. The model then performs attention encoding on the product entity based on an attention mechanism to obtain its semantic features. For instance, the server performs embedding encoding on the product entity to obtain the embedding features of each character. These embedding features include the character's characteristics and positional characteristics, with the positional characteristics representing the character's location within the product entity. The server inputs the embedding features of multiple characters into the semantic recognition model and performs a linear transformation on these features to obtain the query matrix, key matrix, and value matrix for each character. For any given character, the server uses the semantic recognition model to determine the attention weights of the multiple characters on that character, based on the query matrix and the key matrix. These attention weights include the character's own attention weights. The server then uses the semantic recognition model to weight and fuse the value matrices of the multiple characters based on these attention weights to obtain the character's attention features. Finally, the server fuses these attention features to obtain the semantic features of the product entity.

[0121] Process B: The server performs semantic matching in the standard product entity set based on the semantic features of the product entity to obtain at least one standard product entity corresponding to the product entity.

[0122] One product entity may correspond to multiple standard product entities. For example, there may be a product entity "plastic figurine mold", which may correspond to two standard product entities: "plastic mold" and "figurine".

[0123] In one possible implementation, the server determines the similarity between the semantic features of the product entity and the semantic features of multiple standard product entities in the standard product entity set. The server then identifies the standard product entities whose similarity meets the similarity criteria as at least one standard product entity corresponding to the product entity.

[0124] The similarity criteria include a similarity greater than or equal to a similarity threshold and similarity to any one of the top N similarities among the plurality of standard product entities, where N is a positive integer. The similarity threshold and N are set by technicians according to actual circumstances, and this specification does not limit this. In some embodiments, the semantic features of the standard product entity are determined in advance by the semantic recognition model, or determined during the matching process by the semantic recognition model; this specification does not limit this.

[0125] In this implementation, the server can determine the standard product entity corresponding to the product entity based on the similarity between semantic features, and the accuracy of determining the standard product entity is relatively high.

[0126] Taking semantic features as a representation of semantic feature vectors as an example, the server determines the cosine similarity between the semantic feature vector of the product entity and the semantic feature vectors of multiple standard product entities. The server identifies the standard product entities whose cosine similarity is greater than or equal to the probability threshold as the standard product entities of the product entity, or the server identifies the first two standard product entities among the multiple standard product entities whose cosine similarity is less than or equal to the probability threshold as the standard product entities of the product entity.

[0127] Part Two: The server performs a higher-level mapping on the at least one standard product entity to obtain the at least one industry chain name entity.

[0128] The superordinate mapping is used to determine the superordinate entity of an entity, that is, to determine the entity of the category to which an entity belongs.

[0129] In one possible implementation, the server compares the at least one standard product entity with standard product entities under multiple candidate industry chain name entities, and determines at least one industry chain name entity corresponding to the at least one standard product from the multiple candidate industry chain name entities based on the comparison results.

[0130] In this implementation, the server can perform a higher-level mapping on the at least one standard product entity by comparison to obtain the at least one industry chain name entity, and the accuracy of the higher-level mapping is relatively high.

[0131] For example, the server queries the supply chain knowledge graph for at least one standard product entity, obtaining the product node corresponding to that standard product entity. This supply chain knowledge graph includes multiple product nodes and multiple supply chain nodes. Product nodes correspond to standard product entities, and supply chain nodes correspond to supply chain name entities. Connections between product nodes and supply chain nodes indicate a subordinate relationship; that is, the standard product entity corresponding to a product node belongs to the supply chain name entity corresponding to the supply chain node. Based on the connections between product nodes and supply chain nodes in the supply chain knowledge graph, the server determines at least one supply chain node connected to the product node corresponding to the at least one standard product entity. The server then identifies the supply chain name entity corresponding to the at least one supply chain node as the at least one supply chain name entity corresponding to the at least one standard product.

[0132] 308. The server queries the industry map based on the name entity of at least one industry chain to obtain multiple candidate cooperative enterprises. The industry map includes multiple enterprises belonging to different industry chains, and the multiple candidate cooperative enterprises belong to the same industry chain as the enterprise.

