Supplier's recommended methods and related devices

Abstract

The present application discloses a supplier recommendation method and related devices in the field of artificial intelligence. In the technical solution provided by the present application, the characteristic information of the enterprise to be recommended and the characteristic information of each of the multiple suppliers are input into the supplier recommendation model to obtain the supplier recommendation information of the enterprise to be recommended, thereby improving the accuracy of recommending suppliers for the enterprise. Furthermore, the neural network model is trained using the characteristic information of each of the multiple enterprises and the characteristic information of the suppliers that have a contractual relationship with each of the multiple enterprises to obtain the supplier recommendation model, thereby improving the accuracy of the supplier recommendation model.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a supplier recommendation method and related devices. Background Art

[0002] Artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is also the study of the design principles and implementation methods of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making. Research in the field of artificial intelligence includes robotics, natural language processing, computer vision, decision-making and reasoning, human-computer interaction, recommendation and search, basic AI theory, etc.

[0003] In the context of supply chain management, if an enterprise wants to achieve efficient and sustainable development, it must ensure the quality of its products and raw materials. Therefore, for an enterprise, the selection of suppliers is very critical, and how to choose a supplier that meets the needs of the enterprise among many suppliers has become an important issue.

[0004] There are two main methods for recommending suppliers in the existing technology. One is to collect contract data between enterprises and suppliers through manual research and screen and exclude them one by one. This method consumes a lot of manpower and material resources, is inefficient, and is easily affected by subjective factors of researchers, resulting in low accuracy of recommendation results. The other is to use information processing technology to screen according to the needs of the enterprise in big data to obtain suppliers that the system believes best meet the needs of the enterprise. However, this processing method is relatively simple and only starts from the requirements of the enterprise for suppliers. It is difficult to accurately obtain suppliers that meet the needs of the enterprise. The enterprise needs to conduct evaluation and analysis later, and the accuracy is low.

[0005] Therefore, how to improve the accuracy of suppliers recommended to enterprises has become an urgent problem to be solved. Summary of the invention

[0006] This application proposes a supplier recommendation method and related devices in the field of artificial intelligence, which improves the accuracy of recommending suppliers to enterprises.

[0007] In a first aspect, the present application provides a method for recommending suppliers, the method comprising: obtaining characteristic information of an enterprise to be recommended; obtaining characteristic information of each of a plurality of suppliers; inputting the characteristic information of the enterprise to be recommended and the characteristic information of each of the suppliers into a supplier recommendation model to obtain supplier recommendation information of the enterprise to be recommended, the supplier recommendation model being a neural network model trained based on the characteristic information of the enterprise and the characteristic information of the supplier.

[0008] In this method, the characteristic information of the enterprise to be recommended and the characteristic information of each of the multiple suppliers are input into the supplier recommendation model to obtain the supplier recommendation information of the enterprise to be recommended, and the supplier is recommended to the enterprise based on the supplier recommendation information. Compared with recommending suppliers to enterprises through manual research and using information processing technology to screen suppliers according to the needs of enterprises in big data, the accuracy of recommending suppliers to enterprises is improved.

[0009] In one possible implementation, the supplier recommendation model includes a graph convolutional neural network model and a classification model.

[0010] In a possible implementation, the characteristic information of the enterprise includes one or more of the following information: business needs, marketing model, region, cost budget, and payment method.

[0011] In a possible implementation, the characteristic information of the supplier includes one or more of the following information: product type, product price, production cycle, payment method, and region.

[0012] In a second aspect, the present application provides a method for training a supplier recommendation model, the method comprising: obtaining characteristic information of each of a plurality of enterprises and characteristic information of suppliers having a contractual relationship with each of the plurality of enterprises; training a neural network model using the characteristic information of each of the plurality of enterprises and the characteristic information of suppliers having a contractual relationship with each of the plurality of enterprises to obtain the supplier recommendation model.

[0013] In this method, the neural network model is trained using the characteristic information of each of the multiple enterprises and the characteristic information of the suppliers having a contractual relationship with each of the multiple enterprises to obtain a supplier recommendation model, thereby improving the accuracy of the supplier recommendation model.

