Product recommendation method and device and electronic equipment

Abstract

The embodiment of the invention provides a product recommendation method and device, electronic equipment and a readable storage medium, and relates to the technical field of data processing. According to the embodiment of the invention, by constructing the logistics behavior relation data, after the to-be-recommended product and the logistics behavior relation data of the to-be-recommended product are obtained, the to-be-recommended product can be recommended according to the posting relation between the associated product associated with the to-be-recommended product in the logistics behavior relation data and each candidate recommendation object; and determining the relevancy between each candidate recommendation object and the to-be-recommended product, obtaining a target recommendation object according to the relevancy between each candidate recommendation object and the to-be-recommended product, and recommending the to-be-recommended product to the target recommendation object. Thus, based on the posting relationship between the associated product and each candidate recommendation object, the object having potential requirements on the product can be obtained, so that a more comprehensive recommendation object can be obtained, the comprehensiveness of recommendation is improved, and the accuracy and rationality of product recommendation can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a product recommendation method, device and electronic device. Background Art

[0002] With the development of network technology, more and more users are shopping online. In order to improve the user shopping experience, product recommendations are generally made to users. Most existing product recommendation methods generate product portraits based on product data, and then select matching users based on the product portraits and make recommendations to matching users. However, this method only makes recommendations based on product data, and the data is relatively single, resulting in low accuracy and comprehensiveness of product recommendations. Summary of the invention

[0003] Based on the above research, embodiments of the present invention provide a product recommendation method, device and electronic device, which can improve the comprehensiveness and accuracy of product recommendations.

[0004] The embodiments of the present invention can be implemented in the following ways:

[0005] In a first aspect, an embodiment of the present invention provides a product recommendation method, the method comprising:

[0006] Acquire the product to be recommended and the logistics behavior relationship data of the product to be recommended; the logistics behavior relationship data includes the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object;

[0007] Determine the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object;

[0008] According to the relevance between each of the candidate recommendation objects and the product to be recommended, a target recommendation object is obtained, and the product to be recommended is recommended to the target recommendation object.

[0009] In an optional implementation manner, the consignment relationship between the associated product associated with the to-be-recommended product and each of the candidate recommendation objects includes the number of consignments;

[0010] The determining the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated product associated with the product to be recommended and each candidate recommendation object comprises:

[0011] For each of the candidate recommendation objects, determine a weight coefficient of a target-related product corresponding to the consignment relationship of the candidate recommendation object and a consignment frequency index value of the target-related product;

[0012] Obtaining a consignment frequency coefficient of the target-related product according to the number of consignments of the target-related product by the candidate recommendation object and the consignment frequency index value of the target-related product;

[0013] The relevance between the candidate recommendation object and the product to be recommended is determined according to the weight coefficient and the consignment frequency coefficient of the target-related product.

[0014] In an optional implementation manner, the determining of the weight coefficient of the target associated product corresponding to the consignment relationship of the candidate recommendation object and the consignment frequency index value of the target associated product includes:

[0015] Determine the inferred product scenario corresponding to the target-related product, and obtain a weight coefficient of the target-related product according to the inferred product scenario; different inferred product scenarios have different weight coefficients;

[0016] According to the consignment relationship between the target-related product and each of the candidate recommendation objects, determine the number of consignments of each of the candidate recommendation objects to the target-related product;

[0017] According to the number of times each of the candidate recommendation objects has consigned the target-related product, an index value of the number of times the target-related product has been consigned is obtained.

[0018] In an optional implementation manner, before obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the method further includes:

[0019] Obtaining logistics information data; the logistics information data includes the products consigned by each object and the number of consignments of the consigned products;

[0020] Determine the object logistics behavior relationship data according to the products entrusted by each object and the number of times the entrusted products are entrusted; the object logistics behavior relationship data includes the consignment relationship between each product and each object; each consignment relationship includes the number of times the object corresponding to the consignment relationship consigns the product corresponding to the consignment relationship;

[0021] For each of the products, the associated products associated with the product are determined, and the object logistics behavior relationship data is filtered according to the product and the associated products associated with the product to obtain the logistics behavior relationship data of the product.

[0022] In an optional implementation manner, for each of the products, determining the associated products associated with the product includes:

[0023] Acquire product transaction data, where the product transaction data includes products traded by each object;

[0024] For each product, according to the product transaction data, a target transaction object having a demand for the product is obtained;

[0025] Determine the product consigned by the target transaction object according to the logistics information data;

[0026] According to the product entrusted by the target transaction object, the associated products associated with the product are obtained.

[0027] In an optional embodiment, filtering the object logistics behavior relationship data according to the product and the associated products associated with the product to obtain the logistics behavior data of the product includes:

[0028] For each of the products, obtaining a presumed product scenario of the product and a filtering rule corresponding to each of the presumed product scenarios; each of the presumed product scenarios includes associated products of the product under the presumed product scenario;

[0029] According to the product and the products associated with the product, the object logistics behavior relationship data is filtered to obtain the initial logistics behavior relationship data of the product;

[0030] The initial logistics behavior relationship data is filtered according to the filtering rules corresponding to each of the inferred product scenarios to obtain the logistics behavior relationship data of the product.

[0031] In an optional implementation manner, before acquiring, for each of the products, the inferred product scenario of the product and the filtering rules corresponding to each of the inferred product scenarios, the method further includes:

[0032] For each of the inferred product scenarios of the product, determine, based on the logistics information data, a consignment object that has a consignment relationship with the product and associated products under the inferred product scenario;

[0033] Determine the number of times each consignment object consigns the product and the associated products under the inferred product scenario within a set time period;

[0034] According to the number of times each of the consignment objects consigns the product within a set time period and consigns the associated products under the inferred product scenario, a filtering rule for the product under the inferred product scenario is obtained.

[0035] In a second aspect, an embodiment of the present invention provides a product recommendation method, the method comprising:

[0036] Acquire a target recommendation object for which product recommendation is required and logistics behavior relationship data of the target recommendation object; the logistics behavior relationship data includes the consignment relationship between the target recommendation object and each candidate recommendation product;

[0037] Determining the relevance between the target recommendation object and each of the candidate recommendation products according to the consignment relationship between the target recommendation object and each of the candidate recommendation products;

[0038] According to the relevance between the target recommendation object and each of the candidate recommendation products, a target recommendation product is obtained, and the target recommendation product is recommended to the target recommendation object.

[0039] In a third aspect, an embodiment of the present invention provides a product recommendation device, the product recommendation device comprising:

[0040] A data acquisition module, used to acquire the product to be recommended and the logistics behavior relationship data of the product to be recommended; the logistics behavior relationship data includes the consignment relationship between the associated products associated with the product to be recommended and each of the candidate recommendation objects;

[0041] A data processing module, configured to determine the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object;

[0042] The product recommendation module is used to obtain a target recommendation object according to the relevance between each candidate recommendation object and the product to be recommended, and recommend the product to be recommended to the target recommendation object.

[0043] In a fourth aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the product recommendation method described in any of the above-mentioned embodiments when executing the computer program.

[0044] In a fifth aspect, an embodiment of the present invention provides a readable storage medium, wherein the readable storage medium includes a computer program, and when the computer program is executed, the electronic device where the readable storage medium is located is controlled to execute the product recommendation method described in any of the above-mentioned embodiments.

[0045] The embodiments of the present invention provide a product recommendation method, device and electronic device. By constructing logistics behavior relationship data, after obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined based on the consignment relationship between the associated products associated with the product to be recommended in the logistics behavior relationship data and each candidate recommendation object, and the target recommendation object is obtained based on the correlation between each candidate recommendation object and the product to be recommended, and the product to be recommended is recommended to the target recommendation object. In this way, based on the consignment relationship between the associated product and each candidate recommendation object, an object with potential demand for the product can be obtained, so that a more comprehensive recommendation object can be obtained, and the comprehensiveness of the recommendation can be improved. At the same time, the embodiments of the present invention can obtain the real needs of customers through the logistics behavior relationship data, so that product recommendations based on the logistics behavior relationship data can improve the accuracy and rationality of product recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The technical solution and other beneficial effects of the present application will be made apparent by describing in detail the specific implementation methods of the present application in conjunction with the accompanying drawings.

[0047] Figure 1 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0048] Figure 2 A schematic diagram of a flow chart of a product recommendation method provided in an embodiment of the present invention.

[0049] Figure 3 A relationship diagram of object logistics behavior relationship data provided by an embodiment of the present invention.

