Enterprise supply chain recommendation method, device and storage medium based on knowledge representation

Abstract

The present invention discloses a method, device and storage medium for enterprise supply chain recommendation based on knowledge representation. The method includes: collecting the purchase and sale commodity set of the target enterprise, and determining the rough enterprise set related to the target enterprise according to the purchase and sale commodity set; constructing the supply chain network knowledge graph according to the rough enterprise set; determining the knowledge representation of the enterprises in the rough enterprise set according to the supply chain network knowledge graph; determining the target enterprise's set of enterprises to be traded according to the knowledge representation of the enterprises in the rough enterprise set. The technical effect of fine recommendation of supply chain network enterprises is achieved.

Description

Technical Field

[0001] The present invention relates to the field of supply chain technology, and more specifically, to a method, device and storage medium for recommending an enterprise supply chain based on knowledge representation. Background Art

[0002] In their production and operation activities, enterprises provide products and services to upstream and downstream enterprises and users, gradually forming a supply chain system with a network structure. With the continuous development of the economy and the close transactions between enterprises, the supply chain will gradually stabilize. However, due to information constraints and the expansion of enterprise scale, enterprises often encounter bottlenecks in the process of development, resulting in the upstream enterprises being unable to provide raw materials in a timely manner or the downstream enterprises being unable to open sales channels, thereby affecting the development of the enterprises and then affecting the overall economic development and resumption of work and production in the region.

[0003] In the supply chain "enterprise-commodity-enterprise", if the enterprise's transaction object is single and the enterprise transaction chain is too long, it means that the supply chain is weak and the entire supply chain is easily paralyzed due to the operating problems of individual enterprises. Therefore, enterprises need to continuously expand transaction channels, expand the scale of upstream and downstream transaction objects, and enhance the risk resistance of enterprises in order to better enhance the stability of the entire transaction market. Traditional recommendation methods use collaborative filtering or content-based recommendation methods, which can effectively make recommendations based on the user's historical behavior and the potential preferences of multiple users. However, this method has the problems of sparse interaction and cold start, and this method ignores the characteristics of the recommended products themselves. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method, device and storage medium for recommending an enterprise supply chain based on knowledge representation.

[0005] According to one aspect of the present invention, a method for recommending an enterprise supply chain based on knowledge representation is provided, comprising:

[0006] Collect the purchase and sale commodity sets of the target enterprise, and determine the rough set of enterprises related to the target enterprise based on the purchase and sale commodity sets;

[0007] Construct a supply chain network knowledge graph based on the roughly ranked enterprise set;

[0008] According to the supply chain network knowledge graph, determine the knowledge representation of enterprises in the rough-ranked enterprise set;

[0009] According to the knowledge representation of enterprises in the rough-ranked enterprise set, the set of enterprises to be traded of the target enterprise is determined.

[0010] Optionally, the purchase and sales commodity set includes a purchase commodity set and a sales commodity set, and the operation of determining the rough ranking result of the target enterprise according to the purchase and sales commodity set includes:

[0011] Based on the input commodity set, determine the first list of enterprises producing the commodities in the input commodity set;

[0012] According to the sales commodity set, determine the second list of enterprises that purchase the commodities in the sales commodity set;

[0013] Based on the first enterprise list and the second enterprise list, a rough set of enterprises is determined.

[0014] Optionally, the operation of constructing a supply chain network knowledge graph based on the roughly ranked enterprise set includes:

[0015] Based on the target enterprises, the first enterprise list and the second enterprise list, a supply chain network knowledge graph is constructed, wherein the supply chain network knowledge graph includes enterprises, commodities, regions, industries and business scopes.

[0016] Optionally, according to the supply chain network knowledge graph, an operation of determining the knowledge representation of the enterprises in the rough-ranked enterprise set includes:

[0017] According to the supply chain network knowledge graph, determine the upstream and downstream comprehensive representation vectors of enterprises in the rough enterprise set;

[0018] According to the characteristic information of the enterprises in the rough-ranked enterprise set, the characteristic vectors of the enterprises in the rough-ranked enterprise set are determined;

[0019] According to the upstream and downstream comprehensive representation vectors and feature vectors of the enterprises in the rough-ranked enterprise set, the knowledge representation of the enterprises in the rough-ranked enterprise set is determined.