[0133] This industry map includes multiple enterprises across different industry chains. Industry chains reflect the roles of enterprises and the relationships between them, and can be seen as a highly abstract representation of enterprise relationships. By understanding industry chains, one can quickly grasp the situation of enterprises within those chains. The industry map can be viewed as a knowledge collection providing information from multiple industry chains, recording the industry chain to which an enterprise belongs and its role within that chain. Since a product entity reflects the industry chain in which an enterprise operates, querying the industry map based on that product entity can yield multiple candidate partner enterprises in the same industry chain. These candidate partner enterprises are all potential partners of the enterprise. In some embodiments, the industry map is pre-generated by the server and can be directly used during enterprise recommendation. For example, the server retrieves multiple enterprises across multiple industry chains. The server generates the industry map based on these multiple industry chains and the multiple enterprises within them.

[0134] In one possible implementation, the server queries the industry map based on the at least one industry chain name entity to obtain at least one industry chain corresponding to the at least one industry chain name entity. The server then identifies multiple enterprises under the at least one industry chain as multiple candidate cooperative enterprises.

[0135] In this implementation, the at least one industry chain name entity is the industry chain name entity corresponding to the enterprise information. Therefore, the at least one industry chain determined based on the at least one industry chain name entity is also the industry chain corresponding to the enterprise information, or the industry chain to which the enterprise belongs. Identifying candidate partner enterprises in the industry chain to which the enterprise belongs can make full use of industry chain information and improve the accuracy of candidate partner enterprises.

[0136] 310. The server determines at least one target enterprise role entity corresponding to the enterprise role entity, and the at least one target enterprise role entity and the enterprise role entity form an upstream and downstream relationship.

[0137] In this context, upstream and downstream relationships are concepts within the same industry chain, used to represent the relationship between two companies. Company A being upstream of Company B means that Company A's products or services are needed by Company B for its operations. Since the technical solution provided in this specification's embodiments is for company recommendation, focusing on the upstream and downstream relationships within the industry chain can improve the accuracy of the recommendation. The number of target company entities can be one or two. If the target company entity is at the downstream or upstream end of the industry chain, the number of target company entities is one, meaning it is either upstream or downstream of that company entity. If the target company entity is not at the downstream or upstream end of the industry chain, the number of target company entities is two, meaning it is both upstream and downstream of that company entity.

[0138] In one possible implementation, the server performs upstream and downstream mapping on the enterprise role entity to obtain the at least one target enterprise role entity. The upstream and downstream relationships between enterprise role entities are pre-configured, and the at least one target role entity can be quickly determined through upstream and downstream mapping.

[0139] For example, the server queries the upstream and downstream relationship table based on the enterprise role entity to obtain at least one target enterprise role entity corresponding to the enterprise role entity. The upstream and downstream relationship table stores the upstream and downstream positions of different enterprise role entities in the industry chain. For example, if the enterprise role entity is "production", the at least one target enterprise role entity corresponding to the enterprise role entity includes "raw materials" and "sales", where "raw materials" is upstream of "production" and "sales" is downstream of production.

[0140] 312. Based on the at least one target enterprise role entity, the server determines the target cooperative enterprise to recommend to the enterprise from among the multiple candidate cooperative enterprises.

[0141] In this context, the target partner company belongs to the same industry chain as the enterprise, but its role within the industry chain differs from that of the enterprise. The enterprise role entity corresponding to the target partner company is the target enterprise role entity. The number of target partner companies can be one or more; this specification does not limit this.

[0142] In one possible implementation, the server matches the at least one target enterprise role entity among the multiple candidate partner enterprises to obtain the target partner enterprise recommended to that enterprise.

[0143] In this implementation, the server can determine the target partner from multiple candidate partners based on the at least one target enterprise role entity. The process of determining the target partner takes into account the roles of enterprises in the industry chain, and the accuracy of the determined target enterprise is relatively high.

[0144] For example, the server determines the roles of the multiple candidate partner companies in the industry chain. The server then identifies at least one candidate partner company whose role corresponds to the target company's role entity and recommends it as the target partner company to that company.

[0145] In one possible implementation, the server matches the at least one target enterprise role entity among the multiple candidate partner enterprises to obtain multiple first reference partner enterprises. Based on the enterprise profiles of these multiple first reference partner enterprises, the server determines the target partner enterprise for that enterprise from among them.