[0050] Figure 4 A relationship diagram of initial logistics behavior relationship data provided by an embodiment of the present invention.

[0051] Figure 5 A relationship diagram of logistics behavior relationship data provided by an embodiment of the present invention.

[0052] Figure 6 A relationship diagram of logistics behavior relationship data provided by an embodiment of the present invention.

[0053] Figure 7 A block diagram of a product recommendation device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below 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, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0055] In the description of the present application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0056] The disclosure below provides many different embodiments or examples to realize the different structures of the present application. In order to simplify the disclosure of the present application, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present application. In addition, the present application can repeat reference numbers and / or reference letters in different examples, and this repetition is for the purpose of simplification and clarity, which itself does not indicate the relationship between the various embodiments and / or settings discussed. In addition, the various specific processes and material examples provided by the present application, but those of ordinary skill in the art can appreciate the application of other processes and / or the use of other materials.

[0057] As described in the background technology, most existing product recommendation methods generate product portraits based on product data, and then select matching users based on the product portraits and recommend products to matching users. However, this method only recommends based on product data, and the data is relatively single, resulting in low accuracy of product recommendations.

[0058] Based on the above research, the embodiments of the present invention provide a product recommendation method, device and electronic device. By constructing logistics behavior relationship data, after obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined according to the consignment relationship between the associated products associated with the product to be recommended in the logistics behavior relationship data and each candidate recommendation object, and the target recommendation object is obtained according to the correlation between each candidate recommendation object and the product to be recommended, and the product to be recommended is recommended to the target recommendation object. In this way, based on the consignment relationship between the associated product and each candidate recommendation object, the object for product recommendation can be comprehensively and accurately obtained, thereby improving the accuracy of product recommendation. At the same time, the embodiments of the present invention can obtain the real needs of customers through logistics behavior relationship data, so that product recommendations based on logistics behavior relationship data can improve the accuracy and rationality of product recommendations.

[0059] See also Figure 1 , Figure 1 A schematic diagram of the structure of an electronic device provided in this embodiment. Figure 1 As shown, the electronic device 100 includes a product recommendation device 10, a memory 20, a processor 30, and a communication unit 40. The memory 20, the processor 30, and the communication unit 40 are electrically connected to each other directly or indirectly to achieve signal transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0060] In this embodiment, the product recommendation device 10 includes at least one software function module that can be stored in the memory 20 in the form of software or firmware. The processor 30 is used to execute the software function module stored in the memory 20. When the electronic device 100 is running, the processor 30 communicates with the memory 20 through a bus, and the processor 30 executes the software function module to implement the product recommendation method described in this embodiment.

[0061] Among them, the memory 20 can be, but is not limited to, random access memory (Random Access Memory, RAM), read only memory (Read Only Memory, ROM), programmable read-only memory (Programmable Read-Only Memory, PROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable read-only memory (Electric Erasable Programmable Read-Only Memory, EEPROM), etc.

[0062] The processor 30 is used to perform one or more functions described in this embodiment. In some embodiments, the processor 30 may include one or more processing cores (e.g., a single-core processor (S) or a multi-core processor (S)). By way of example only, the processor 30 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), an application specific instruction set processor (ASIP), a graphics processing unit (GPU), a physical processing unit (PPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a controller, a microcontroller unit, a reduced instruction set computer (RISC) or a microprocessor, or any combination thereof.

[0063] For ease of description, only one processor is described in the electronic device 100. However, it should be noted that the electronic device 100 in this embodiment may also include multiple processors, so the steps performed by one processor described in this embodiment may also be performed jointly or individually by multiple processors. For example, if the processor of the electronic device performs step A and step B, it should be understood that step A and step B may also be performed jointly by two different processors or individually in one processor. For example, the processor performs step A and the second processor performs step B, or the processor and the second processor perform steps A and B together.

[0064] In this embodiment, the process definition method disclosed in any implementation manner may be applied to the processor 30 or implemented by the processor 30 .

[0065] The communication unit 40 is used to establish a communication connection between the electronic device 100 and other devices through a network, and to send and receive data through the network.

[0066] In some embodiments, the network may be any type of wired or wireless network, or a combination thereof. By way of example only, the network may include a wired network, a wireless network, a fiber optic network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a wide area network (WAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, or a near field communication (NFC) network, or any combination thereof.

[0067] In this embodiment, the electronic device may be a notebook, an ultra-mobile personal computer (UMPC), a personal digital assistant (PDA), a physical server, or other devices, and may also be a server cluster built based on multiple physical servers. This embodiment does not impose any restrictions on the specific type of electronic device.

[0068] Understandably, Figure 1 The structure shown is for illustration only. The electronic device may also have Figure 1 More or fewer components than shown, or with Figure 1 Different configurations are shown. Figure 1 The components shown may be implemented in hardware, software or a combination thereof.

[0069] based on Figure 1 This embodiment provides a product recommendation method. Figure 1 The steps of the product recommendation method provided in this embodiment are described in detail below. Figure 2 The product recommendation method provided in this embodiment includes steps S101 to S103.

[0070] Step S101: Acquire the product to be recommended and the logistics behavior relationship data of the product to be recommended.

[0071] Among them, the products to be recommended refer to the products that need to be recommended. The products to be recommended can be set according to business needs and are not specifically limited here.

[0072] In an optional implementation, the electronic device may provide an interactive interface to the business party, and the business party may input the product to be recommended through the interactive interface. The electronic device may obtain the product to be recommended by responding to the input operation of the business party.

[0073] In an optional implementation, the electronic device may also receive product information sent by other service terminals, and obtain the product to be recommended based on the product information. The product information may include product identification, product name, product introduction, and the like.

[0074] In this embodiment, the logistics behavior relationship data refers to the consignment relationship data between objects and products. Considering that in actual applications, when an object has a demand for a product, it will often purchase products associated with the product, thereby generating consignment behavior for associated products. Therefore, in this embodiment, when determining the logistics behavior relationship data of a product, the consignment relationship of the associated products associated with the product can also be added to the logistics behavior relationship data. Therefore, in this embodiment, the logistics behavior relationship data of each product also includes the consignment relationship between the associated products associated with the product and the object. In this way, when determining the target recommendation object based on the logistics behavior relationship data, objects with potential demand can be found, so that more comprehensive recommendation objects can be obtained, thereby improving the comprehensiveness of the recommendation.

[0075] The associated products associated with the product to be recommended may be products having the same functions as the product to be recommended, or may be products having the same usage scenarios as the product to be recommended, and are not limited here.

[0076] Among them, the candidate recommendation object represents the candidate recommendation user corresponding to the product to be recommended. In the logistics behavior relationship data, each object corresponds to a user. The object in the logistics behavior relationship data can be represented based on the user identification information of the corresponding user or the user identification information can be encoded. Optionally, the user identification information can be unique identification information such as the user's mobile phone number and the user's ID number. It can be understood that the products in the logistics behavior relationship data are also represented based on the product information of the product or the product information is encoded.

[0077] In this embodiment, the consignment relationship mainly refers to the recipient's behavior of receiving the product.

[0078] Step S102: Determine the relevance between each candidate recommendation object and the product to be recommended based on the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object.

[0079] In this embodiment, the consignment relationship between each object and a product can reflect the object's demand for the product. When an object has multiple consignment relationships with a product, and multiple consignment relationships are receiving relationships, it indicates that the object has demand for the product. Therefore, in this embodiment, target recommended objects with potential demand for the product can be obtained based on the product's logistics behavior relationship data.

[0080] In this embodiment, for the product to be recommended, when determining the target recommendation object that has a demand for the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined based on the consignment relationship between the associated products of the product to be recommended and each candidate recommendation object in the logistics behavior relationship data. Based on the correlation between each candidate recommendation object and the object to be recommended, the target recommendation object can be determined from each candidate recommendation object.

[0081] In this embodiment, the relevance is used to characterize the demand relevance between the product and the object. When the relevance between the candidate recommendation object and the associated product is higher, it means that the demand relevance between the candidate recommendation object and the product to be recommended is higher, and the authenticity of the candidate recommendation object as the target recommendation object is higher. Correspondingly, when the relevance between the candidate recommendation object and the associated product is lower, it means that the demand relevance between the candidate recommendation object and the product to be recommended is lower, and the authenticity of the candidate recommendation object as the target recommendation object is lower.