[0020] Optionally, according to the supply chain network knowledge graph, the operation of determining the upstream and downstream comprehensive representation vectors of the enterprises in the rough-ranked enterprise set includes:

[0021] The Word2Vec model is used to convert the upstream and downstream node information of the enterprises in the rough-ranked enterprise set into vectors, and the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set is determined;

[0022] According to the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set, the upstream and downstream comprehensive characterization vectors of the enterprises in the rough-ranked enterprise set are determined.

[0023] Optionally, the characteristic information of the enterprises in the rough-ranked enterprise set includes enterprise scale, enterprise operating conditions, enterprise credit level, enterprise capital, enterprise age, enterprise staff and enterprise revenue, and the operation of determining the characteristic vectors of the enterprises in the rough-ranked enterprise set based on the characteristic information of the enterprises in the rough-ranked enterprise set includes:

[0024] One-hot encoding is used to encode the feature information separately and then concatenated to determine the feature vectors of the enterprises in the rough-ranked enterprise set.

[0025] Optionally, the method further comprises:

[0026] The rough-ranked enterprise set is divided according to the supply chain network knowledge graph to determine the transaction enterprise set and the enterprise set to be recommended, where the transaction enterprise set refers to the enterprises that have a transaction relationship with the target enterprise.

[0027] Optionally, the operation of determining the set of enterprises to be traded of the target enterprise according to the knowledge representation of the enterprises in the roughly ranked enterprise set includes:

[0028] Sum the knowledge representations of the enterprises in the transaction enterprise set to determine the vector sum of the knowledge representations of the transaction enterprises;

[0029] The knowledge representation vector of the transaction enterprise is summed with the knowledge representation of the enterprises in the enterprise set to be recommended, and the enterprise characteristic knowledge representation vectors of the enterprises in the enterprise set to be recommended are determined respectively;

[0030] The enterprise feature knowledge representation vectors are respectively input into a pre-set multi-layer perceptron for calculation, and the recommendation probability of the enterprise in the set of recommended enterprises is respectively determined;

[0031] The recommendation probabilities of the enterprises in the set of recommended enterprises are ranked, and a set of enterprises to be traded with the target enterprise, which has a predetermined number of transactions, is determined.

[0032] According to another aspect of the present invention, there is provided an enterprise supply chain recommendation device based on knowledge representation, comprising:

[0033] The first determination module is used to collect the purchase and sale commodity set of the target enterprise, and determine the rough ranking enterprise set related to the target enterprise based on the purchase and sale commodity set;

[0034] The construction module is used to construct a supply chain network knowledge graph based on a rough set of enterprises;

[0035] The second determination module is used to determine the knowledge representation of enterprises in the rough-ranked enterprise set based on the supply chain network knowledge graph;

[0036] The third determination module is used to determine the set of target enterprises to be traded based on the knowledge representation of the enterprises in the rough-ranked enterprise set.

[0037] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is used to execute the method described in any one of the above aspects of the present invention.

[0038] According to another aspect of the present invention, an electronic device is provided, comprising: a processor; a memory for storing instructions executable by the processor; the processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of the above aspects of the present invention.

[0039] Therefore, the enterprise supply chain recommendation method based on knowledge representation provided by the present invention collects the purchase and sale commodity set of the target enterprise, and determines the rough-ranked enterprise set related to the target enterprise according to the purchase and sale commodity set; constructs the supply chain network knowledge graph according to the rough-ranked enterprise set; determines the knowledge representation of the enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph; determines the target enterprise's set of enterprises to be traded according to the knowledge representation of the enterprises in the rough-ranked enterprise set. Thus, through the two steps of the recall method based on the transaction commodity association query and the accurate sorting based on the knowledge representation, the potential trading partners of each enterprise in the market are predicted, and the technical effect of refined recommendation of supply chain network enterprises is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] A more complete understanding of exemplary embodiments of the present invention may be obtained by referring to the following drawings:

[0041] Figure 1 is a flow chart of a method for recommending an enterprise supply chain based on knowledge representation provided by an exemplary embodiment of the present invention;

[0042] Figure 2 is a schematic diagram of a supply chain network knowledge graph provided by an exemplary embodiment of the present invention;

[0043] Figure 3 It is a schematic diagram of a supply chain enterprise recommendation process based on knowledge representation provided by an exemplary embodiment of the present invention;

[0044] Figure 4 is a schematic diagram of the structure of an enterprise supply chain recommendation device based on knowledge representation provided by an exemplary embodiment of the present invention;

[0045] Figure 5 This is a structure of an electronic device provided by an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0046] Below, the exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described here.