[0146] Among them, the enterprise profile is used to describe the situation of the enterprise. For example, the enterprise profile includes at least one of the following: enterprise risk, enterprise size, enterprise qualifications, and enterprise evaluation.

[0147] In this implementation, the server can identify the target partner from multiple candidate partners based on the target enterprise role entity and enterprise profile. The identification process combines industry chain information with enterprise profile, which improves the accuracy of the identified target partner.

[0148] For example, the server determines the roles of multiple candidate partner companies in the industry chain. The server identifies the candidate partner companies whose roles correspond to the target company's role entity and designates them as first reference partner companies. Based on the company profiles of these first reference partner companies, the server determines the company score for each first reference partner company. The server sorts the multiple first reference partner companies according to their company scores. The server then identifies the top target number of first reference partner companies as the target partner companies for that company. The target number is set by technical personnel based on actual circumstances, and this embodiment does not limit this setting.

[0149] The following explains the method by which the server determines enterprise scores based on enterprise profiles.

[0150] In some embodiments, where the enterprise profile includes enterprise risk, enterprise size, enterprise qualifications, and enterprise evaluation, the server determines an enterprise risk score, an enterprise size score, an enterprise qualifications score, and an enterprise evaluation score based on these factors. Specifically, higher enterprise risk results in a lower enterprise risk score, and vice versa. Higher enterprise size results in a higher enterprise size score, and vice versa. Better enterprise qualifications result in a higher enterprise qualifications score, and vice versa. A better enterprise evaluation results in a higher enterprise evaluation score, and vice versa. The server then performs a weighted sum of the enterprise risk score, enterprise size score, enterprise qualifications score, and enterprise evaluation score to obtain the enterprise score corresponding to the enterprise profile.

[0151] In one possible implementation, the product entity includes a product name entity and a product specification entity. The product name entity is used to obtain the plurality of candidate partner companies. The server matches the plurality of candidate partner companies based on the at least one target company role entity to obtain a plurality of first reference partner companies. Based on the product specification entity and the company profiles of the plurality of first reference partner companies, the server determines the target partner company of the enterprise from the plurality of first reference partner companies.

[0152] Among them, the product specification entity is used to represent the specifications of the product, such as the product's size, shape, color, and material.

[0153] In this implementation, the server can identify a target partner from multiple candidate partners based on the target enterprise role entity, enterprise profile, and product specification entity. The identification process combines industry chain information and enterprise profile with product specifications, which improves the accuracy of the identified target partners.

[0154] For example, the server determines the roles of multiple candidate partners in the industry chain. The server then identifies the candidate partners whose roles correspond to the target partner's role entity as first reference partners. Based on the product specification entity, the server matches these first reference partners to obtain multiple second reference partners. Based on the enterprise profiles of these second reference partners, the server identifies the target partner from among them.

[0155] The following describes the method by which the server determines the target partner company from among the multiple second reference partner companies based on their corporate profiles.

[0156] In some embodiments, the server determines the enterprise score of each of the plurality of second reference partners based on their enterprise profiles. The server then sorts the plurality of second reference partners according to their enterprise scores. The server identifies the top target number of the sorted second reference partners as the target partners for that enterprise.

[0157] In some embodiments, after identifying a target partner company, the server recommends that company to the terminal. For example, the server sends the target partner company's information to the terminal, which then displays that information.

[0158] The following will combine Figure 5 The technical solution provided in the embodiments of this specification is illustrated through an example.

[0159] See Figure 5 The server receives a company recommendation request, which carries the company information "producing plastic figurine molds, length x width Y height Z". The server performs product named entity recognition, role named entity recognition, and specification named entity recognition on this information, obtaining the product name entity "plastic figurine mold", the company role entity "production", and the product specification entity "length x width Y height Z", thus structuring the company information. The server standardizes the product name entity "plastic figurine mold", obtaining two standard product entities: "plastic mold" and "figurine". The server performs a higher-level mapping on these two standard product entities, mapping "plastic mold" to the industry chain name entity "mold" and "figurine" to the industry chain name entity "toy". The server identifies two industry chains corresponding to the industry chain name entities "mold" and "toy" in the industry graph, and obtains multiple candidate partner companies from these two industry chains. Based on the enterprise role entity "Production", the server identifies at least one target enterprise role entity "Raw Materials" and "Sales". The server then matches these at least one target enterprise role entity "Raw Materials" and "Sales" with multiple candidate partner enterprises to obtain multiple first reference partner enterprises. Based on the product specification entity, the server matches these multiple first reference partner enterprises to obtain multiple second reference partner enterprises. Based on the enterprise profiles of these multiple second reference partner enterprises, the server determines the target partner enterprise for that enterprise from among them.