[0082] In an optional implementation, when determining the relevance between each candidate recommendation object and the product to be recommended, for each candidate recommendation object, the total number of all consignment behaviors (including consignment behaviors that represent collection of items and consignment behaviors that represent mailing) between the candidate recommendation object and the associated products associated with the product to be recommended, as well as the consignment quantity of the consignment relationship (consignment behaviors that represent collection of items) between the candidate recommendation object and the associated products associated with the product to be recommended, can be determined first, and then the ratio of the consignment quantity to the total quantity is determined, and the ratio is determined as the relevance between the candidate recommendation object and the product to be recommended. For example, for product A to be recommended, the associated associated products are product a, assuming that candidate recommendation object X1 has 10 consignment behaviors for product a, of which there are 2 consignment relationships that represent collection of items, then the relevance between candidate recommendation object X1 and product A to be recommended is 1 / 5.

[0083] It can be understood that if a candidate recommendation object has a consignment relationship with multiple related products, the proportion corresponding to a single related product can be determined first, and then the proportions corresponding to multiple related products can be accumulated to obtain the relevance between the candidate recommendation object and the product to be recommended. For example, for product A to be recommended, the related products are product a and product b. Assuming that candidate recommendation object X1 has 10 consignment behaviors for product a, of which 2 are consignment relationships for collection, and candidate recommendation object X1 has 5 consignment behaviors for product b, of which 1 is consignment relationship for collection, then the relevance between candidate recommendation object X1 and product A to be recommended is 2 / 10+1 / 5=2 / 5.

[0084] In an optional implementation, for each candidate recommendation object, the sum of the number of consignment relationships between the candidate recommendation object and the associated products associated with the product to be recommended can also be directly determined as the correlation between the candidate recommendation object and the product to be recommended. For example, for product A to be recommended, the associated products are product a and product b. Assuming that candidate recommendation object X1 has 10 consignment relationships with product a, of which 2 are consignment relationships that represent collection, and candidate recommendation object X1 has 5 consignment relationships with product b, of which 1 is consignment relationship that represents collection, then the correlation between candidate recommendation object X1 and product A to be recommended is 3.

[0085] Step S103: Obtain a target recommendation object based on the relevance between each candidate recommendation object and the product to be recommended, and recommend the product to be recommended to the target recommendation object.

[0086] In an optional implementation, when determining the target recommendation object based on the relevance between each candidate recommendation object and the product to be recommended, the relevance between each candidate recommendation object and the product to be recommended can be ranked, and the candidate recommendation objects ranked in the top N relevances are selected as the target recommendation objects. It can be understood that N is a positive integer and can be set according to business needs, and this embodiment does not impose any specific restrictions.

[0087] In an optional implementation, when determining the target recommendation object based on the relevance between each candidate recommendation object and the product to be recommended, the relevance between each candidate recommendation object and the product to be recommended can also be compared with a set relevance threshold, and the candidate recommendation object with a relevance greater than the set relevance threshold can be determined as the target recommendation object.

[0088] After the target recommendation object is determined, the product to be recommended can be recommended to the target recommendation object. It can be understood that recommending the product to be recommended to the target recommendation object refers to recommending the product to be recommended to the client corresponding to the target recommendation object.

[0089] In an optional implementation, the logistics behavior relationship data may also include the consignment relationship between the product to be recommended and the object. When recommending the product to be recommended to the target recommendation object, the direct object that has a direct consignment relationship with the product to be recommended can be obtained based on the logistics behavior relationship data, and then the product to be recommended can be recommended to the direct object and the target recommendation object.

[0090] This embodiment constructs logistics behavior relationship data. After obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined based on the consignment relationship between the associated products associated with the product to be recommended in the logistics behavior relationship data and each candidate recommendation object. According to the correlation between each candidate recommendation object and the product to be recommended, the target recommendation object is obtained, and the product to be recommended is recommended to the target recommendation object. In this way, based on the consignment relationship between the associated products and each candidate recommendation object, an object with potential demand for the product can be obtained, so that a more comprehensive recommendation object can be obtained, and the comprehensiveness of the recommendation can be improved. At the same time, this embodiment can obtain the real needs of customers through the logistics behavior relationship data, so that product recommendations based on the logistics behavior relationship data can improve the accuracy and rationality of product recommendations.

[0091] In this embodiment, the accuracy of the logistics behavior relationship data of each product affects the accuracy of product recommendation. Therefore, in this embodiment, the construction of the logistics behavior relationship data of each product is a relatively important key point. In order to improve the accuracy of the logistics behavior relationship data, in this embodiment, the logistics behavior relationship data can be constructed through the following process.

[0092] (1) Obtain logistics information data.

[0093] The logistics information data refers to the data that generates logistics behavior. In this embodiment, the logistics information data includes the products entrusted by each object and product attribute information.

[0094] Among them, the product attribute information includes the number of times the product is consigned, the consignment weight of each consignment, the consignment time, and the consignment amount, etc.

[0095] In an optional implementation, the logistics information data may be obtained according to a set time period. For example, if the set time period is 1 year, the logistics information data within 1 year may be obtained according to the consignment time.

[0096] In an optional implementation, after obtaining the logistics information data, the logistics information data may be divided according to time periods to obtain logistics information data of different time periods, and logistics behavior relationship data of different time periods may be constructed based on the logistics information data of different time periods.

[0097] (2) Determine the object logistics behavior relationship data based on the products consigned by each object and the number of consignments of the consigned products.

[0098] Among them, after obtaining the logistics information data, the object logistics behavior relationship can be constructed according to the products consigned by each object in the logistics information data and the number of consignments of the consigned products.

[0099] In this embodiment, for each product, if there is any object that has a logistics behavior of receiving the product, a consignment relationship is established between the product and the object. For example, if there is a logistics behavior of receiving the product between object A and product a, a consignment relationship of object A→(receiving) product a can be established.

[0100] In this embodiment, for each product, when constructing the consignment relationship between the product and the object, the product attribute information of the product is added to the consignment relationship, so that the consignment relationship includes the object and the number of consignments of the product.

[0101] In this embodiment, the object logistics behavior relationship data can also be embodied in the form of an object logistics behavior relationship network, such as Figure 3 As shown, the nodes in the object logistics behavior relationship network are used to represent products and objects, and the edges in the object logistics behavior relationship network are used to represent the consignment relationship between products and objects.

[0102] Optionally, the edge may be associated with the product attribute information of the product, and the product attribute information of the product may be obtained through the edge. The product attribute information may also be associated with a product node or an object node, and the product attribute information may be obtained through the node.

[0103] In an optional implementation, when constructing the object logistics behavior relationship network, the object node can also be used to store the image information of the object, such as gender, age, etc. For the product node, the basic information of the product, such as the type and weight of the product, etc. can also be stored.

[0104] (3) For each product, determine the associated products associated with the product, and filter the object logistics behavior relationship data based on the product and the associated products associated with the product to obtain the logistics behavior relationship data of the product.

[0105] In this embodiment, the object logistics behavior relationship data includes the consignment relationship between each product and each object. For a single product, if you want to obtain the logistics behavior relationship data of the product, you need to filter the object logistics behavior relationship data to obtain the logistics behavior relationship data of the single product.

[0106] Optionally, in this embodiment, for each product, the product type of the product may be determined first, and then, based on the product type of the product, products belonging to the same product type may be used to determine associated products associated with the product.

[0107] Among them, product types can be divided according to usage scenarios and product uses, without specific restrictions.

[0108] For example, taking the usage scenario as an example, assuming that the product is infant formula, its usage scenario can be divided into the mother-infant scenario, so the associated products associated with this product can be diapers, children's clothing, baby strollers, etc. in the same usage scenario. For example, taking the product function as an example, assuming that the product is milk, its function is to provide calcium intake, so the associated products associated with this product can be calcium tablets with the same function or other products with calcium supplement function.

[0109] In an optional implementation, a mapping relationship table between product types and product information may be established in advance. Thus, for each product, when determining the product type of the product, the product information of the product may be matched with the product information in the mapping relationship table to obtain target product information that matches the product information of the product, and the product type corresponding to the target product information may be used as the product type of the product. Then, products corresponding to all product information associated with the product type may be determined as associated products.

[0110] In an optional implementation, a product type classifier may be pre-trained. Thus, when determining the product type of the product, the product information of the product may be input into the product type classifier, and the product information may be classified by the product type classifier to obtain the product type of the product. Then, according to the product type of the product, the products belonging to the same product type may be used to determine the associated products associated with the product. The product type classifier may be trained by machine learning or deep learning.

[0111] In an optional implementation, for each product, the associated products of the product may also be determined based on expert experience.