[0047] It should be noted that the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.

[0048] Those skilled in the art can understand that the terms "first" and "second" in the embodiments of the present invention are only used to distinguish different steps, devices or modules, etc., and neither represent any specific technical meaning nor indicate the necessary logical order between them.

[0049] It should also be understood that, in the embodiments of the present invention, “plurality” may refer to two or more than two, and “at least one” may refer to one, two or more than two.

[0050] It should also be understood that any component, data or structure mentioned in the embodiments of the present invention can generally be understood as one or more, unless explicitly limited or otherwise indicated in the context.

[0051] In addition, the term "and / or" in the present invention is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects before and after are in an "or" relationship.

[0052] It should also be understood that the description of the various embodiments of the present invention focuses on the differences between the various embodiments, and the same or similar aspects thereof can be referenced to each other, and for the sake of brevity, they will not be described one by one.

[0053] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0054] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0055] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0056] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0057] Embodiments of the present invention can be applied to electronic devices such as terminal devices, computer systems, servers, etc., which can operate with many other general or special computing system environments or configurations. Examples of well-known terminal devices, computing systems, environments and / or configurations suitable for use with electronic devices such as terminal devices, computer systems, servers, etc. include, but are not limited to: personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network personal computers, small computer systems, large computer systems, and distributed cloud computing technology environments including any of the above systems, etc.

[0058] Electronic devices such as terminal devices, computer systems, servers, etc. can be described in the general context of computer system executable instructions (such as program modules) executed by computer systems. Generally, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in a distributed cloud computing environment, where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules can be located on local or remote computing system storage media including storage devices.

[0059] Exemplary Methods

[0060] Figure 1 FIG. 1 is a flow chart of a method for recommending enterprise supply chains based on knowledge representation provided by an exemplary embodiment of the present invention. This embodiment can be applied to electronic devices, such as Figure 1 As shown, the enterprise supply chain recommendation method 100 based on knowledge representation includes the following steps:

[0061] Step 101, collect the purchase and sale commodity set of the target enterprise, and determine the rough set of enterprises related to the target enterprise based on the purchase and sale commodity set.

[0062] Optionally, the purchase and sales commodity set includes a purchase commodity set and a sales commodity set, and the operation of determining the rough ranking result of the target enterprise according to the purchase and sales commodity set includes:

[0063] Based on the input commodity set, determine the first list of enterprises producing the commodities in the input commodity set;

[0064] According to the sales commodity set, determine the second list of enterprises that purchase the commodities in the sales commodity set;

[0065] Based on the first enterprise list and the second enterprise list, a rough set of enterprises is determined.

[0066] Specifically, first, based on the enterprise transaction information in the entire supply chain network, the set of purchase and sales items of the enterprise to be recommended is obtained. Based on the enterprise's transaction commodity information, the list of enterprises that produce and purchase the enterprise's purchase and sales items in the entire transaction market is queried to form a potential recall customer set (the process of screening the initial potential object set of the enterprise in the recommendation is called recall). That is, the purchase and sales item sets of the enterprise to be recommended are collected to form I (Cn) and O (Cn) sub-tables to represent the enterprise's purchase and sales item sets, and the commodities in the purchase and sales item sets are queried respectively to obtain the list of enterprises that produce I (Cn) and purchase O (Cn) commodities in the entire market transaction network, forming a rough ranking result of the enterprise to be recommended.

[0067] Step 102, construct a supply chain network knowledge graph based on the roughly ranked enterprise set.

[0068] Optionally, the operation of constructing a supply chain network knowledge graph based on the roughly ranked enterprise set includes:

[0069] Based on the target enterprises, the first enterprise list and the second enterprise list, a supply chain network knowledge graph is constructed, wherein the supply chain network knowledge graph includes enterprises, commodities, regions, industries and business scopes.

[0070] Specifically, refer to Figure 2 As shown in Figure 1, a supply chain network knowledge graph is constructed, which includes five types of nodes: enterprise, commodity, region, industry, and business scope. The simple structure diagram of the supply chain knowledge graph is as follows: Figure 3 As shown (due to the large area of ​​the spectrum Figure 3 Only the knowledge graph with enterprise A as the central node is shown.) For example, if the rough enterprise set contains 1,000 enterprises, then a supply chain network knowledge graph is drawn based on the target enterprise and the 1,000 enterprises.