[0014] In one possible implementation, the neural network model includes a graph convolutional neural network model and a classification model; wherein, the use of feature information of each of the multiple enterprises and feature information of suppliers having a contractual relationship with each of the multiple enterprises to train the neural network model includes: inputting the feature information of each of the multiple enterprises and the feature information of suppliers having a contractual relationship with each of the multiple enterprises into the graph convolutional neural network model to obtain feature vectors of each of the multiple enterprises and feature vectors of suppliers having a contractual relationship with each of the multiple enterprises; inputting the feature vectors of each of the multiple enterprises and the feature vectors of suppliers having a contractual relationship with each of the multiple enterprises into the classification model to obtain classification results; and adjusting the parameters of the graph convolutional neural network model and the classification model based on the classification results.

[0015] In this implementation, the characteristic information of each of the multiple enterprises and the characteristic information of the suppliers having a contractual relationship with each of the multiple enterprises are input into the graph convolutional neural network model to obtain the characteristic vector of each of the multiple enterprises and the characteristic vector of the suppliers having a contractual relationship with each of the multiple enterprises; the characteristic vector of each of the multiple enterprises and the characteristic vector of the suppliers having a contractual relationship with each of the multiple enterprises are input into the classification model to obtain the classification result, and the graph convolutional neural network model and the classification model are trained based on the classification result, thereby further improving the accuracy of the supplier recommendation model.

[0016] In a possible implementation, the inputting feature information of each of the multiple enterprises and the feature information of the suppliers having a contractual relationship with each of the multiple enterprises into the graph convolutional neural network model to obtain the feature vector of each of the multiple enterprises and the feature vector of the suppliers having a contractual relationship with each of the multiple enterprises includes: obtaining invoice information of each of the multiple enterprises and invoice information of the suppliers having a contractual relationship with each of the multiple enterprises; calculating the embedding vector of each of the multiple enterprises based on the invoice information of each of the multiple enterprises; calculating the embedding vector of the suppliers having a contractual relationship with each of the multiple enterprises based on the invoice information of the suppliers having a contractual relationship with each of the multiple enterprises; inputting the embedding vector of each of the multiple enterprises and the embedding vector of the suppliers having a contractual relationship with each of the multiple enterprises into the graph convolutional neural network model to obtain the feature vector of each of the multiple enterprises and the feature vector of the suppliers having a contractual relationship with each of the multiple enterprises.

[0017] In this implementation, pre-processing learning is performed on the acquired invoice information of each of the multiple enterprises and the invoice information of suppliers having a contractual relationship with each of the multiple enterprises to obtain feature vectors of each of the multiple enterprises and the feature vectors of suppliers having a contractual relationship with each of the multiple enterprises, thereby improving the accuracy of the feature vectors of each of the multiple enterprises and the feature vectors of suppliers having a contractual relationship with each of the multiple enterprises, and at the same time improving the accuracy of the supplier recommendation model.

[0018] In a possible implementation, the characteristic information of the enterprise includes one or more of the following information: business needs, marketing model, region, cost budget, and payment method.

[0019] In a possible implementation, the characteristic information of the supplier includes one or more of the following information: product type, product price, production cycle, payment method, and region.

[0020] In a third aspect, the present application provides a supplier recommendation device, which may include various modules for implementing the method in the first aspect, and these modules may be implemented in software and / or hardware.

[0021] In a fourth aspect, the present application provides a training device for a supplier recommendation model, which may include modules for implementing the method in the second aspect, and these modules may be implemented in software and / or hardware.

[0022] In a fifth aspect, the present application provides a supplier recommendation device. The device may include a processor coupled to a memory. The memory is used to store program code, and the processor is used to execute the program code in the memory to implement the method in the first aspect or any one of the implementations thereof.

[0023] Optionally, the device may further include the memory.

[0024] In a sixth aspect, the present application provides a training device for a supplier recommendation model. The device may include a processor coupled to a memory. The memory is used to store program code, and the processor is used to execute the program code in the memory to implement the method in the second aspect or any one of the implementations thereof.

[0025] Optionally, the device may further include the memory.

[0026] In the seventh aspect, the present application provides a chip comprising at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, and the at least one processor is used to run a computer program or instruction to execute the method described in the first aspect or the second aspect or any possible implementation method therein.

[0027] In an eighth aspect, the present application provides a computer-readable medium storing a program code for execution by a device, wherein the program code includes a method for executing the method described in the first aspect or the second aspect or any possible implementation method thereof.

[0028] In a ninth aspect, the present application provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the method as described in the first aspect or the second aspect or any possible implementation thereof.