[0160] All the above-mentioned optional technical solutions can be combined in any way to form the optional embodiments of this specification, and will not be described in detail here.

[0161] The technical solution provided in the embodiments of this specification obtains a company recommendation request, which carries the company's information. The company information is structured to obtain product entities and company role entities. Based on the product entities, a query is performed in the industry map to obtain multiple candidate partner companies belonging to the same industry chain as the company. These multiple candidate partner companies are companies that may cooperate with the company. Based on the company role entities, a target partner company is determined from these multiple candidate partner companies; this target partner company is also the partner company recommended to the company. The above-described company recommendation method is highly efficient.

[0162] Figure 6 This is a schematic diagram of the structure of a recommended enterprise device provided in the embodiments of this specification. See also... Figure 6 The device includes: a request acquisition module 601, a structured module 602, a query module 603, and an enterprise determination module 604.

[0163] The request acquisition module 601 is used to acquire enterprise recommendation requests from enterprises. The enterprise recommendation request carries the enterprise information, which represents the enterprise's products and its role in the industry chain. The enterprise recommendation request is used to request recommendations of cooperative enterprises to the enterprise.

[0164] The structuring module 602 is used to structure the enterprise information to obtain the product entities and enterprise role entities in the enterprise information.

[0165] The query module 603 is used to perform a query in the industry map based on the product entity to obtain multiple candidate cooperative enterprises. The industry map includes multiple enterprises belonging to different industry chains, and the multiple candidate cooperative enterprises belong to the same industry chain as the product.

[0166] The enterprise identification module 604 is used to identify target partner enterprises to recommend to the enterprise from among the multiple candidate partner enterprises based on the enterprise role entity.

[0167] In one possible implementation, the structured module 602 is used to perform named entity recognition on the enterprise information to obtain the product entity and the enterprise role entity in the enterprise information.

[0168] In one possible implementation, the structured module 602 is used to input the enterprise information into an entity recognition model. The entity recognition model encodes multiple characters in the enterprise information to obtain the encoded features of the enterprise information. The entity recognition model then decodes these encoded features to obtain labels for each character in the enterprise information; these labels represent the character type of the corresponding character. Based on the character types of these multiple characters, the characters are combined to obtain the product entity and the enterprise role entity.

[0169] In one possible implementation, the query module 603 is used to determine at least one industry chain name entity corresponding to the product entity. Based on the at least one industry chain name entity, a query is performed in the industry map to obtain the multiple candidate cooperative enterprises.

[0170] In one possible implementation, the query module 603 is used to standardize the product entity to obtain at least one standard product entity. A higher-level mapping is then performed on the at least one standard product entity to obtain the at least one industry chain name entity.

[0171] In one possible implementation, the query module 603 is used to perform semantic recognition on the product entity to obtain the semantic features of the product entity. Based on the semantic features of the product entity, semantic matching is performed in a set of standard product entities to obtain at least one standard product entity corresponding to the product entity. The set of standard product entities includes multiple standard product entities.

[0172] In one possible implementation, the enterprise determination module 604 is used to determine at least one target enterprise role entity corresponding to the enterprise role entity, wherein the at least one target enterprise role entity and the enterprise role entity constitute an upstream and downstream relationship. Based on the at least one target enterprise role entity, the enterprise's target cooperative enterprise is determined from the plurality of candidate cooperative enterprises.

[0173] In one possible implementation, the product entity includes a product name entity and a product specification entity. The product name entity is used to obtain the plurality of candidate partner companies. The company determination module 604 is used to match the plurality of candidate partner companies based on the at least one target company role entity to obtain a plurality of first reference partner companies. Based on the product specification entity and the company profiles of the plurality of first reference partner companies, the target partner company is determined from the plurality of first reference partner companies.

