Commodity recommendation method and device based on label vectorization

Abstract

The present invention discloses a commodity recommendation method and device based on label vectorization, wherein the method comprises: vectorizing a commodity recommendation text to be placed to obtain a commodity recommendation text vector to be placed; obtaining a historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, and obtaining a historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector; inputting the commodity recommendation text vector to be placed and the historical placement user response information into a trained deep neural network to obtain a placement user label of the commodity to be placed; sorting the historical commodity placement user labels and the commodity placement user labels to be placed according to weights, and determining the user corresponding to the commodity to be placed according to the sorting result.

Description

Technical Field

[0001] The present invention relates to the technical field of text classification, and in particular to a product recommendation method and device based on label vectorization. Background Art

[0002] The existing methods for recommending products generally adopt the following methods:

[0003] 1. Collect user evaluation information on different types of goods, then conduct portrait analysis and processing on different types of goods based on user evaluation information, and then conduct portrait analysis and processing on different users based on user transaction records and social information. Then, match the portraits of different types of goods with the portraits of different users to recommend products to users.

[0004] 2. Build a product model based on all product information to form a product vector. Then obtain the purchased product information based on the user's purchase record. Calculate the user's preference vector through the purchased products and the product attributes in the product model. Recommend products to the user based on the degree of proximity between the product vector and the preference vector from the products that the user has not purchased yet.

[0005] 3. Create user profiles through user behavior records (browsing, ordering, collecting, paying, and adding to shopping carts, etc.); generate user ratings for products based on the relationship between users and products, and then generate a user rating matrix for products based on collaborative filtering; combine user profiles and user rating matrix for products to generate recommendation data.

[0006] However, there are the following problems when recommending products:

[0007] 1. Matching product demand based on user evaluation information, purchase history, behavior history, etc., which requires that users have purchased products or reviewed them on the platform, or have browsed or placed orders, etc., and cannot identify and match newly launched products or newly registered users;

[0008] 2. It can only identify and match recommended products for platform users, but cannot identify unregistered users or users who have not purchased products on the platform, and cannot tap into potential customer needs. It has a single technical function and low economic benefits. Summary of the invention

[0009] In view of the above problems, the present invention is proposed to provide a product recommendation method and device based on label vectorization that overcome the above problems or at least partially solve the above problems.

[0010] According to one aspect of the present invention, a product recommendation method based on label vectorization is provided, which includes:

[0011] Vectorize the product recommendation text to be placed to obtain a product recommendation text vector;

[0012] Obtaining historical product recommendation text vectors similar to the product recommendation text vector to be placed, and obtaining historical product placement user labels and historical placement user response information of the historical product recommendation text vectors;

[0013] Input the recommended text vector of the product to be launched and the historical user response information into the trained deep neural network to obtain the user label of the product to be launched;

[0014] The historical product placement user tags and the product placement user tags to be placed are sorted according to the weights, and the user corresponding to the product to be placed is determined according to the sorting result.

[0015] According to another aspect of the present invention, a product recommendation device based on label vectorization is provided, comprising:

[0016] A vectorization module, adapted to perform vectorization processing on the product recommendation text to be placed, and obtain a vector of the product recommendation text to be placed;

[0017] An acquisition module, adapted to acquire a historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, and acquire a historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector;

[0018] A label module, adapted to input the recommendation text vector of the product to be placed and the historical user response information of the placement into the trained deep neural network to obtain the user label of the product to be placed;

[0019] The sorting module is adapted to sort the historical product placement user tags and the product placement user tags to be placed according to the weights, and determine the user corresponding to the product to be placed according to the sorting result.

[0020] According to another aspect of the present invention, there is provided an electronic device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0021] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned product recommendation method based on label vectorization.

[0022] According to yet another aspect of the present invention, a computer storage medium is provided, wherein the storage medium stores at least one executable instruction, and the executable instruction enables a processor to perform operations corresponding to the above-mentioned product recommendation method based on label vectorization.