[0029] In a tenth aspect, the present application provides a computing device comprising at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, the communication interface communicates with a target system, and the at least one processor is used to run a computer program or instruction to execute a method as described in the first aspect or the second aspect or any possible implementation method thereof.

[0030] In the eleventh aspect, the present application provides a computing system comprising at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, the communication interface communicates with a target system, and the at least one processor is used to run a computer program or instruction to execute the method described in the first aspect or the second aspect or any possible implementation method thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A schematic diagram of a system architecture of an embodiment of the present application;

[0032] Figure 2 A schematic diagram of a flow chart of a supplier recommendation method according to an embodiment of the present application;

[0033] Figure 3 A flowchart of a method for training a supplier recommendation model according to an embodiment of the present application;

[0034] Figure 4 A flowchart of another method for training a supplier recommendation model according to an embodiment of the present application;

[0035] Figure 5 A schematic structural diagram of a supplier recommendation device according to an embodiment of the present application;

[0036] Figure 6 A schematic structural diagram of a training device for a supplier recommendation model according to an embodiment of the present application;

[0037] Figure 7 A schematic diagram of the structure of an apparatus provided in one embodiment of the present application. DETAILED DESCRIPTION

[0038] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0039] In order to facilitate the understanding of the present application, the concepts involved in the present application are first explained.

[0040] Graph convolutional network (GCN): refers to a neural network that can perform deep learning on graph data. Among them, convolutional neural network refers to a feedforward neural network that contains convolution calculations and has a deep structure. Graph convolutional neural network is a natural extension of convolutional neural network in the graph domain.

[0041] Embedding vector: refers to a real number vector used to represent the attributes of things (such as companies and suppliers). Embedding vectors are widely used in the field of machine learning for predictive reasoning. For example, the matching degree between a company and a supplier can be predicted through the embedding vector of a company and the embedding vector of a supplier. Embedding vectors are continuously optimized and learned during the model training process.

[0042] Figure 1 Schematic diagram of a system architecture of an embodiment of the present application. Figure 1 The data collection device 160 is used to collect characteristic information of the enterprise and the characteristic information of the supplier and store them in the database 130; the training device 120 generates a supplier recommendation model 101 based on the characteristic information of the enterprise and the characteristic information of the supplier maintained in the database 130, wherein the supplier recommendation model can also be called a supplier recommendation rule.

[0043] The method for the training device 120 to obtain the supplier recommendation model 101 based on the characteristic information of the enterprise and the characteristic information of the supplier can be found in Figure 3 to Figure 4 The embodiment shown in any figure.

[0044] The supplier recommendation model 101 obtained by the training device 120 can be applied to different systems or devices, for example, applied to the execution device 110 .

[0045] The execution device 110 is configured with an I / O interface 112 for data exchange with external devices. The user can input feature information of the enterprise to be recommended into the I / O interface 112 through the client device 140 .

[0046] The execution device 110 may call data, codes, etc. in the data storage system 150 , and may also store data, instructions, etc. in the data storage system 150 .

[0047] The calculation module 111 uses the supplier recommendation model 101 to process the characteristic information of the enterprise to be recommended and the characteristic information of the supplier, thereby obtaining the supplier recommendation information of the enterprise to be recommended.

[0048] Finally, the I / O interface 112 returns the processing result to the client device 140 to provide it to the user.

[0049] exist Figure 1 In the case shown, the user can manually specify the data to be input into the execution device 110, for example, by operating in the interface provided by the I / O interface 112. In another case, the client device 140 can automatically input data into the I / O interface 112 and obtain the result. If the automatic input of data by the client device 140 requires the authorization of the user, the user can set the corresponding authority in the client device 140. The user can view the result output by the execution device 110 in the client device 140, and the specific presentation form can be a specific method such as display, sound, action, etc. The client device 140 can also serve as a data collection terminal to store the collected characteristic information of the enterprise and the characteristic information of the supplier into the database 130.

[0050] It is worth noting that Figure 1 This is only a schematic diagram of a system architecture provided by an embodiment of the present application. The positional relationship between the devices, components, modules, etc. shown in the figure does not constitute any limitation. For example, in Figure 1 In the embodiment, the data storage system 150 is an external memory relative to the execution device 110. In other cases, the data storage system 150 may also be placed in the execution device 110.

[0051] Figure 2 FIG. 1 is a flow chart of a supplier recommendation method according to an embodiment of the present application. Figure 2 As shown, the method at least includes S201 to S203.