[0071] Therefore, for the supply chain network, a supply chain knowledge graph is constructed, which includes node information such as enterprises, commodities, regions, industries, as well as basic enterprise information such as enterprise qualification type, enterprise credit level, and enterprise scale.

[0072] Step 103, determining the knowledge representation of the enterprises in the rough-ranked enterprise set based on the supply chain network knowledge graph.

[0073] Optionally, according to the supply chain network knowledge graph, an operation of determining the knowledge representation of the enterprises in the rough-ranked enterprise set includes:

[0074] According to the supply chain network knowledge graph, determine the upstream and downstream comprehensive representation vectors of enterprises in the rough enterprise set;

[0075] According to the characteristic information of the enterprises in the rough-ranked enterprise set, the characteristic vectors of the enterprises in the rough-ranked enterprise set are determined;

[0076] According to the upstream and downstream comprehensive representation vectors and feature vectors of the enterprises in the rough-ranked enterprise set, the knowledge representation of the enterprises in the rough-ranked enterprise set is determined.

[0077] Optionally, according to the supply chain network knowledge graph, the operation of determining the upstream and downstream comprehensive representation vectors of the enterprises in the rough-ranked enterprise set includes:

[0078] The Word2Vec model is used to convert the upstream and downstream node information of the enterprises in the rough-ranked enterprise set into vectors, and the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set is determined;

[0079] According to the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set, the upstream and downstream comprehensive characterization vectors of the enterprises in the rough-ranked enterprise set are determined.

[0080] Optionally, the characteristic information of the enterprises in the rough-ranked enterprise set includes enterprise scale, enterprise operating conditions, enterprise credit level, enterprise capital, enterprise age, enterprise staff and enterprise revenue, and the operation of determining the characteristic vectors of the enterprises in the rough-ranked enterprise set based on the characteristic information of the enterprises in the rough-ranked enterprise set includes:

[0081] One-hot encoding is used to encode the feature information separately and then concatenated to determine the feature vectors of the enterprises in the rough-ranked enterprise set.

[0082] Specifically, refer to Figure 2 and Figure 3 As shown in the figure, for each enterprise, the knowledge representation of the enterprise is composed of the basic characteristics of the enterprise itself, such as enterprise scale, enterprise operating conditions, enterprise credit level, etc., and the embedding (low-dimensional vector) mean of the upstream and downstream nodes of the enterprise. The characteristic information of the enterprise itself is composed of one-hot encoding, and the one-hot encoding of each feature is w. Multiple feature information is spliced ​​into the enterprise's own embedding information W = [w1, w2, w3, ..., w n ], the node information of commodities, regions, industries, etc. is represented by Word2Vec, and the upstream and downstream comprehensive representation method of enterprises is as shown in formula (1):

[0083]

[0084] Where e is the enterprise node to be recommended, contextnode(e) is the node that has direct connection with the enterprise in the knowledge graph, and e i is the enterprise node, n i These are non-enterprise nodes such as products and regions.

[0085] In addition, the purpose of Word2Vec is to convert text into vectors. This patent uses the open source general Word2Vec tool, and the conversion result is the vector representation of ei and ni in Formula 1. context(e) is the modulus, and its function is normalization.

[0086] After linear transformation of the enterprise’s own embedding information and the enterprise’s upstream and downstream comprehensive embedding information, the final enterprise representation e(t) is:

[0087]

[0088] Where g(e) is the mapping, g(e)=Me, M is the mapping matrix, which can be obtained through training. For isolated nodes, g(e) is set to 0.

[0089] Thus, a new way to represent supply chain network nodes is constructed, which completes the embedding representation of the enterprise through the node's own knowledge and the aggregation of adjacent nodes.

[0090] Step 104, determining the set of enterprises to be traded of the target enterprise based on the knowledge representation of the enterprises in the rough-ranked enterprise set.

[0091] Optionally, the method further comprises:

[0092] The rough-ranked enterprise set is divided according to the supply chain network knowledge graph to determine the transaction enterprise set and the enterprise set to be recommended, where the transaction enterprise set refers to the enterprises that have a transaction relationship with the target enterprise.