[0112] To improve accuracy, in this embodiment, for each product, a target user group that has a real demand for the product may be determined first, and then related products associated with the product may be obtained based on the products entrusted by the target user group.

[0113] Based on this, in this embodiment, for each product, the step of determining the associated products associated with the product may also include:

[0114] Step a1: Obtain product transaction data.

[0115] Step a2: For each product, determine the target transaction object that has demand for the product based on the product transaction data.

[0116] Step a3: Determine the target product consigned by the target transaction object based on the logistics information data.

[0117] Step a4: Determine the associated products associated with the target product consigned by the target transaction object.

[0118] Considering that in actual applications, when a user has a demand for a product, he or she will often purchase the product or a product related to the product. Therefore, this embodiment can determine the target user group that has a real demand for the product, that is, the target transaction object, based on the product transaction data.

[0119] The product transaction data can be obtained from the product transaction platform. In this embodiment, the product transaction data includes the products traded by each object and the transaction time.

[0120] Optionally, in this embodiment, for each product, when determining the target transaction object that has a real demand for the product based on the product transaction data, the product can be matched with each product in the product transaction data to obtain the transaction object that trades the product, and the transaction object that trades the product can be determined as the target transaction object that has a real demand for the product.

[0121] Optionally, in this embodiment, for each product, when determining the target transaction object that has a real demand for the product based on the product transaction data, the target of the product can be determined, and based on the target of the product, the target transaction object that has a real demand for the product can be obtained from the product transaction data.

[0122] When determining the target of a product, it can be determined based on the product's usage scenario and function. For example, for infant formula products, its usage scenario is the mother-infant scenario, and its function is to provide nutrition for infants. After analyzing the product's usage scenario and function introduction, it can be concluded that the target of the product is infants.

[0123] It is understandable that the analysis and processing of product usage scenarios and function introductions can be processed through network models, such as the language representation model (Bidirectional Encoder Representation from Transformers, BERT).

[0124] In this embodiment, when the target transaction object with a real demand for the product is obtained from the product transaction data based on the target of the product, the target of the product traded by each object in the product transaction data can be first determined based on the same method, and then the target of the product with the target of the product traded by each object in the product transaction data can be matched to obtain matching products whose target in the product transaction data is the same as the target of the product, and then the object of the transaction matching product is determined as the target transaction object with a real demand for the product.

[0125] For example, assuming that the product is infant formula, and its target audience is infants, it is assumed that there are products a, b, and c in the product transaction data, where product a is an infant insurance product, and its target audience is infants; product b is a fitness equipment product, and its target audience is adults; and product c is a nutritional product for the elderly, and its target audience is the elderly. By matching according to the target audience, product a can be obtained as a matching product that matches the target audience of the product. In this way, all objects that trade product a in the product transaction data can be determined as target transaction objects that have a real demand for the product.

[0126] After obtaining the target transaction object, the target product entrusted by the target transaction object can be determined based on the logistics information data.

[0127] Optionally, when determining the target product entrusted by the target transaction object based on the logistics information data, the target transaction object can be matched with each object in the logistics information data to obtain the target object in the logistics information data that is the same as the target transaction object, and then all products entrusted by the target object can be determined as the target products entrusted by the target transaction object.

[0128] After obtaining the target product entrusted by the target transaction object, the associated products associated with the product can be determined according to the target product entrusted by the target transaction object.

[0129] Optionally, when determining the associated products associated with the target product entrusted by the target transaction object, the correlation between the target product and the product can be determined, and the target products ranked in the top N in correlation can be determined as the associated products associated with the product. Wherein, N is a positive integer. When determining the correlation between the target product and the product, it can be obtained by Pearson correlation coefficient method, Euclidean distance, etc.

[0130] In this embodiment, for each product, after determining the associated products associated with the product, the object logistics behavior relationship data can be filtered according to the product and the associated products associated with the product to obtain the logistics behavior relationship data of the product.

[0131] Among them, when filtering the object logistics behavior relationship data according to the product and the associated products associated with the product, the consignment relationship including the product and the consignment relationship including the associated products can be retained, and the consignment relationship not including the product and the consignment relationship not including the associated products can be filtered, so that the logistics behavior relationship data of the product can be obtained. It can be understood that the logistics behavior relationship data of each product obtained includes the consignment relationship between the product and the object and the consignment relationship between the associated product and the object.

[0132] Considering that in actual application, there may be interference items in the consignment relationship of each product, in order to improve the accuracy of the logistics behavior data of each product, in this embodiment, when constructing the logistics behavior data of each product, the interference items in the logistics behavior data of each product can be filtered. Based on this, according to the product and the associated products associated with the product, the object logistics behavior relationship data is filtered, and the process of obtaining the logistics behavior data of the product can also include:

[0133] According to the product and the products associated with the product, the object logistics behavior relationship data is filtered to obtain the initial logistics behavior relationship data of the product.

[0134] The initial logistics behavior relationship data is filtered according to the preset filtering rules to obtain the logistics behavior relationship data of the product.

[0135] Among them, for each product, according to the product and the associated products associated with the product, the object logistics behavior relationship data is filtered, that is, the consignment relationship including the product and the consignment relationship including the associated products are retained, and the consignment relationship not including the product and the consignment relationship not including the associated products are filtered, so that the initial logistics behavior relationship data of the product can be obtained. It can be understood that the initial logistics behavior relationship data of each product obtained includes the consignment relationship between the object consigning the product and the product, and the consignment relationship between the object consigning the associated products associated with the product and the associated products associated with the product.

[0136] For each product, after obtaining the initial logistics behavior relationship data of the product, the initial logistics behavior relationship data of the product can be filtered according to the filtering rules.

[0137] Among them, the filtering rules refer to the rules set when filtering the logistics behavior relationship data, which can be set according to actual needs without specific restrictions.

[0138] In this embodiment, each consignment relationship in the logistics behavior relationship data includes the number of times the object corresponding to the consignment relationship consigns the product corresponding to the consignment relationship and the consignment weight of the product corresponding to the consignment relationship. Therefore, this embodiment can formulate filtering rules based on the number of consignments and the consignment weight.

[0139] For example, the filtering rules include whether the number of consignments within a set time period meets a set number threshold, and whether the consignment weight of each consignment meets a set weight threshold.

[0140] When filtering, for each associated product in the initial logistics behavior relationship data, first, according to the number of consignments and consignment time included in the consignment relationship between each object and the associated product, the number of consignments of each object to the associated product within the set time period is obtained, and according to the consignment weight included in the consignment relationship between each object and the associated product, the consignment weight of each object consigning the associated product each time is obtained, and then the number of consignments of each object to the associated product within the set time period is compared with the set number threshold, and the consignment weight of each object consigning the associated product each time is compared with the set weight threshold. For each object, if the number of consignments of the object to the associated product within the set time period is less than the set number threshold, and the consignment weight of the object consigning the associated product each time is less than the set weight threshold, then the consignment relationship between the object and the associated product is determined to be interference data, indicating that the object is unlikely to have a demand for the associated product, and the demand for the recommended product corresponding to the associated product is small, and then the consignment relationship between the object and the product can be filtered. If the number of times the object has consigned the associated product within the set time period is not less than the set number threshold, or the object has consigned the associated product once with a weight that is not less than the set weight threshold, then it is determined that the consignment relationship between the object and the associated product is not interference data, indicating that the object is very likely to have a demand for the product, and thus has a greater demand for the recommended product corresponding to the associated product, and there is no need to filter the consignment relationship between the object and the associated product.

[0141] In order to improve the accuracy of the data, when formulating filtering rules based on the number of consignments and consignment weight, for each associated product, the number of consignments corresponding to the associated product can be determined based on the number of consignments of the associated product by the object consigning the associated product, and the consignment weight index value of each consignment of the associated product can be determined based on the consignment weight of each consignment of the associated product by the object consigning the associated product. Then, the filtering rules for the product are determined based on the number of index values ​​and the consignment weight index values.

[0142] Optionally, the number of consignments corresponding to the associated product is determined based on the number of consignments of the associated product by the object consigning the associated product. This may be to determine the average number of consignments of the associated product by the object consigning the associated product, and determine the average number as the number of consignments. This may also be to determine the median number of consignments of the associated product by the object consigning the associated product, and determine the median number as the number of consignments. This may also be to determine the mode value of the number of consignments of the associated product by the object consigning the associated product, and determine the mode value as the number of consignments.