[0052] S201, obtaining characteristic information of the enterprise to be recommended.

[0053] The enterprises to be recommended can also be called core enterprises, including enterprises that occupy a core position in the supply chain and have payment responsibility for the raw materials, goods or services provided by their upstream first-tier suppliers.

[0054] Exemplarily, the characteristic information of the enterprise to be recommended may include one or more of the following information: business needs, marketing model, region, cost budget and payment method, etc.

[0055] S202, obtaining characteristic information of each supplier among the multiple suppliers.

[0056] Suppliers can include direct suppliers and indirect suppliers. Direct suppliers can also be called 1st-level suppliers, including enterprises that directly provide raw materials, goods or services to core enterprises on credit; indirect suppliers can also be called N-level suppliers, including enterprises that directly provide raw materials, goods or services to their downstream suppliers on credit. N-level suppliers are indirect suppliers of core enterprises, and N is a positive integer greater than or equal to 2.

[0057] Exemplarily, the multiple suppliers may include suppliers with whom the enterprise to be recommended has the intention to cooperate and / or suppliers with whom the enterprise to be recommended has the intention to cooperate, etc.

[0058] Exemplarily, the characteristic information of the supplier may include one or more of the following information: product type, product price, production cycle, payment method, and region, etc.

[0059] S203, input the characteristic information of the enterprise to be recommended and the characteristic information of each supplier into the supplier recommendation model to obtain the supplier recommendation information of the enterprise to be recommended. The supplier recommendation model is a neural network model trained based on the characteristic information of the enterprise and the characteristic information of the supplier.

[0060] Before inputting the characteristic information of the enterprise to be recommended and the characteristic information of each supplier into the supplier recommendation model, it is necessary to first determine the embedding vector of the enterprise to be recommended and the embedding vector of each supplier based on the characteristic information of the enterprise to be recommended and the characteristic information of each supplier.

[0061] Exemplarily, a method for determining an embedding vector of an enterprise to be recommended based on characteristic information of the enterprise to be recommended may include: extracting enterprise fields from the characteristic information of the enterprise to be recommended and performing preprocessing, the preprocessing including calculating the missing rate of each enterprise field and removing enterprise fields with a missing rate higher than a preset missing rate; determining the type of the enterprise field, if the enterprise field is a sequence type, performing one-hot processing on the enterprise field, and if the enterprise field is a numeric type, performing normalization processing on the enterprise field.

[0062] The embedding vector in the form of One-hot may include N data values, which may include data value 0 and data value 1. The enterprise field is input into the best XGBoost model to obtain the feature importance ranking of the enterprise field, wherein the best XGBoost model is obtained by training the XGBoost algorithm. The enterprise fields whose rankings are greater than the preset ranking in the feature importance ranking are connected to obtain the embedding vector of the enterprise to be recommended.

[0063] Exemplarily, a method for determining a supplier's embedding vector based on the supplier's characteristic information may include: obtaining all supplier fields used to distinguish suppliers, performing unified dimension processing on the values ​​corresponding to each supplier field, and obtaining characteristic values ​​of the supplier field; sorting the values ​​of all supplier fields in ascending order to obtain an ascending supplier field queue; arranging the characteristic values ​​corresponding to each supplier field in the ascending supplier field queue in reverse order to the supplier field queue to obtain a reverse characteristic value queue; connecting each supplier field in the ascending supplier field queue with each characteristic value in the reverse characteristic value queue of the corresponding serial number to obtain the supplier's embedding vector.

[0064] After determining the embedding vector of the enterprise to be recommended and the embedding vector of each supplier according to the characteristic information of the enterprise to be recommended and the characteristic information of each supplier, the embedding vector of the enterprise to be recommended and the embedding vector of each supplier are input into the supplier recommendation model to obtain the supplier recommendation information of the enterprise to be recommended.

[0065] In one possible implementation, the supplier recommendation model includes a graph convolutional neural network model and a classification model.

[0066] As an example, the embedding vector of the enterprise to be recommended and the embedding vector of each supplier are input into the trained graph convolutional neural network model to obtain the feature vector of the enterprise to be recommended and the feature vector of each supplier, and then the feature vector of the enterprise to be recommended and the feature vector of each supplier are input into the trained classification model to obtain the supplier recommendation information of the enterprise to be recommended.