[0093] Specifically, refer to Figure 2 As shown in the figure, for example, if enterprise A is the target enterprise, then enterprises B, C, D, E, F, and G are the transaction enterprise set, where the enterprises in the transaction enterprise set are the enterprises that have a transaction relationship with the target enterprise. Therefore, the transaction enterprise set and the enterprise set to be recommended can be clearly divided through the supply chain network knowledge graph. For example, the rough enterprise set contains 1,000 enterprises, and there are 100 enterprises that have a transaction relationship with the target enterprise. At this time, the transaction enterprise set is 100, and the enterprise set to be recommended is 900.

[0094] Optionally, the operation of determining the set of enterprises to be traded of the target enterprise according to the knowledge representation of the enterprises in the roughly ranked enterprise set includes:

[0095] Sum the knowledge representations of the enterprises in the transaction enterprise set to determine the vector sum of the knowledge representations of the transaction enterprises;

[0096] The knowledge representation vector of the transaction enterprise is summed with the knowledge representation of the enterprises in the enterprise set to be recommended, and the enterprise characteristic knowledge representation vectors of the enterprises in the enterprise set to be recommended are determined respectively;

[0097] The enterprise feature knowledge representation vectors are respectively input into a pre-set multi-layer perceptron for calculation, and the recommendation probability of the enterprise in the set of recommended enterprises is respectively determined;

[0098] The recommendation probabilities of the enterprises in the set of recommended enterprises are ranked, and a set of enterprises to be traded with the target enterprise, which has a predetermined number of transactions, is determined.

[0099] Specifically, refer to Figure 3As shown in the figure, the W obtained by enterprise knowledge representation is used as the dual-channel input of the convolutional neural network CNN to obtain the embedded representation of all enterprises, and the enterprises in the recall set are summed with the existing transaction enterprises of the recommended enterprise one by one to obtain the attention of the candidate enterprise for each transaction enterprise of the target enterprise. The attention is an attention mechanism, and its purpose is to find the enterprise features (such as credit, capital, etc.) that the target enterprise is most interested in during the training process.

[0100] Among them, attention is reflected in the form of summation, adding the pooling layer result [1,1,1,0,1,0] of the recalled enterprise and the transaction enterprise vector [1,1,0,0,1,1] of the target enterprise to become [2,2,1,0,2,1].

[0101] Here, splicing means concatenating two vectors into one vector, such as concatenating [1,1,1,0,1,0] and [1,1,0,0,1,1] to become [1,1,1,0,1,0,1,1,0,0,1,1].

[0102] For example, the rough-ranked enterprise set has 10,000 enterprises, the target enterprise has 100 trading enterprises, and the enterprise set to be recommended has 900 enterprises. The comprehensive characterization of enterprises requires the upstream and downstream comprehensive characterization of the 900 enterprises in the recommended enterprise set and the 100 trading enterprises in the target enterprise set that have trading relations. Each enterprise has a characterization vector W.

[0103] The summed embedding representation is sent to the MLP multi-layer perceptron for calculation, or the recommendation probability of each enterprise in the recall set of the target enterprise is sorted according to the probability. For example, the top 100 enterprises can be returned as the final accurate recommendation result, that is, the set of enterprises to be traded. For example, each enterprise in the recall set (the set of enterprises to be recommended) is a potential recommendation object. Each enterprise in the recall set is represented, convolved, and pooled to form a vector, which is concatenated with the vector calculated by the attention mechanism. For example, (1,1,1,0,1) and (1,1,0,1) are concatenated to (1,1,1,0,1,1,1,0,1). This vector is put into the MLP calculation to obtain a value between 0 and 1. The value represents the degree of recommendation. For example, the probability value of a roughly ranked enterprise is 0.95, indicating that it is a high-quality enterprise that needs to be recommended to the target enterprise. If the probability is 0.1, it means that this roughly ranked enterprise is not recommended to be recommended to the target enterprise. Sort by this value and recommend the top N (for example, the top 100) enterprises to the target enterprise.

[0104] Thus, rough recommendation (recall) is accomplished through product similarity calculation, and refined enterprise recommendation is accomplished through knowledge representation using a multi-layer perceptron.