[0174] In one possible implementation, the enterprise determination module 604 is used to match the plurality of first reference partner enterprises based on the product specification entity to obtain a plurality of second reference partner enterprises. Based on the enterprise profiles of the plurality of second reference partner enterprises, the target partner enterprise of the enterprise is determined from the plurality of second reference partner enterprises.

[0175] In one possible implementation, the enterprise determination module 604 is used to determine the enterprise score of each of the plurality of second reference partners based on their enterprise profiles. The plurality of second reference partners are then sorted according to their enterprise scores. The top [number] ranked second reference partners are then determined as the enterprise's target partners.

[0176] In one possible implementation, the enterprise determining module 604 is also configured to perform at least one of the following:

[0177] The company's historical partners are identified as its target partners.

[0178] At least one enterprise of the same type as the historical partner enterprise is identified as the target partner enterprise of this enterprise.

[0179] It should be noted that the enterprise recommendation device provided in the above embodiments is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the enterprise recommendation device and the enterprise recommendation method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0180] The technical solution provided in the embodiments of this specification obtains a company recommendation request, which carries the company's information. The company information is structured to obtain product entities and company role entities. Based on the product entities, a query is performed in the industry map to obtain multiple candidate partner companies belonging to the same industry chain as the company. These multiple candidate partner companies are companies that may cooperate with the company. Based on the company role entities, a target partner company is determined from these multiple candidate partner companies; this target partner company is also the partner company recommended to the company. The above-described company recommendation method is highly efficient.

[0181] This specification provides a computer device for performing the above-described method. This computer device can be implemented as a terminal or a server. The structure of a terminal will be described below:

[0182] Figure 7 This is a schematic diagram of the structure of a terminal provided in an embodiment of this specification. The terminal 700 can be a smartphone, tablet computer, laptop computer, or desktop computer. The terminal 700 may also be referred to as user equipment, portable terminal, laptop terminal, desktop terminal, or other names.

[0183] Typically, terminal 700 includes one or more processors 701 and one or more memories 702.

[0184] Processor 701 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. Processor 701 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 701 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 701 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 701 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0185] Memory 702 may include one or more computer-readable storage media, which may be non-transitory. Memory 702 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in memory 702 are used to store at least one computer program for execution by processor 701 to implement the enterprise-recommended method provided in the method embodiments of this specification.

[0186] In some embodiments, the terminal 700 may also optionally include a peripheral device interface 703 and at least one peripheral device. The processor 701, memory 702, and peripheral device interface 703 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 703 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 704, a display screen 705, a camera assembly 706, an audio circuit 707, and a power supply 708.

[0187] Peripheral device interface 703 can be used to connect at least one I / O (Input / Output) related peripheral device to processor 701 and memory 702. In some embodiments, processor 701, memory 702 and peripheral device interface 703 are integrated on the same chip or circuit board; in some other embodiments, any one or two of processor 701, memory 702 and peripheral device interface 703 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0188] The radio frequency (RF) circuit 704 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 704 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 704 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 704 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc.

[0189] Display screen 705 is used to display a user interface (UI). This UI may include graphics, text, icons, video, and any combination thereof. When display screen 705 is a touch display screen, it also has the ability to collect touch signals on or above its surface. These touch signals can be input as control signals to processor 701 for processing. In this case, display screen 705 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard.

[0190] The camera assembly 706 is used to capture images or videos. Optionally, the camera assembly 706 includes a front-facing camera and a rear-facing camera. Typically, the front-facing camera is located on the front panel of the terminal, and the rear-facing camera is located on the back of the terminal.

[0191] The audio circuit 707 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, and convert the sound waves into electrical signals that are input to the processor 701 for processing, or input to the radio frequency circuit 704 to realize voice communication.

[0192] The power supply 708 is used to supply power to the various components in the terminal 700. The power supply 708 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery.

[0193] In some embodiments, the terminal 700 further includes one or more sensors 709. The one or more sensors 709 include, but are not limited to: an accelerometer 710, a gyroscope 711, a pressure sensor 712, an optical sensor 713, and a proximity sensor 714.

[0194] Accelerometer 710 can detect the magnitude of acceleration on the three coordinate axes of a coordinate system established with terminal 700.

[0195] The gyroscope sensor 711 can detect the orientation and rotation angle of the terminal 700. The gyroscope sensor 711 can work in conjunction with the accelerometer sensor 710 to collect the user's 3D movements on the terminal 700.