[0067] Exemplarily, the method of inputting the feature vector of the enterprise to be recommended and the feature vector of each supplier into the trained classification model to obtain the supplier recommendation information of the enterprise to be recommended may include: respectively calculating the inner product of the feature vector of the enterprise to be recommended and the feature vector of each supplier, and the supplier corresponding to the supplier whose inner product with the feature vector of the enterprise to be recommended is greater than or equal to the recommendation threshold β is the supplier recommended to the enterprise to be recommended, and the supplier recommendation information of the enterprise to be recommended includes all the suppliers recommended to the enterprise to be recommended. Among them, the recommendation threshold β is a parameter of the classification model, which can be obtained by training the classification model.

[0068] The technical solution provided by the present application inputs the characteristic information of the enterprise to be recommended and the characteristic information of each of multiple suppliers into a supplier recommendation model, obtains the supplier recommendation information of the enterprise to be recommended, and recommends suppliers to the enterprise based on the supplier recommendation information, thereby improving the accuracy of recommending suppliers to the enterprise.

[0069] Figure 3 FIG. 1 is a flow chart of a method for training a supplier recommendation model according to an embodiment of the present application. Figure 3 As shown, the method at least includes S301 to S302.

[0070] S301, acquiring characteristic information of each of the multiple enterprises and characteristic information of suppliers having a contractual relationship with each of the multiple enterprises.

[0071] The multiple enterprises may include enterprises that have contract data information with one or more suppliers. The contract data information between an enterprise and a supplier is data information indicating that the enterprise has a contract relationship with the supplier. The contract relationship between an enterprise and a supplier means that within the validity period of the contract data information between the two parties, the supplier is a supplier of the enterprise and provides raw materials, goods or services to the enterprise.

[0072] Each of the multiple enterprises has at least one supplier with which it has a contractual relationship. A supplier can have contractual relationships with multiple enterprises. Any of the multiple enterprises can be a supplier with which another enterprise has a contractual relationship, and any of the multiple suppliers can be an enterprise with which another supplier has a contractual relationship.

[0073] Exemplarily, the characteristic information of an enterprise may include one or more of the following information: business needs, marketing model, region, cost budget, payment method, etc.

[0074] Exemplarily, the characteristic information of the supplier may include one or more of the following information: product type, product price, production cycle, payment method, and region, etc.

[0075] S302, using the characteristic information of each of the multiple enterprises and the characteristic information of suppliers having a contractual relationship with each of the multiple enterprises to train a neural network model to obtain a supplier recommendation model.

[0076] Before using the characteristic information of each enterprise among the multiple enterprises and the characteristic information of the suppliers having a contractual relationship with each enterprise among the multiple enterprises to train the neural network and obtain the supplier recommendation model, it is necessary to first determine the embedding vector of each enterprise and the embedding vector of the supplier having a contractual relationship with each enterprise based on the characteristic information of each enterprise among the multiple enterprises and the characteristic information of the suppliers having a contractual relationship with each enterprise.

[0077] It should be noted that the method for determining the embedding vector of each enterprise and the embedding vector of the supplier having a contractual relationship with each enterprise based on the characteristic information of each enterprise and the characteristic information of the supplier having a contractual relationship with each enterprise can be referred to S203, which will not be repeated here.

[0078] In one possible implementation, the supplier recommendation model includes a graph convolutional neural network model and a classification model.

[0079] As an example, the feature information of each of the multiple enterprises and the feature information of the suppliers having a contractual relationship with each of the multiple enterprises are input into the graph convolutional neural network model to obtain the feature vectors of each of the multiple enterprises and the feature vectors of the suppliers having a contractual relationship with each of the multiple enterprises; the feature vectors of each of the multiple enterprises and the feature vectors of the suppliers having a contractual relationship with each of the multiple enterprises are input into the classification model to obtain the classification results; and the parameters of the graph convolutional neural network model and the classification model are adjusted based on the obtained classification results and the actual contract results between the enterprises and the suppliers.

[0080] Exemplarily, a loss function is used to calculate the loss value between the classification result obtained by the supplier recommendation model and the actual contract result between the enterprise and the supplier, and the weight parameters in the graph neural network model and the classification model are adjusted based on the loss value.

[0081] The technical solution provided in this application uses the characteristic information of each of a plurality of enterprises and the characteristic information of suppliers having a contractual relationship with each of the plurality of enterprises to train a neural network model to obtain a supplier recommendation model, thereby improving the accuracy of the supplier recommendation model.