[0023] According to the product recommendation method and device based on label vectorization of the present invention, the product recommendation text to be placed is vectorized to obtain the product recommendation text vector to be placed; the historical product recommendation text vector similar to the product recommendation text vector to be placed is obtained, and the historical product placement user label and historical placement user response information of the historical product recommendation text vector are obtained; the product recommendation text vector to be placed and the historical placement user response information are input into the trained deep neural network to obtain the user label of the product to be placed; the historical product placement user label and the product placement user label to be placed are sorted according to the weight, and the user corresponding to the product to be placed is determined according to the sorting result. Through the product recommendation text to be placed itself and the historical product recommendation text vector similar to the product recommendation text vector to be placed, the user corresponding to the product to be placed is obtained based on the information of historical products, historical product placement users, etc., without the user having to have historical behavior on the product, and the product having to have historical data of being placed, etc., thereby solving the user or product cold start problem.

[0024] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0026] Figure 1 A flowchart of a method for recommending products based on label vectorization according to an embodiment of the present invention is shown;

[0027] Figure 2 A functional block diagram of a commodity recommendation device based on label vectorization according to an embodiment of the present invention is shown;

[0028] Figure 3 A schematic structural diagram of an electronic device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0029] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0030] Figure 1 FIG. 4 is a flowchart of a method for recommending products based on label vectorization according to an embodiment of the present invention. Figure 1 As shown in FIG. 1 , the product recommendation method based on label vectorization specifically includes the following steps:

[0031] Step S101 , vectorizing the product recommendation text to be placed to obtain a product recommendation text vector to be placed.

[0032] When vectorizing the product recommendation text to be placed, the Bert network can be used to perform vectorization processing to obtain the product recommendation text vector to be placed.

[0033] The Bert network can use seq2seq sequence mode and attention mechanism to vectorize the product recommendation text to be released, read each word in the product recommendation text to be released in turn, without segmenting it, and directly vectorize the entire product recommendation text to be released to improve the accuracy of vectorization. The vectorization process is not affected by special words such as stop words and modal particles that may be contained in the product recommendation text to be released, and there is no need to pre-set the word segmentation dictionary, which is more accurate and efficient. The Bert network can directly use existing general corpus, Wikipedia corpus and other text information for training, so that it can be applied to different product recommendation texts. It can also add new product recommendation texts according to actual needs, etc. By accumulating and analyzing historical product recommendation texts, the vectorization processing accuracy of the Bert network can be continuously optimized.

[0034] Furthermore, before vectorizing the product recommendation text to be placed, in order to ensure that the product recommendation text to be placed can better match the user, the product recommendation text to be placed can be first checked. The text check can be set according to the specific implementation situation, such as whether the text check includes whether the words and sentences used are standardized, whether they meet the recommendation standards, whether they contain non-standard terms, etc., which are not limited here.

[0035] Step S102, obtaining historical product recommendation text vectors similar to the product recommendation text vector to be placed, and obtaining historical product placement user tags and historical placement user response information of the historical product recommendation text vectors.

[0036] After obtaining the product recommendation text vector to be placed, the vector distance between the historical product recommendation text vector and the product recommendation text vector to be placed is calculated based on the historical product recommendation text vector. The vector distance can be calculated using methods such as the vector distance formula, which will not be explained in detail here. Determine the similarity between the historical product recommendation text vector and the product recommendation text vector to be placed based on the vector distance. Among them, the similarity is inversely proportional to the vector distance. When the vector distance is smaller, the similarity is higher, and when the vector distance is larger, the similarity is lower. Obtain the historical product recommendation text vector whose similarity is higher than the preset threshold, and at the same time obtain the historical product placement user label and historical placement user response information of the historical product recommendation text vector.