[0196] The pressure sensor 712 can be installed on the side bezel of the terminal 700 and / or on the lower layer of the display screen 705. When the pressure sensor 712 is installed on the side bezel of the terminal 700, it can detect the user's grip signal on the terminal 700, and the processor 701 can perform left / right hand recognition or quick operation based on the grip signal collected by the pressure sensor 712. When the pressure sensor 712 is installed on the lower layer of the display screen 705, the processor 701 can control the operable controls on the UI interface based on the user's pressure operation on the display screen 705.

[0197] An optical sensor 713 is used to collect ambient light intensity. In one embodiment, the processor 701 can control the display brightness of the display screen 705 based on the ambient light intensity collected by the optical sensor 713.

[0198] The proximity sensor 714 is used to detect the distance between the user and the front of the terminal 700.

[0199] Those skilled in the art will understand that Figure 7 The structure shown does not constitute a limitation on terminal 700, and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0200] The aforementioned computer equipment can also be implemented as a server. The structure of a server is described below:

[0201] Figure 8This is a schematic diagram of a server structure provided in an embodiment of this specification. The server 800 can vary significantly due to differences in configuration or performance. It may include one or more Central Processing Units (CPUs) 801 and one or more memories 802. The one or more memories 802 store at least one computer program, which is loaded and executed by the one or more processors 801 to implement the methods provided in the various method embodiments described above. Of course, the server 800 may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server 800 may also include other components for implementing device functions, which will not be elaborated upon here.

[0202] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including a computer program that can be executed by a processor to perform the enterprise recommendation method described above. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage device, etc.

[0203] In an exemplary embodiment, a computer program product or computer program is also provided, which includes program code stored in a computer-readable storage medium. A processor of a computer device reads the program code from the computer-readable storage medium and executes the program code, causing the computer device to perform the above-described enterprise-recommended method.

[0204] In some embodiments, the computer program described in this specification may be deployed and executed on a single computer device, or on multiple computer devices located in one location, or on multiple computer devices distributed across multiple locations and interconnected via a communication network. These multiple computer devices distributed across multiple locations and interconnected via a communication network may constitute a blockchain system.

[0205] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0206] The above are merely optional embodiments of this specification and are not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification shall be included within the scope of protection of this specification.

Claims

1. A method for enterprise recommendation, the method comprising: Obtain a company recommendation request, the company recommendation request carries the company information, the company information is used to represent the company's products and its role in the industry chain, and the company recommendation request is used to request recommendations of cooperative companies to the company; In the absence of any historical partners, the enterprise information is structured to obtain product entities and enterprise role entities within the enterprise information. Based on the product entity, a query is performed in the industry map to obtain multiple candidate cooperative enterprises. The industry map includes multiple enterprises belonging to different industry chains, and the multiple candidate cooperative enterprises belong to the same industry chain as the enterprise. Based on the enterprise role entity, a target cooperative enterprise recommended to the enterprise is determined from the plurality of candidate cooperative enterprises; The process of querying the industry map based on the product entity yields multiple candidate partner companies, including: Identify at least one industry chain name entity corresponding to the product entity; Based on the at least one industry chain name entity, a query is performed in the industry map to obtain at least one industry chain corresponding to the at least one industry chain name entity; Multiple companies under at least one industrial chain are identified as multiple candidate cooperative companies.

2. The method according to claim 1, wherein structuring the enterprise information to obtain the product entities and enterprise role entities in the enterprise information includes: Named entity recognition is performed on the enterprise information to obtain the product entity and the enterprise role entity in the enterprise information.

3. The method according to claim 2, wherein performing named entity recognition on the enterprise information to obtain the product entity and the enterprise role entity in the enterprise information includes: The enterprise information is input into an entity recognition model. The entity recognition model encodes multiple characters in the enterprise information to obtain the encoded features of the enterprise information. The entity recognition model decodes the encoded features to obtain the label of each character in the enterprise information. The label is used to represent the character type of the corresponding character. The multiple characters are combined based on their character types to obtain the product entity and the enterprise role entity.

4. The method according to claim 1, wherein determining at least one industry chain name entity corresponding to the product entity includes: The product entity is standardized to obtain at least one standard product entity; A higher-level mapping is performed on the at least one standard product entity to obtain the at least one industry chain name entity.