[0143] According to the consignment weight of the associated product consigned each time by the object consigning the associated product, the consignment weight index value of each consignment of the associated product is determined. It can be that according to the consignment weight of the associated product consigned each time by the object consigning the associated product, the average weight of the associated product consigned each time by the object consigning the associated product is determined, and the average weight is determined as the consignment weight index value. It can also be that according to the weight of the associated product consigned each time by the object consigning the associated product, the median weight of the associated product consigned each time by the object consigning the associated product is determined, and the median weight is determined as the consignment weight index value. It can also be that according to the weight of the associated product consigned each time by the object consigning the associated product, the mode value of the weight of the associated product consigned each time by the object consigning the associated product is determined, and the mode value of the weight is determined as the consignment weight index value.

[0144] When determining the filtering rule of the product based on the number of times index value and the consignment weight index value, the filtering rule may be determined by setting the number of times threshold value of each object consigning the associated product according to the number of times index value, and the weight threshold value of each object consigning the associated product each time according to the consignment weight index value. The filtering rule is determined based on the number of times threshold value of each object consigning the associated product and the weight threshold value of each consignment of the associated product. The number of times threshold value of each object consigning the associated product may be the same as the number of times index value, and the weight threshold value of each object consigning the associated output may be the same as the consignment weight index value.

[0145] For example, for a certain associated product of a product, the threshold number of times each object can consign the associated product is set to 5 times, and the weight threshold of each consignment of the associated product is set to 1kg. Then, the filtering rule can be set that the number of times each object can consign the associated product cannot be less than 5 times, and the weight of each consignment of the associated product cannot be less than 1kg.

[0146] In an optional implementation, in this embodiment, the filtering rules can also be determined according to the number index value and the consignment weight index value within the set time period. In detail, the number threshold of consignment of the associated product for each object within the set time period and the weight threshold of each consignment of the associated product can be set according to the number index value and the consignment weight index value within the set time period. Then, the filtering rules are determined according to the number threshold of consignment of the associated product for each object within the set time period and the weight threshold of each consignment of the associated product.

[0147] For example, assuming that the set time period is January, for a certain associated product of the product, the number of times each object consigns the associated product in each month is set to 3 times, and the weight threshold of each consignment of the associated product is set to 0.5kg. Then, the filtering rules can be set such that the number of times each object consigns the associated product in each month cannot be less than 3 times, and the weight of each consignment of the associated product cannot be less than 0.5kg, or the weight threshold of each object consigning the associated product in each month cannot be less than 1.5kg.

[0148] Considering that in actual applications, in different scenarios, for each product, the object will have a demand for the product and related products at the same time, so there will be a consignment relationship between the product and the related products at the same time. In order to improve accuracy, in this embodiment, for each product, the related products associated with the product can be divided into scenarios, and filtering rules are determined based on the scenarios for filtering. Based on this,

[0149] Based on the product and the associated products associated with the product, the step of filtering the object logistics behavior relationship data to obtain the logistics behavior data of the product includes:

[0150] Step b1: For each product, obtain the inferred product scenario of the product and the filtering rules corresponding to the inferred product scenario.

[0151] Step b2: Filter the object logistics behavior relationship data according to the product and the products associated with the product to obtain the initial logistics behavior relationship data of the product.

[0152] Step b3: Filter the initial logistics behavior relationship data according to the filtering rules corresponding to each inferred product scenario to obtain the logistics behavior relationship data of the product.

[0153] Among them, for each product, after obtaining the associated products of the product, the inferred product scenario of the product and the associated products under different inferred product scenarios can be determined according to the expert experience method. In this embodiment, the inferred product scenario refers to the possible usage scenario of the product. For example, for the product of baby milk powder, its inferred product scenario can be scenarios such as having a baby at home, having a baby yourself, and having a baby with a friend. For the inferred product scenarios of having a baby at home and having a baby with a friend, the associated products can be diapers, children's clothing and other products. For the inferred product scenario of having a baby with a friend, its associated products can be baby toys, baby jewelry and other products.

[0154] In this embodiment, different inferred product scenarios may include different associated products, or some or all of the associated products may be the same. For example, for product A, its corresponding inferred product scenarios include scenario A, scenario B, and scenario C, where the associated products under scenario A are product a, product b, and product c, the associated products under scenario B are product a, product b, and product d, and the associated products under scenario C are product e and product f.

[0155] In this embodiment, each inferred product scenario has a corresponding filtering rule, and the filtering rule corresponding to each inferred product scenario can be set according to expert experience.

[0156] To improve accuracy, the filtering rules corresponding to each inferred product scenario can be obtained through the following steps:

[0157] Step c1: for each inferred product scenario of the product, determine the consignment object that has a consignment relationship with the product and the related products under the inferred product scenario according to the logistics information data.

[0158] Step c2: Determine the number of times each consignment object consigns the product and the associated products under the inferred product scenario within a set time period.

[0159] Step c3: according to the number of times each consignment object consigns the product within a set time period and the number of times the associated products under the inferred product scenario are consigned, a filtering rule for the product under the inferred product scenario is obtained.

[0160] Among them, after obtaining the inferred product scenario of each product, for each inferred product scenario of each product, the consignment object that has a consignment relationship with the product and the related products under the inferred product scenario can be determined based on the logistics information data.

[0161] For example, for the inferred product scenario A of product A, assuming that the related products under the inferred product scenario A are product a and product b, the consignment objects for which consignment behavior exists for product A and both product a and product b can be determined first.

[0162] After obtaining the consignment object, the number of consignments of the product and the associated products under the inferred product scenario by each consignment object within the set time period can be determined based on the logistics information data. The set time period can be set according to actual needs, for example, the set time period can be set by month or by year.

[0163] Optionally, when determining the filtering rules for the product under the inferred product scenario based on the number of times each consignment object consigns the product within a set time period and the number of times related products under the inferred product scenario have been consigned, the first number index value of the product within the set time period and the second number index value of the related products under the inferred product scenario can be determined based on the number of times each consignment object consigns the product within the set time period and the number of times related products under the inferred product scenario have been consigned. Based on the first number index value of the product within the set time period and the second number index value of the related products, the filtering rules for the product under the inferred product scenario can be obtained.

[0164] Optionally, the average number of times each consignment object consigns the product within a set time period may be determined as the first number index value, and the average number of times each consignment object consigns the associated product under the inferred product scenario within the set time period may be determined as the second number index value. Alternatively, the median number of times each consignment object consigns the product within a set time period may be determined as the first number index value, and the median number of times each consignment object consigns the associated product under the inferred product scenario within the set time period may be determined as the second number index value. Alternatively, the mode value of the number of times each consignment object consigns the product within a set time period may be determined as the first number index value, and the mode value of the number of times each consignment object consigns the associated product under the inferred product scenario within the set time period may be determined as the second number index value.

[0165] Among them, when obtaining the filtering rule of the product in the inferred product scenario based on the first number index value of the product within the set time period and the second number index value of the associated product, the first number index value of the product within the set time period and the second number index value of the associated product can be used to set the number threshold of the number of times each object entrusts the product within the set time period and the number threshold of the associated product, and then determine the filtering rule based on the number threshold of the number of times each object entrusts the product within the set time period and the number threshold of the associated product.

[0166] For example, assuming that the set time period is January, for product A, the associated products are product a and product b, and the number threshold for each object to consign product A in each month is set to 5 times, the number threshold for product a is set to 3 times, and the number threshold for product b is set to 2 times. Then the filtering rules can be set that the number of times each object consigns product A in each month is not less than 5 times, the number of times product a is not less than 3 times, and the number of times product b is not less than 2 times.

[0167] For each product, after obtaining the filtering rules for the product in different inferred product scenarios, the object logistics behavior relationship data can be filtered according to the product and its associated products, that is, the consignment relationship including the product and the consignment relationship including the associated products are retained, and the consignment relationship not including the product and the consignment relationship not including the associated products are filtered to obtain the initial logistics behavior relationship data of the product. Figure 4 As shown, Figure 4 is the initial logistics behavior relationship data of product P.

[0168] After obtaining the initial logistics relationship data of the product, the initial logistics behavior relationship data can be filtered according to the filtering rules corresponding to each inferred product scenario to obtain the logistics behavior relationship data of the product.