[0105] Construct a supply chain transaction network, calculate the enterprise's own knowledge representation based on the enterprise's own knowledge and the enterprise's transaction commodity information, regional information, industry information, etc., and finally use a multi-layer perceptron to calculate the recommendation degree between different enterprises to complete the refined recommendation of supply chain network enterprises.

[0106] Therefore, from the perspective of broadening the supply chain network and enhancing the stability of enterprise development, the present invention extracts data such as enterprises and commodities in the market to construct a supply chain knowledge graph, obtains the dual-channel feature information of the enterprise by fusing the basic information of the enterprise and the knowledge of the enterprise-related nodes, and uses neural network algorithms such as CNN and MLP to complete the extraction and construction of enterprise representation. Finally, the upstream and downstream enterprises closest to the target enterprise in the candidate set are calculated and recalled to complete refined enterprise transaction recommendations, which can effectively help enterprises broaden upstream and downstream sales channels, enhance market stability, and help the government speed up the resumption of work and production.

[0107] Exemplary Devices

[0108] Figure 4 FIG. 1 is a schematic diagram of a structure of a knowledge representation-based enterprise supply chain recommendation device provided by an exemplary embodiment of the present invention. Figure 4 As shown, the device 400 includes:

[0109] The first determination module 410 is used to collect the purchase and sale commodity set of the target enterprise, and determine the rough ranking enterprise set related to the target enterprise according to the purchase and sale commodity set;

[0110] A construction module 420 is used to construct a supply chain network knowledge graph based on the roughly ranked enterprise set;

[0111] The second determination module 430 is used to determine the knowledge representation of the enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph;

[0112] The third determination module 440 is used to determine the set of target enterprise to be traded based on the knowledge representations of the enterprises in the roughly ranked enterprise set.

[0113] Optionally, the purchase and sales commodity set includes a purchase commodity set and a sales commodity set, and the first determination module 410 includes:

[0114] The first determination submodule is used to determine the first list of enterprises producing the commodities in the input commodity set according to the input commodity set;

[0115] The second determination submodule is used to determine, based on the sales commodity set, a second list of enterprises that purchase commodities in the sales commodity set;

[0116] The third determination submodule is used to determine a rough set of enterprises according to the first enterprise list and the second enterprise list.

[0117] Optionally, the construction module 420 includes:

[0118] A sub-module is constructed to construct a supply chain network knowledge graph based on the target enterprise, the first enterprise list and the second enterprise list, wherein the supply chain network knowledge graph includes enterprises, commodities, regions, industries and business scopes.

[0119] Optionally, the second determining module 430 includes:

[0120] The fourth determination submodule is used to determine the upstream and downstream comprehensive representation vectors of enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph;

[0121] A fifth determination submodule, used to determine a feature vector of an enterprise in the rough-ranked enterprise set according to feature information of the enterprise in the rough-ranked enterprise set;

[0122] The sixth determination submodule is used to determine the knowledge representation of the enterprises in the rough-ranked enterprise set according to the upstream and downstream comprehensive representation vectors and feature vectors of the enterprises in the rough-ranked enterprise set.

[0123] Optionally, the fourth determining submodule includes:

[0124] The first determination unit is used to perform vector conversion on the upstream and downstream node information of the enterprises in the rough-ranked enterprise set by using the Word2Vec model to determine the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set;

[0125] The second determining unit is used to determine the upstream and downstream comprehensive characterization vectors of the enterprises in the rough-ranked enterprise set according to the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set.

[0126] Optionally, the characteristic information of the enterprises in the roughly ranked enterprise set includes enterprise scale, enterprise operating conditions, enterprise credit level, enterprise capital, enterprise years, enterprise staff and enterprise revenue, and the fifth determination submodule includes:

[0127] The third determining unit is used to encode the feature information separately by one-hot encoding and then concatenate them to determine the feature vectors of the enterprises in the rough-ranked enterprise set.

[0128] Optionally, the method further comprises:

[0129] The fourth determination module is used to divide the rough enterprise set according to the supply chain network knowledge graph, determine the transaction enterprise set and the enterprise set to be recommended, where the transaction enterprise set is the enterprise that has a transaction relationship with the target enterprise.