5. The method according to claim 4, wherein standardizing the product entity to obtain at least one standard product entity comprises: Semantic recognition is performed on the product entity to obtain its semantic features; Based on the semantic features of the product entity, semantic matching is performed in a set of standard product entities to obtain at least one standard product entity corresponding to the product entity. The set of standard product entities includes multiple standard product entities.

6. The method according to claim 1, wherein determining the target partner enterprise to recommend to the enterprise from the plurality of candidate partner enterprises based on the enterprise role entity comprises: Identify at least one target enterprise role entity corresponding to the enterprise role entity, wherein the at least one target enterprise role entity and the enterprise role entity constitute an upstream and downstream relationship; Based on the at least one target enterprise role entity, the target cooperative enterprise of the enterprise is determined from the plurality of candidate cooperative enterprises.

7. The method according to claim 6, wherein the product entity includes a product name entity and a product specification entity, the product name entity is used to obtain the plurality of candidate partner companies, and determining the target partner company of the enterprise from the plurality of candidate partner companies based on the at least one target enterprise role entity includes: Based on the at least one target enterprise role entity, multiple first reference cooperative enterprises are obtained by matching among the multiple candidate cooperative enterprises; Based on the product specification entity and the enterprise profiles of the plurality of first reference cooperative enterprises, the target cooperative enterprise of the enterprise is determined from the plurality of first reference cooperative enterprises.

8. The method according to claim 7, wherein determining the target partner of the enterprise from the plurality of reference partner enterprises based on the product specification entity and the enterprise profiles of the plurality of reference partner enterprises comprises: Based on the product specification entity, the plurality of first reference cooperative enterprises are matched to obtain a plurality of second reference cooperative enterprises; Based on the corporate profiles of the plurality of second reference partners, the target partner of the enterprise is determined from the plurality of second reference partners.

9. The method according to claim 8, wherein determining the target partner of the enterprise from the plurality of second reference partner enterprises based on the enterprise profiles of the plurality of second reference partner enterprises comprises: Based on the enterprise profiles of the multiple second reference partners, determine the enterprise score of each second reference partner; The multiple second reference partners are ranked according to their enterprise scores. The first number of second reference partners after sorting are determined as the target partners of the enterprise.

10. The method according to any one of claims 1-9, wherein after obtaining the enterprise recommendation request, the method further comprises at least one of the following: The company's historical partners are identified as the company's target partners. At least one enterprise of the same type as the historical partner enterprise is identified as the target partner enterprise of the enterprise.

11. A business recommendation device, the device comprising: The request acquisition module is used to acquire enterprise recommendation requests from enterprises. The enterprise recommendation requests carry enterprise information, which represents the enterprise's products and its role in the industry chain. The enterprise recommendation requests are used to request recommendations of cooperative enterprises to the enterprise. The structuring module is used to structure the enterprise information when the enterprise has no historical partners, so as to obtain the product entities and enterprise role entities in the enterprise information; The query module is used to query the industry map based on the product entity to obtain multiple candidate cooperative enterprises. The industry map includes multiple enterprises belonging to different industry chains, and the multiple candidate cooperative enterprises belong to the same industry chain as the enterprise. The enterprise determination module is used to determine, based on the enterprise role entity, a target cooperative enterprise to be recommended to the enterprise from among the multiple candidate cooperative enterprises; The process of querying the industry map based on the product entity yields multiple candidate partner companies, including: Identify at least one industry chain name entity corresponding to the product entity; Based on the at least one industry chain name entity, a query is performed in the industry map to obtain at least one industry chain corresponding to the at least one industry chain name entity; Multiple companies under at least one industrial chain are identified as multiple candidate cooperative companies.

12. A computer device comprising one or more processors and one or more memories, wherein at least one computer program is stored in the one or more memories, the computer program being loaded and executed by the one or more processors to implement the enterprise recommendation method as described in any one of claims 1 to 10.

13. A computer-readable storage medium storing at least one computer program, the computer program being loaded and executed by a processor to implement the enterprise recommendation method as described in any one of claims 1 to 10.

14. A computer program product comprising a computer program that, when executed by a processor, implements the enterprise recommendation method according to any one of claims 1 to 10.

Images