[0169] Among them, when filtering the initial logistics behavior relationship data according to the filtering rules corresponding to each inferred product scenario to obtain the logistics behavior relationship data of the product, for each associated product in the initial logistics behavior relationship data, it can be determined whether to filter the consignment relationship between the associated product and the object based on the inferred product scenario corresponding to the associated product and then the filtering rules corresponding to the inferred product scenario. Figure 5 As shown, for the associated product b, the corresponding inferred product scenario is L, then the consignment relationship between the associated product b and the object B can be detected, and whether the consignment relationship between the associated product b and the object D satisfies the filtering rule R1 corresponding to the inferred product scenario L. Assuming that the consignment relationship between the associated product b and the object B does not satisfy the filtering rule R1, the consignment relationship between the associated product b and the object B is filtered.

[0170] After obtaining the logistics behavior relationship data of each product, the logistics behavior relationship data of each product and the product can be stored and updated according to the set cycle. When product recommendation is required, the product information of the product to be recommended can be determined first, and then the logistics behavior relationship data of the product to be recommended can be obtained based on the product information of the product to be recommended. According to the logistics behavior relationship data of the product to be recommended, the relevance between each candidate recommendation object and the product to be recommended is determined, and the target recommendation object is obtained based on the relevance between each candidate recommendation object and the product to be recommended, and the product to be recommended is recommended to the target recommendation object.

[0171] In order to improve the accuracy of product recommendation, in this embodiment, the step of determining the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object includes:

[0172] For each candidate recommendation object, a weight coefficient of a target-related product corresponding to the consignment relationship of the candidate recommendation object and a consignment frequency index value of the target-related product are determined.

[0173] According to the number of times the candidate recommendation object has consigned the target associated product and the consignment number index value of the target associated product, a consignment number coefficient of the target associated product is obtained.

[0174] The relevance between the candidate recommendation object and the product to be recommended is determined based on the weight coefficient and the consignment frequency coefficient of the target associated product.

[0175] In an optional implementation, the weight coefficient corresponding to each associated product can be determined based on the correlation between the associated product and the product to be recommended. The higher the correlation, the greater the weight coefficient. Conversely, the lower the correlation, the smaller the weight coefficient. The correlation can be obtained by Pearson correlation coefficient method, Euclidean distance, etc.

[0176] In this embodiment, the weight coefficient of each associated product can be obtained according to the weight coefficient corresponding to the inferred product scenario corresponding to the associated product, wherein the weight coefficient corresponding to each inferred product scenario can be set according to expert experience and business needs. It can be understood that the sum of the weights of the inferred product scenarios is 1.

[0177] In this embodiment, the consignment times index value of each associated product refers to a reference value of the consignment times of each object in the logistics behavior relationship data for the associated product, which can be calculated based on the consignment times of each object in the logistics behavior relationship data for the associated product.

[0178] Therefore, in this embodiment, for each candidate recommendation object, the step of determining the weight coefficient of the target associated product corresponding to the consignment relationship of the candidate recommendation object and the consignment frequency index value of the target associated product may include:

[0179] (1) Determine the inferred product scenario corresponding to the target-related product, and obtain the weight coefficient of the target-related product based on the inferred product scenario.

[0180] (2) According to the consignment relationship between the target-related product and each candidate recommendation object, the number of consignments of each candidate recommendation object to the target-related product is determined.

[0181] (3) According to the number of times each candidate recommendation object has consigned the target-related product, an index value of the number of times the target-related product has been consigned is obtained.

[0182] In this embodiment, different inferred product scenarios have different weight coefficients. For each associated product, if the associated product corresponds to only one inferred product scenario, the weight coefficient of the associated product is the same as the weight of the corresponding inferred product scenario. If the associated product corresponds to multiple inferred product scenarios, that is, the associated product exists in multiple inferred product scenarios, then the weight coefficient of the associated product is the sum of the weights of the corresponding launched product scenarios.

[0183] In this embodiment, since the consignment relationship between each candidate recommended object and the associated product in the logistics behavior relationship data includes the number of consignments of the candidate recommended object to the associated product, for each target associated product corresponding to the candidate recommended object, the consignment relationship between the target associated product and each candidate recommended object is analyzed to obtain the number of consignments of each candidate recommended object to the target associated product. After obtaining the number of consignments of each candidate recommended object to the target associated product, the index value of the number of consignments of each candidate recommended object to the target associated product is calculated to obtain the index value of the number of consignments of the target associated product.

[0184] The index value of the number of consignments of each candidate recommendation object to the target-related product can be calculated by averaging the number of consignments of each candidate recommendation object to the target-related product, and using the average number of consignments as the index value of the number of consignments of the target-related product. Alternatively, the median number of consignments of each candidate recommendation object to the target-related product can be calculated, and the median number of consignments can be used as the index value of the number of consignments of the target-related product. Alternatively, the mode number of consignments of each candidate recommendation object to the target-related product can be calculated, and the mode number of consignments can be used as the index value of the number of consignments of the target-related product.

[0185] For each candidate recommendation object, after determining the weight coefficient of the target-related product corresponding to the consignment relationship of the candidate recommendation object and the consignment frequency index value of the target-related product, the consignment frequency coefficient of the target-related product can be obtained based on the number of consignments of the target-related product by the candidate recommendation object and the consignment frequency index value of the target-related product.

[0186] Optionally, in this embodiment, the consignment frequency coefficient of the target-related product is obtained according to the number of consignments of the candidate recommended object to the target-related product and the consignment frequency index value of the target-related product. The consignment frequency coefficient of the target-related product can be determined as the ratio of the number of consignments of the candidate recommended object to the target-related product and the consignment frequency index value of the target-related product. Alternatively, the consignment frequency coefficient of the target-related product can be determined as the difference between the number of consignments of the candidate recommended object to the target-related product and the consignment frequency index value of the target-related product.

[0187] Optionally, in this embodiment, the correlation between the candidate recommendation object and the product to be recommended is determined based on the weight coefficient and the consignment frequency coefficient of the target associated product. The correlation between the candidate recommendation object and the product to be recommended can be determined as the product of the weight coefficient and the consignment frequency coefficient of the target associated product.

[0188] It should be noted that, for each candidate recommendation object, if the candidate recommendation object corresponds to multiple target-related products, the sub-correlation corresponding to each target-related product can be obtained based on the weight coefficient of each target-related product and the number of consignments of the target-related product, and then the sub-correlation corresponding to each target-related product can be summed to obtain the correlation between the candidate recommendation object and the product to be recommended.

[0189] For example, Figure 6 As shown, Figure 6 Nodes A, B, and C are candidate recommendation nodes, nodes a, b, and c are related products, and node P is a product to be recommended. Node A has a consignment relationship with nodes b and c, node B has a consignment relationship with node b, and node C has a consignment relationship with node a. The inferred product scenario corresponding to node b is scenario L, the inferred product scenario corresponding to node a is scenario S, and the inferred product scenario corresponding to node c is scenario Y. Assuming that the weight coefficients of scenario L, scenario S, and scenario Y are 0.2, 0.3, and 0.5 respectively, the consignment index value of node b is 3 times, the consignment index value of node c is 4 times, and the consignment index value of node a is 2 times.

[0190] When determining the correlation between node A and node P, assuming that node A and node b have 4 consignments, and node A and node c have 5 consignments, then the consignment coefficient between node A and node b is 4 / 3, and the consignment coefficient between node A and node c is 5 / 4. The correlation between node A and node P is calculated to be: 4 / 3*0.2+5 / 4*0.5=0.8912.

[0191] When determining the correlation between node B and node P, assuming that there are 6 consignments between node B and node P, the consignment coefficient between node B and node P is 6 / 3, and the correlation between node B and node P is calculated to be: 6 / 3*0.2=0.4.

[0192] When determining the correlation between node C and node P, assuming that node C and node a have 4 consignments, the consignment coefficient between node C and node a is 4 / 2, and the correlation between node C and node P is calculated to be: 4 / 2*0.3=0.6.

[0193] After obtaining the relevance between the nodes of each candidate recommendation object and the node of the product to be recommended, the relevance can be sorted, so as to select the target recommendation objects with the top N relevance rankings from each candidate recommendation object, such as Figure 6 In the example, the correlation between node A and node P is greater than the correlation between node C and node P and greater than the correlation between node B and node P. If the top two candidate recommendation objects are selected as the target recommendation objects, then the candidate recommendation objects corresponding to node A and node C can be selected as the target recommendation objects.

[0194] It should be noted that Figure 6 Node D in is an object node that has a direct entrustment relationship with node P. Therefore, the object corresponding to node D can be used as a direct recommendation object. When making product recommendations, recommendations can be made directly to node D.