[0037] Among them, the user tags for commodity placement include the industry tag to which the user belongs, the user behavior tag, the user location tag, the time tag, etc. The user tags for commodity placement can obtain the user's basic attributes, consumption habits, and commodity preferences with the help of the operator's big data, with the user's authorization and permission. Based on the above user information, a user tag library can be created to record the user's industry tag, user behavior tag, user location tag, time tag, etc. After the commodity is placed, the user tags for commodity placement can be recorded according to the user to whom the commodity is placed. The industry tag to which the user belongs can be determined based on the user's occupation, user preference, user interest, etc.; the user behavior tag can be determined based on the user's browsing, clicking, purchasing, etc. behaviors of the commodity; the user location tag can be determined based on the corresponding location information obtained by the mobile device used by the user; the time tag can determine the appropriate time to place the commodity based on the time when the user uses the mobile device to perform the behavior operation on the commodity, so as to avoid the user ignoring the placed commodity and increase the possibility of the user browsing the commodity. The time tag can record the time of commodity placement. The historical user response information for placement is the user's response information to the placed commodity after the commodity is placed to the user. Specifically, the historical delivery user response information includes the user's browsing, clicking, purchasing and other behavior information on the delivered products, such as browsing time, browsing dwelling time, browsing frequency, clicking time, purchasing time, purchase times, purchase quantity, etc. The above is an example, and during implementation, appropriate product delivery user tags and historical delivery user response information can be selected according to the specific implementation situation.

[0038] Product placement user tags and historical placement user response information are determined after historical products are placed, with the help of operator big data, by collecting various information of users who placed the products.

[0039] Step S103: input the recommendation text vector of the product to be placed and the historical user response information into the trained deep neural network to obtain the user label of the product to be placed.

[0040] The deep neural network can be trained in advance, and the training process of the deep neural network specifically includes: first obtaining the historical product recommendation text, vectorizing the historical product recommendation text, and obtaining the historical product recommendation text vector. Here, when performing the vectorization process, the Bert network can be used for vectorization to obtain the historical product recommendation text vector. Then, the historical product recommendation text vector and the collected historical delivery user response information are used as sample data and input into the deep neural network to be trained, and the training parameters of the deep neural network are adjusted according to the historical product delivery user labels to obtain the trained deep neural network.

[0041] After obtaining the trained deep neural network, the product recommendation text vector to be placed and the historical placement user response information are input into the trained deep neural network to obtain the user label of the product to be placed. Here, based on the historical placement user response information of the historical product recommendation text vector similar to the product recommendation text vector to be placed, the user label of the product to be placed is predicted taking into account the similarity between the product recommendation text vector to be placed and the historical product recommendation text vector.

[0042] Step S104, sorting the historical product placement user tags and the product placement user tags to be placed according to the weights, and determining the user corresponding to the product to be placed according to the sorting result.

[0043] After obtaining the user tag of the product to be placed, first calculate the weight corresponding to each tag in the historical product placement user tag based on the similarity between the historical product recommendation text vector and the product recommendation text vector to be placed, combined with the weight of each tag in the historical product placement user tag. Then, sort the weights corresponding to each tag in the historical product placement user tag and the weight of the product placement user tag to be placed to obtain the user tag sequence corresponding to the product to be placed, from which the user industry tag, user behavior tag, user location tag, time tag, etc. corresponding to the product to be placed can be determined. According to the above tags contained in the user tag sequence corresponding to the product to be placed, combined with the user tag library, the user corresponding to the product to be placed can be determined, so that the recommendation text of the product to be placed can be delivered to the corresponding user.