[0082] Figure 4 FIG. 1 is a flow chart of another method for training a supplier recommendation model according to an embodiment of the present application. Figure 4 As shown, the method at least includes S401 to S404.

[0083] S401, obtaining invoice information of each of the multiple enterprises and invoice information of suppliers having a contractual relationship with each of the multiple enterprises.

[0084] Invoice information is used to represent the business voucher data issued and collected by enterprises and suppliers when buying and selling goods, providing or receiving services, and engaging in other business activities. Invoice information can include the name, quality, agreed price, and prepayment amount of the product or service provided.

[0085] More abundant characteristic information of the enterprise and the supplier can be obtained from the invoice information.

[0086] S402, calculating an embedding vector of each of the multiple enterprises according to the invoice information of each of the multiple enterprises.

[0087] First, the feature information of each enterprise is extracted from its invoice information, and then the embedding vector of the enterprise is calculated based on the extracted feature information.

[0088] In a possible implementation, a term frequency-inverse document frequency (TF-IDF) method is used to extract characteristic information of the enterprise from the invoice information of the enterprise.

[0089] As an example, the invoice information of an enterprise is segmented to obtain multiple text information data, the word frequency and inverse file frequency of each text information data are calculated, the weight of the text information data is calculated according to the word frequency and inverse file frequency of each text information data, the text information data in the invoice information of each enterprise is sorted from large to small according to the calculated weight, and the characteristic vector of the enterprise is selected according to the weight sorting. For example, the text information data with the top 10 weight values ​​are used as the characteristic information of the enterprise.

[0090] Exemplarily, the word frequency of text information data can be calculated by the following formula:

[0091]

[0092] Among them, tf w Represents the word frequency of the text data information, M represents the number of times the text information data appears in the invoice information, and N represents the number of all text information data in the invoice information.

[0093] Exemplarily, the inverse file frequency of text information data can be calculated by the following formula:

[0094]

[0095] Among them, idf w It represents the inverse file frequency of the text information data, Q represents the number of all invoice information, and S represents the number of all invoice information including the text information data.

[0096] For example, tf can be calculated w *idf w Get the weight of each text information data.

[0097] It should be noted that the method for calculating the embedding vector of an enterprise according to the characteristic information of the enterprise can refer to S203, which will not be described in detail here.

[0098] S403, calculating an embedding vector of a supplier having a contractual relationship with each of the multiple enterprises according to the invoice information of the supplier having a contractual relationship with each of the multiple enterprises.

[0099] The method of calculating the embedding vector of the supplier having a contractual relationship with each of the multiple enterprises according to the invoice information of the supplier having a contractual relationship with each of the multiple enterprises can be referred to S402, which will not be described in detail here.

[0100] S404, using the embedding vector of each enterprise among the multiple enterprises and the embedding vector of the supplier having a contractual relationship with each enterprise among the multiple enterprises to train the neural network model, to obtain a supplier recommendation model.

[0101] It should be noted that the method of training the neural network model using the embedding vector of each of the multiple enterprises and the embedding vector of the supplier having a contractual relationship with each of the multiple enterprises to obtain the supplier recommendation model can be referred to S302 and will not be repeated here.

[0102] The technical solution provided by the present application obtains invoice information of each of a plurality of enterprises and invoice information of suppliers having contractual relationships with each of the plurality of enterprises for preprocessing and learning, thereby obtaining an embedding vector of each of the plurality of enterprises and an embedding vector of suppliers having contractual relationships with each of the plurality of enterprises, thereby improving the accuracy of the embedding vector of each of the plurality of enterprises and the embedding vector of suppliers having contractual relationships with each of the plurality of enterprises, and using the embedding vectors of the enterprises and the embedding vectors of the suppliers to train a supplier recommendation model, thereby improving the accuracy of the supplier recommendation model.

[0103] Figure 5 FIG. 1 is a schematic structural diagram of a supplier recommendation device according to an embodiment of the present application. Figure 5 As shown, the apparatus 500 may include an acquisition module 501 and an input module 502 .

[0104] Any module in the acquisition module and the input module in the embodiment of the present application can be implemented in whole or in part by software and / or hardware. Among them, the part implemented by software can be run on a processor to realize the corresponding function, and the part implemented by hardware can be a component of the processor.