[0130] Optionally, the third determining module 440 includes:

[0131] A seventh determination submodule is used to sum the knowledge representations of the enterprises in the transaction enterprise set to determine the vector sum of the transaction enterprise knowledge representations;

[0132] An eighth determination submodule is used to sum the transaction enterprise knowledge representation vector and the knowledge representation of the enterprises in the enterprise set to be recommended, respectively, to determine the enterprise characteristic knowledge representation vectors of the enterprises in the enterprise set to be recommended;

[0133] A ninth determination submodule is used to input the enterprise feature knowledge representation vectors into a pre-set multi-layer perceptron for calculation, and respectively determine the recommendation probability of an enterprise in the set of recommended enterprises;

[0134] The tenth determination submodule is used to sort the recommendation probabilities of the enterprises in the set of recommended enterprises, and determine a predetermined number of sets of enterprises to be traded with the target enterprise.

[0135] Exemplary Electronic Devices

[0136] Figure 5 The electronic device provided by an exemplary embodiment of the present invention may be any one or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the collected input signals from them. Figure 5 FIG. 2 is a block diagram of an electronic device according to an embodiment of the present invention. Figure 5 As shown, the electronic device 500 includes one or more processors 501 and a memory 502 .

[0137] The processor 501 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0138] The memory 502 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 501 may run the program instructions to implement the method for information mining of historical change records and / or other desired functions of the software program of each embodiment of the present invention described above. In one example, the electronic device may also include: an input device 503 and an output device 504, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown).

[0139] In addition, the input device 503 may also include, for example, a keyboard, a mouse, and the like.

[0140] The output device 504 can output various information to the outside, and can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto.

[0141] Of course, to simplify, Figure 5 Only some of the components related to the present invention in the electronic device are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application conditions, the electronic device may also include any other appropriate components.

[0142] Exemplary computer program products and computer-readable storage media

[0143] In addition to the above-mentioned methods and devices, an embodiment of the present invention may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method for performing information mining on historical change records according to various embodiments of the present invention described in the above "Exemplary Method" section of this specification.

[0144] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the embodiments of the present invention, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0145] In addition, an embodiment of the present invention may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enables the processor to execute the steps of the method for information mining of historical change records according to various embodiments of the present invention described in the above “Exemplary Method” section of this specification.

[0146] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, system or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0147] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details disclosed above are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0148] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0149] The block diagrams of the devices, systems, equipment, and systems involved in the present invention are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, systems, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The word "such as" used here refers to the phrase "such as but not limited to", and can be used interchangeably with it.

[0150] The method and system of the present invention may be implemented in many ways. For example, the method and system of the present invention may be implemented by software, hardware, firmware or any combination of software, hardware, firmware. The above order of steps for the method is only for illustration, and the steps of the method of the present invention are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present invention may also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present invention. Thus, the present invention also covers a recording medium storing a program for executing the method according to the present invention.

[0151] It should also be noted that in the system, device and method of the present invention, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. The above description of the disclosed aspects is provided to enable any technician in the field to make or use the present invention. Various modifications to these aspects are very obvious to those skilled in the art, and the general principles defined here can be applied to other aspects without departing from the scope of the present invention. Therefore, the present invention is not intended to be limited to the aspects shown here, but in accordance with the widest range consistent with the principles and novel features disclosed here.

[0152] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present invention to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for recommending enterprise supply chains based on knowledge representation, characterized in that: include: Collecting the purchase and sale commodity set of the target enterprise, and determining a rough set of enterprises related to the target enterprise based on the purchase and sale commodity set; Constructing a supply chain network knowledge graph based on the rough-ranked enterprise set; Determining knowledge representations of enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph; Determining a set of enterprises to be traded of the target enterprise according to the knowledge representations of the enterprises in the rough-ranked enterprise set; The method further comprises: The rough-ranked enterprise set is divided according to the supply chain network knowledge graph to determine a transaction enterprise set and a to-be-recommended enterprise set, wherein the transaction enterprise set is an enterprise that has a transaction relationship with the target enterprise; wherein, The operation of determining the set of enterprises to be traded of the target enterprise according to the knowledge representation of the enterprises in the rough-ranked enterprise set includes: Summing the knowledge representations of the enterprises in the transaction enterprise set to determine a vector sum of the transaction enterprise knowledge representations; Summing the transaction enterprise knowledge representation vector and the knowledge representation of the enterprises in the enterprise set to be recommended, respectively, to determine the enterprise characteristic knowledge representation vectors of the enterprises in the enterprise set to be recommended; Inputting the enterprise feature knowledge representation vectors into a pre-set multi-layer perceptron for calculation respectively, and determining the recommendation probability of the enterprises in the set of enterprises to be recommended respectively; The recommendation probabilities of the enterprises in the set of enterprises to be recommended are sorted to determine a predetermined number of sets of enterprises to be traded with the target enterprise.