[0195] The product recommendation method provided in this embodiment constructs logistics behavior relationship data. After obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined based on the consignment relationship between the associated products associated with the product to be recommended in the logistics behavior relationship data and each candidate recommendation object, and the target recommendation object is obtained based on the correlation between each candidate recommendation object and the product to be recommended, and the product to be recommended is recommended to the target recommendation object. In this way, based on the consignment relationship between the associated products and each candidate recommendation object, an object with potential demand for the product can be obtained, so that a more comprehensive recommendation object can be obtained, and the comprehensiveness of the recommendation can be improved. At the same time, this embodiment can obtain the real needs of customers through the logistics behavior relationship data, so that product recommendations based on the logistics behavior relationship data can improve the accuracy and rationality of product recommendations.

[0196] It should be noted that in the specific implementation of the present invention, the logistics information data and product transaction information related to user information, when the above embodiments of the present application are applied to specific products or technologies, the user's permission or consent is required, and the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions. For example, during the development process, all users' sensitive information will be desensitized, and a series of data mining and inference will be carried out in the desensitized state. The final result is limited to internal use to ensure the confidentiality and security of the data.

[0197] Based on the same inventive concept, an embodiment of the present invention also provides a product recommendation method, which determines the target recommended product to be recommended to the recommended object by using logistics behavior relationship data when the recommended object is known. In detail, it can be implemented by the following steps:

[0198] (A) Obtain the target recommendation object for which product recommendation is required and the logistics behavior relationship data of the target recommendation object.

[0199] (B) Determine the relevance between the target recommendation object and each candidate recommendation product based on the consignment relationship between the target recommendation object and each candidate recommendation product.

[0200] (C) Obtaining a target recommended product based on the relevance between the target recommended object and each candidate recommended product, and recommending the target recommended product to the target recommended object.

[0201] Among them, the target recommendation object represents the target recommendation user who needs product recommendation. The electronic device can provide an interactive interface to the business party. The business party can input the object for which product recommendation is required through the interactive interface. The electronic device can obtain the target recommendation object by responding to the input operation of the business party.

[0202] In an optional implementation manner, the electronic device may also receive object information sent by other service terminals, and obtain a target recommended object based on the object information.

[0203] After obtaining the target recommended object, the object logistics behavior relationship data can be filtered according to the user identification information corresponding to the target recommended object to obtain the logistics behavior relationship data of the target recommended object. In this embodiment, the logistics behavior relationship data of the target recommended object includes the consignment relationship between the target recommended object and each candidate recommended product.

[0204] After obtaining the logistics behavior relationship data of the target recommendation object, the relevance between the target recommendation object and each candidate recommendation product can be determined based on the consignment relationship between the target recommendation object and each candidate recommendation product in the logistics behavior relationship data of the target recommendation object.

[0205] Among them, when determining the correlation between the target recommended object and each candidate recommended product based on the consignment relationship between the target recommended object and each candidate recommended product, for each candidate recommended product, the number of times the target recommended object has consigned the candidate recommended product can be obtained based on the consignment relationship between the target recommended object and the candidate recommended product, and the number of consignments coefficient of the candidate recommended product can be obtained based on the number of consignments of the target recommended object to the candidate recommended product and the number of consignments index value of the candidate recommended product, and the weight coefficient of the candidate recommended product can be obtained based on the inferred product scenario corresponding to the candidate recommended product, and then the correlation between the target recommended object and the candidate recommended product can be determined based on the weight coefficient and the number of consignments coefficient of the candidate recommended product.

[0206] Among them, according to the number of times the target recommended object has entrusting the candidate recommended product and the entrustment number index value of the candidate recommended product, the entrustment number coefficient of the candidate recommended product is obtained, and according to the inferred product scenario corresponding to the candidate recommended product, the weight coefficient of the candidate recommended product is obtained. According to the weight coefficient and the entrustment number coefficient of the candidate recommended product, the relevance between the target recommended object and the candidate recommended product is determined. The process of determining the relevance between each candidate recommended object and the product to be recommended based on the entrustment relationship between the associated products associated with each product to be recommended and each candidate recommended object can be referred to the above-mentioned process, which will not be repeated here.

[0207] After obtaining the relevance between the target recommendation object and each candidate recommendation product, the relevance between the target recommendation object and each candidate recommendation product can be ranked, and the candidate recommendation products ranked in the top N relevance rankings are selected as the target recommendation products. It can be understood that N is a positive integer and can be set according to business needs, and this embodiment does not make specific restrictions.

[0208] In an optional implementation, the relevance between the target recommendation object and each candidate recommendation product may be compared with a set relevance threshold, and the candidate recommendation product with a relevance greater than the set relevance threshold may be determined as the target recommendation product.

[0209] After obtaining the target recommended product, the target recommended product can be recommended to the target recommendation object.

[0210] The product recommendation method provided by the embodiment of the present invention constructs logistics behavior relationship data. After obtaining the logistics behavior relationship data of the target recommendation object, the relevance between the target recommendation object and each candidate recommendation product can be determined according to the consignment relationship between the target recommendation object and each candidate recommendation product. According to the relevance between the target recommendation object and each candidate recommendation product, the target recommendation product is obtained, and the target recommendation product is recommended to the target recommendation object. In this way, based on the consignment relationship between the candidate recommendation product and the target recommendation object, the product that the target object really needs can be obtained, so that the target recommendation product can be recommended to the target recommendation object, which can improve the accuracy and rationality of the product recommendation.

[0211] Based on the same inventive concept, the embodiment of the present invention also provides a product recommendation device, which is applied to Figure 1 The electronic device shown is executed by the electronic device, such as Figure 7 As shown, the product recommendation device 10 provided in this embodiment includes a data acquisition module 11 , a data processing module 12 and a product recommendation module 13 .

[0212] The data acquisition module 11 is used to acquire the product to be recommended and the logistics behavior relationship data of the product to be recommended; the logistics behavior relationship data includes the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object.

[0213] The data processing module 12 is used to determine the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object.

[0214] The product recommendation module 13 is used to obtain a target recommendation object according to the relevance between each candidate recommendation object and the product to be recommended, and recommend the product to be recommended to the target recommendation object.

[0215] In an optional implementation manner, the consignment relationship between the associated product associated with the to-be-recommended product and each candidate recommendation object includes the number of consignments; the data processing module 12 is used to:

[0216] For each candidate recommendation object, a weight coefficient of a target-related product corresponding to the consignment relationship of the candidate recommendation object and a consignment frequency index value of the target-related product are determined.

[0217] Obtaining a consignment frequency coefficient of the target associated product according to the consignment frequency of the candidate recommendation object and the consignment frequency index value of the target associated product;

[0218] The relevance between the candidate recommendation object and the product to be recommended is determined based on the weight coefficient and the consignment frequency coefficient of the target associated product.

[0219] In an optional implementation, the data processing module 12 is used to:

[0220] Determine the inferred product scenario corresponding to the target associated product, and obtain the weight coefficient of the target associated product according to the inferred product scenario; different inferred product scenarios have different weight coefficients;

[0221] According to the consignment relationship between the target related product and each candidate recommended object, the number of consignments of each candidate recommended object to the target related product is determined;

[0222] According to the number of times each candidate recommendation object has consigned the target-related product, a consignment number index value of the target-related product is obtained.

[0223] In an optional implementation, before obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the data acquisition module 11 is used to:

[0224] Obtaining logistics information data; the logistics information data includes the products consigned by each object and the number of consignments of the consigned products;

[0225] According to the products entrusted by each object and the number of times the entrusted products are entrusted, the object logistics behavior relationship data is determined; the object logistics behavior relationship data includes the consignment relationship between each product and each object; each consignment relationship includes the number of times the object corresponding to the consignment relationship consigns the product corresponding to the consignment relationship;

[0226] For each product, the associated products associated with the product are determined, and according to the product and the associated products associated with the product, the object logistics behavior relationship data is filtered to obtain the logistics behavior relationship data of the product.

[0227] In an optional implementation manner, the data acquisition module 11 is used to:

[0228] Obtain product transaction data, which includes the products traded by each object;

[0229] For each product, determine the target transaction object that has demand for the product based on the product transaction data;

[0230] Determine the target products consigned by the target transaction object based on the logistics information data;

[0231] According to the target product entrusted by the target transaction object, the associated products associated with the product are obtained.

[0232] In an optional embodiment, the data acquisition module 11 is used to:

[0233] For each product, obtain the inferred product scenario of the product and the filtering rules corresponding to each inferred product scenario; each inferred product scenario includes the associated products of the product under the inferred product scenario;

[0234] According to the product and the products associated with the product, the object logistics behavior relationship data is filtered to obtain the initial logistics behavior relationship data of the product;

[0235] According to the filtering rules corresponding to each inferred product scenario, the initial logistics behavior relationship data is filtered to obtain the logistics behavior relationship data of the product.