[0105] In one implementation, the apparatus 500 may be used to implement Figure 2 For example, the acquisition module 501 can be used to implement S201 and S202, and the input module 502 can be used to implement S203.

[0106] Figure 6 This is a schematic structural diagram of a training device for a supplier recommendation model according to an embodiment of the present application. Figure 6 As shown, the apparatus 600 may include an acquisition module 601 and a training module 602 .

[0107] Any module in the acquisition module and the training module in the embodiment of the present application can be implemented in whole or in part by software and / or hardware. Among them, the part implemented by software can be run on a processor to implement the corresponding function, and the part implemented by hardware can be a component of the processor.

[0108] In one implementation, the apparatus 600 may be used to implement Figure 3 For example, the acquisition module 601 can be used to implement S301, and the training module 602 can be used to implement S302.

[0109] In another possible implementation, the apparatus 600 may further include a processing module. The apparatus 600 in this implementation may be used to implement Figure 4 For example, the acquisition module 601 can be used to implement S401, the processing module can be used to implement S402 to S403, and the training module 602 can be used to implement S404.

[0110] Figure 7 A schematic diagram of the structure of an apparatus provided in one embodiment of the present application. Figure 7 The device 700 shown can be used to execute the method described in any of the above embodiments.

[0111] like Figure 7 As shown, the device 700 of this embodiment includes: a memory 701, a processor 702, a communication interface 703 and a bus 704. The memory 701, the processor 702 and the communication interface 703 are connected to each other through the bus 704.

[0112] The memory 701 may be a read only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM). The memory 701 may store a program. When the program stored in the memory 701 is executed by the processor 702, the processor 702 may be used to execute Figures 2 to 4The individual steps of the method are shown.

[0113] Processor 702 can adopt a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits to execute relevant programs to implement the supplier recommendation method and the supplier recommendation model training method of the method embodiment of the present application.

[0114] The processor 702 may also be an integrated circuit chip with signal processing capability. In the implementation process, each step of the method of each embodiment of the present application may be completed by an integrated logic circuit of hardware in the processor 702 or an instruction in software form.

[0115] The processor 702 may also be a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The methods, steps, and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. The general purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0116] The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 701, and the processor 702 reads the information in the memory 701 and completes the functions required to be executed by each method in the embodiments of the present application in combination with its hardware. For example, it can execute Figures 2 to 4 The various steps / functions of the illustrated embodiment.

[0117] The communication interface 703 may use, but is not limited to, a transceiver or other transceiver device to implement communication between the apparatus 700 and other devices or a communication network.

[0118] The bus 704 may include a path for transmitting information between various components of the device 700 (eg, the memory 701 , the processor 702 , and the communication interface 703 ).

[0119] It should be understood that the device 700 shown in the embodiment of the present application may be an electronic device, or may also be a chip configured in the electronic device.

[0120] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0121] It should be understood that the term "and / or" in this article is only a description of the association relationship of 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. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0122] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can be represented by: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0123] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0124] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for recommending a supplier, characterized in that: The method comprises: Obtain characteristic information of the enterprise to be recommended; Obtain characteristic information of each of the multiple suppliers; Inputting the characteristic information of the enterprise to be recommended and the characteristic information of each supplier into a supplier recommendation model to obtain the supplier recommendation information of the enterprise to be recommended, wherein the supplier recommendation model is a neural network model trained based on the characteristic information of the enterprise and the characteristic information of the supplier; the supplier recommendation model includes a graph convolutional neural network model and a classification model; Inputting the characteristic information of the enterprise to be recommended and the characteristic information of each supplier into the supplier recommendation model to obtain the supplier recommendation information of the enterprise to be recommended, including: obtaining the embedding vector of the enterprise to be recommended according to the characteristic information of the enterprise to be recommended, and obtaining the embedding vector of each supplier according to the characteristic information of each supplier; inputting the embedding vector of the enterprise to be recommended and the embedding vector of each supplier into the graph convolutional neural network model to obtain the characteristic vector of the enterprise to be recommended and the characteristic vector of each supplier, and inputting the characteristic vector of the enterprise to be recommended and the characteristic vector of each supplier into the classification model to obtain the supplier recommendation information of the enterprise to be recommended; The step of obtaining the embedding vector of the enterprise to be recommended according to the characteristic information of the enterprise to be recommended includes: Extract enterprise fields from the feature information of the enterprise to be recommended; determine the feature importance ranking of the enterprise fields; connect the enterprise fields whose rankings are greater than a preset ranking in the feature importance ranking to obtain an embedding vector of the enterprise to be recommended; The acquiring the embedding vector of each supplier according to the characteristic information of each supplier includes: Determine the supplier fields that can distinguish suppliers, unify the dimensions of the values ​​corresponding to each supplier field, and obtain the eigenvalues ​​of the supplier fields; sort the values ​​of all supplier fields in ascending order to obtain an ascending supplier field queue; arrange the eigenvalues ​​corresponding to each supplier field in the ascending supplier field queue in the reverse order of the supplier field queue to obtain a reverse eigenvalue queue; connect each supplier field in the ascending supplier field queue with each eigenvalue in the reverse eigenvalue queue of the corresponding sequence number to obtain the supplier's embedding vector.