2. The method according to claim 1, characterized in that The purchase and sale commodity set includes a purchase commodity set and a sales commodity set, and the operation of determining the rough ranking result of the target enterprise according to the purchase and sale commodity set includes: Determine, based on the imported commodity set, a list of first enterprises producing commodities in the imported commodity set; Determining, based on the sales item set, a second list of enterprises that purchase the products in the sales item set; The rough-ranked enterprise set is determined based on the first enterprise list and the second enterprise list.

3. The method according to claim 2, characterized in that The operation of constructing a supply chain network knowledge graph according to the rough-sorted enterprise set includes: The supply chain network knowledge graph is constructed based on the target enterprise, the first enterprise list and the second enterprise list, wherein the supply chain network knowledge graph includes enterprises, commodities, regions, industries and business scopes.

4. The method according to claim 1, characterized in that The operation of determining the knowledge representation of the enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph includes: Determine the upstream and downstream comprehensive representation vectors of the enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph; Determining feature vectors of the enterprises in the roughly ranked enterprise set according to feature information of the enterprises in the roughly ranked enterprise set; The knowledge representation of the enterprises in the rough-ranked enterprise set is determined according to the upstream and downstream comprehensive representation vectors and the feature vectors of the enterprises in the rough-ranked enterprise set.

5. The method according to claim 4, characterized in that The operation of determining the upstream and downstream comprehensive representation vectors of the enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph includes: Using the Word2Vec model to perform vector conversion on the upstream and downstream node information of the enterprises in the rough-ranked enterprise set, and determining the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set; According to the vector information of the upstream and downstream nodes of the enterprises in the rough-ranked enterprise set, the upstream and downstream comprehensive characterization vectors of the enterprises in the rough-ranked enterprise set are determined.

6. The method according to claim 4, characterized in that The characteristic information of the enterprises in the rough-ranked enterprise set includes enterprise scale, enterprise operating conditions, enterprise credit level, enterprise capital, enterprise age, enterprise staff and enterprise revenue, and the operation of determining the characteristic vector of the enterprises in the rough-ranked enterprise set based on the characteristic information of the enterprises in the rough-ranked enterprise set includes: The feature information is encoded separately using one-hot encoding and then concatenated to determine the feature vectors of the enterprises in the rough-ranked enterprise set.

7. An enterprise supply chain recommendation device based on knowledge representation, characterized in that: include: A first determination module is used to collect the purchase and sale commodity set of the target enterprise, and determine a rough ranking enterprise set related to the target enterprise based on the purchase and sale commodity set; A construction module, used to construct a supply chain network knowledge graph based on the rough-ranked enterprise set; A second determination module is used to determine the knowledge representation of the enterprises in the rough-ranked enterprise set according to the supply chain network knowledge graph; A third determination module is used to determine the set of enterprises to be traded of the target enterprise according to the knowledge representations of the enterprises in the rough-ranked enterprise set; The fourth determination module is used to divide the rough enterprise set according to the supply chain network knowledge graph to determine a transaction enterprise set and a to-be-recommended enterprise set, wherein the transaction enterprise set is an enterprise that has a transaction relationship with the target enterprise; wherein, The third determination module includes: A seventh determination submodule is used to sum the knowledge representations of the enterprises in the transaction enterprise set to determine a vector sum of the transaction enterprise knowledge representations; An eighth determination submodule is used to sum the transaction enterprise knowledge representation vector and the knowledge representations of the enterprises in the enterprise set to be recommended, respectively, to determine the enterprise characteristic knowledge representation vectors of the enterprises in the enterprise set to be recommended; A ninth determination submodule is used to input the enterprise feature knowledge representation vectors into a preset multi-layer perceptron for calculation, and respectively determine the recommendation probability of the enterprises in the set of enterprises to be recommended; The tenth determination submodule is used to sort the recommendation probabilities of the enterprises in the set of enterprises to be recommended, and determine a predetermined number of sets of enterprises to be traded with the target enterprise.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the method according to any one of claims 1 to 6.

9. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is used to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1 to 6.

Images