[0236] In an optional embodiment, before obtaining the inferred product scenario of each product and the filtering rules corresponding to each inferred product scenario, the data acquisition module 11 is used to:

[0237] For each of the inferred product scenarios of the product, determine, based on the logistics information data, a consignment object that has a consignment relationship with the product and associated products under the inferred product scenario;

[0238] Determine the number of times each consignment object consigns the product and the associated products under the inferred product scenario within a set time period;

[0239] According to the number of times each consignment object consigns the product within a set time period and consigns the associated products under the inferred product scenario, a filtering rule for the product under the inferred product scenario is obtained.

[0240] The embodiments of the present invention provide a product recommendation method, device, electronic device and readable storage medium. By constructing logistics behavior relationship data, after obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined based on the consignment relationship between the associated products associated with the product to be recommended in the logistics behavior relationship data and each candidate recommendation object, and the target recommendation object is obtained based on the correlation between each candidate recommendation object and the product to be recommended, and the product to be recommended is recommended to the target recommendation object. In this way, based on the consignment relationship between the associated product and each candidate recommendation object, an object with potential demand for the product can be obtained, so that a more comprehensive recommendation object can be obtained, and the comprehensiveness of the recommendation can be improved. At the same time, the embodiments of the present invention can obtain the real needs of customers through the logistics behavior relationship data, so that product recommendations based on the logistics behavior relationship data can improve the accuracy and rationality of product recommendations.

[0241] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the product recommendation device described above can refer to the corresponding process in the aforementioned method, and will not be described in detail here.

[0242] Based on the above, this embodiment provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the product recommendation method of any of the aforementioned embodiments is implemented.

[0243] The readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0244] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the readable storage medium described above can refer to the corresponding process in the aforementioned method, and will not be described in detail here.

[0245] In summary, the embodiments of the present invention provide a product recommendation method, device, electronic device and readable storage medium. By constructing logistics behavior relationship data, after obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the correlation between each candidate recommendation object and the product to be recommended can be determined based on the consignment relationship between the associated products associated with the product to be recommended in the logistics behavior relationship data and each candidate recommendation object, and the target recommendation object is obtained based on the correlation between each candidate recommendation object and the product to be recommended, and the product to be recommended is recommended to the target recommendation object. In this way, based on the consignment relationship between the associated product and each candidate recommendation object, an object with potential demand for the product can be obtained, so that a more comprehensive recommendation object can be obtained, and the comprehensiveness of the recommendation can be improved. At the same time, the embodiments of the present invention can obtain the real needs of customers through the logistics behavior relationship data, so that product recommendations based on the logistics behavior relationship data can improve the accuracy and rationality of product recommendations.

[0246] The above is a detailed introduction to a product recommendation method, device and electronic device provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the technical solution and its core idea of ​​the present application. Ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some of the technical features therein with equivalents; and these modifications or replacements do not make the essence of the corresponding technical solution deviate from the scope of the technical solution of the embodiments of the present application.

Claims

1. A product recommendation method, characterized in that: The method comprises: Acquire the product to be recommended and the logistics behavior relationship data of the product to be recommended; the logistics behavior relationship data includes the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object; Determining the relevance between each of the candidate recommendation objects and the product to be recommended according to the consignment relationship between the associated products associated with the product to be recommended and each of the candidate recommendation objects; According to the relevance between each of the candidate recommendation objects and the product to be recommended, a target recommendation object is obtained, and the product to be recommended is recommended to the target recommendation object.

2. The product recommendation method according to claim 1, characterized in that: The consignment relationship between the associated product associated with the product to be recommended and each of the candidate recommendation objects includes the number of consignments; The determining the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated product associated with the product to be recommended and each candidate recommendation object comprises: For each of the candidate recommendation objects, determine a weight coefficient of a target-related product corresponding to the consignment relationship of the candidate recommendation object and a consignment frequency index value of the target-related product; Obtaining a consignment frequency coefficient of the target-related product according to the number of consignments of the target-related product by the candidate recommendation object and the consignment frequency index value of the target-related product; The relevance between the candidate recommendation object and the product to be recommended is determined according to the weight coefficient and the consignment frequency coefficient of the target-related product.

3. The product recommendation method according to claim 2, characterized in that: The step of determining the weight coefficient of the target-related product corresponding to the consignment relationship of the candidate recommendation object and the consignment frequency index value of the target-related product includes: Determine the inferred product scenario corresponding to the target-related product, and obtain a weight coefficient of the target-related product according to the inferred product scenario; different inferred product scenarios have different weight coefficients; According to the consignment relationship between the target-related product and each of the candidate recommendation objects, determine the number of consignments of each of the candidate recommendation objects to the target-related product; According to the number of times each of the candidate recommendation objects has consigned the target-related product, an index value of the number of times the target-related product has been consigned is obtained.

4. The product recommendation method according to claim 1, characterized in that: Before obtaining the product to be recommended and the logistics behavior relationship data of the product to be recommended, the method further includes: Obtaining logistics information data; the logistics information data includes the products consigned by each object and the number of consignments of the consigned products; Determine the object logistics behavior relationship data according to the products entrusted by each object and the number of times the entrusted products are entrusted; the object logistics behavior relationship data includes the consignment relationship between each product and each object; each consignment relationship includes the number of times the object corresponding to the consignment relationship consigns the product corresponding to the consignment relationship; For each of the products, the associated products associated with the product are determined, and the object logistics behavior relationship data is filtered according to the product and the associated products associated with the product to obtain the logistics behavior relationship data of the product.

5. The product recommendation method according to claim 4, characterized in that: The step of determining, for each of the products, associated products associated with the product includes: Acquire product transaction data, where the product transaction data includes products traded by each object; For each product, according to the product transaction data, determine the target transaction object that has demand for the product; Determine the target product consigned by the target transaction object according to the logistics information data; According to the target product entrusted by the target transaction object, associated products associated with the product are obtained.

6. The product recommendation method according to claim 4, characterized in that: The filtering of the object logistics behavior relationship data according to the product and the associated products associated with the product to obtain the logistics behavior data of the product includes: For each of the products, obtaining a presumed product scenario of the product and a filtering rule corresponding to each of the presumed product scenarios; each of the presumed product scenarios includes associated products of the product under the presumed product scenario; According to the product and the products associated with the product, the object logistics behavior relationship data is filtered to obtain the initial logistics behavior relationship data of the product; The initial logistics behavior relationship data is filtered according to the filtering rules corresponding to each of the inferred product scenarios to obtain the logistics behavior relationship data of the product.

7. The product recommendation method according to claim 6, characterized in that: Before obtaining, for each of the products, the inferred product scenarios of the product and the filtering rules corresponding to each of the inferred product scenarios, the method further includes: For each of the inferred product scenarios of the product, determine, based on the logistics information data, a consignment object that has a consignment relationship with the product and associated products under the inferred product scenario; Determine the number of times each consignment object consigns the product and the associated products under the inferred product scenario within a set time period; According to the number of times each of the consignment objects consigns the product within a set time period and consigns the associated products under the inferred product scenario, a filtering rule for the product under the inferred product scenario is obtained.

8. A product recommendation method, characterized in that: The method comprises: Acquire a target recommendation object for which product recommendation is required and logistics behavior relationship data of the target recommendation object; the logistics behavior relationship data includes the consignment relationship between the target recommendation object and each candidate recommendation product; Determining the relevance between the target recommendation object and each of the candidate recommendation products according to the consignment relationship between the target recommendation object and each of the candidate recommendation products; According to the relevance between the target recommendation object and each of the candidate recommendation products, a target recommendation product is obtained, and the target recommendation product is recommended to the target recommendation object.

9. A product recommendation device, characterized in that: The product recommendation device comprises: A data acquisition module, used to acquire the product to be recommended and the logistics behavior relationship data of the product to be recommended; the logistics behavior relationship data includes the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object; A data processing module, configured to determine the relevance between each candidate recommendation object and the product to be recommended according to the consignment relationship between the associated products associated with the product to be recommended and each candidate recommendation object; The product recommendation module is used to obtain a target recommendation object according to the relevance between each candidate recommendation object and the product to be recommended, and recommend the product to be recommended to the target recommendation object.

10. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the product recommendation method according to any one of claims 1 to 7 or the product recommendation method according to claim 8 is implemented.

Images