2. The method according to claim 1, characterized in that The characteristic information of the enterprise includes one or more of the following information: business needs, marketing model, region, cost budget and payment method.

3. The method according to claim 1 or 2, characterized in that: The characteristic information of the supplier includes one or more of the following information: product type, product price, production cycle, payment method and region.

4. A method for training a supplier recommendation model, characterized in that: The method comprises: Acquire characteristic information of each of the plurality of enterprises and characteristic information of suppliers having a contractual relationship with each of the plurality of enterprises; Acquire invoice information of each of the multiple enterprises and invoice information of suppliers having a contractual relationship with each of the multiple enterprises; Calculate an embedding vector for each of the multiple enterprises based on the invoice information of each of the multiple enterprises; Calculate the embedding vector of the supplier having a contractual relationship with each of the multiple enterprises according to the invoice information of the supplier having a contractual relationship with each of the multiple enterprises; Inputting the embedding vector of each of the multiple enterprises and the embedding vector of the supplier having a contractual relationship with each of the multiple enterprises into a graph convolutional neural network model to obtain a feature vector of each of the multiple enterprises and a feature vector of the supplier having a contractual relationship with each of the multiple enterprises; Inputting the feature vector of each of the multiple enterprises and the feature vector of the supplier having a contractual relationship with each of the multiple enterprises into a classification model to obtain a classification result; Adjusting parameters of the graph convolutional neural network model and the classification model based on the classification result; The step of calculating the embedding vector of each of the plurality of enterprises according to the invoice information of each of the plurality of enterprises comprises: Extracting enterprise fields from the invoice information of the enterprise to be recommended; determining the feature importance ranking of the enterprise fields; connecting the enterprise fields whose rankings are greater than a preset ranking in the feature importance ranking to obtain an embedding vector of the enterprise to be recommended; The step of calculating the embedding vector of the supplier having a contractual relationship with each of the multiple enterprises according to the invoice information of the supplier having a contractual relationship with each of the multiple enterprises comprises: The supplier fields that can distinguish suppliers are determined from the supplier's invoice information, and the values ​​corresponding to each supplier field are processed in a unified dimension to obtain the eigenvalues ​​of the supplier fields; the values ​​of all supplier fields are sorted in ascending order to obtain an ascending supplier field queue; the eigenvalues ​​corresponding to each supplier field in the ascending supplier field queue are arranged in the reverse order of the supplier field queue to obtain a reverse eigenvalue queue; each supplier field in the ascending supplier field queue is connected with each eigenvalue in the reverse eigenvalue queue of the corresponding serial number to obtain the supplier's embedding vector.

5. The method according to claim 4, characterized in that The characteristic information of the enterprise includes one or more of the following information: business needs, marketing model, region, cost budget and payment method.

6. The method according to claim 4, characterized in that The characteristic information of the supplier includes one or more of the following information: product type, product price, production cycle, payment method and region.

7. A supplier recommendation device, characterized in that: The invention comprises various functional modules for implementing the method according to any one of claims 1 to 3.

8. A training device for a supplier recommendation model, characterized in that: The method comprises various functional modules for implementing the method according to any one of claims 4 to 6.

9. A computer-readable medium, characterized in that The computer-readable medium stores a program code for computer execution, wherein the program code includes instructions for executing the method according to any one of claims 1 to 3 or instructions for executing the method according to any one of claims 4 to 6.

10. A computer program product, comprising a computer program, which, when executed by a processor, implements the instructions of the method according to any one of claims 1 to 3 or the instructions of the method according to any one of claims 4 to 6.

Images