[0044] According to the commodity recommendation method based on label vectorization provided by the present invention, the commodity recommendation text to be placed is vectorized to obtain the commodity recommendation text vector to be placed; the historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed is obtained, and the historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector are obtained; the commodity recommendation text vector to be placed and the historical placement user response information are input into the trained deep neural network to obtain the user label of the commodity to be placed; the historical commodity placement user label and the commodity placement user label to be placed are sorted according to the weight, and the user corresponding to the commodity to be placed is determined according to the sorting result. Through the commodity recommendation text to be placed itself and the historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, the user corresponding to the commodity to be placed is obtained based on the information of historical commodities, historical commodity placement users, etc., without the need for the user to have historical behavior on the commodity, and the commodity to have historical data of being placed, etc., thereby solving the user or commodity cold start problem.

[0045] Figure 2 FIG. 4 shows a functional block diagram of a commodity recommendation device based on label vectorization according to an embodiment of the present invention. Figure 2 As shown, the product recommendation device based on label vectorization includes the following modules:

[0046] The vectorization module 210 is adapted to: perform vectorization processing on the commodity recommendation text to be placed, and obtain a commodity recommendation text vector to be placed;

[0047] The acquisition module 220 is adapted to: acquire a historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, and acquire a historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector;

[0048] The label module 230 is adapted to: input the recommendation text vector of the product to be placed and the historical user response information of the placement into the trained deep neural network to obtain the user label of the product to be placed;

[0049] The sorting module 240 is adapted to sort the historical commodity placement user tags and the commodity placement user tags to be placed according to the weights, and determine the user corresponding to the commodity to be placed according to the sorting result.

[0050] Optionally, the device further includes: a training module 250.

[0051] The training module 250 is suitable for: obtaining historical product recommendation texts, vectorizing the historical product recommendation texts, and obtaining historical product recommendation text vectors; using the historical product recommendation text vectors and historical delivery user response information as sample data, inputting them into the deep neural network to be trained, adjusting the training parameters of the deep neural network according to the historical product delivery user labels, and obtaining the trained deep neural network.

[0052] Optionally, the acquisition module 220 is further suitable for: calculating the vector distance between the historical product recommendation text vector and the product recommendation text vector to be placed; determining the similarity between the historical product recommendation text vector and the product recommendation text vector to be placed based on the vector distance; the similarity is inversely proportional to the vector distance; obtaining historical product recommendation text vectors whose similarity is higher than a preset threshold, and obtaining historical product placement user labels and historical placement user response information of the historical product recommendation text vectors.

[0053] Optionally, the sorting module 240 is further adapted to: calculate the weight corresponding to each tag in the historical product placement user tags according to the similarity between the historical product recommendation text vector and the product recommendation text vector to be placed, combined with the weight of each tag in the historical product placement user tags; perform sorting according to the weight corresponding to each tag in the historical product placement user tags and the weight of the product placement user tags to be placed, to obtain the user tag sequence corresponding to the product to be placed; determine the user corresponding to the product to be placed according to the user tag sequence corresponding to the product to be placed.

[0054] Optionally, the device further includes: a checking module 260 .

[0055] The checking module 260 is adapted to perform text checking on the commodity recommendation text to be placed.

[0056] The description of each module above refers to the corresponding description in the method embodiment and will not be repeated here.

[0057] The present application also provides a non-volatile computer storage medium, which stores at least one executable instruction, and the computer executable instruction can execute the product recommendation method based on label vectorization in any of the above method embodiments.

[0058] Figure 3 A schematic structural diagram of an electronic device according to an embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the electronic device.

[0059] like Figure 3As shown, the electronic device may include: a processor (processor) 302 , a communication interface (Communications Interface) 304 , a memory (memory) 306 , and a communication bus 308 .

[0060] in:

[0061] The processor 302 , the communication interface 304 , and the memory 306 communicate with each other via the communication bus 308 .

[0062] The communication interface 304 is used to communicate with other devices such as clients or other servers.

[0063] The processor 302 is used to execute the program 310, and specifically can execute the relevant steps in the above-mentioned embodiment of the product recommendation method based on label vectorization.

[0064] Specifically, the program 310 may include program codes, which include computer operation instructions.

[0065] The processor 302 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the electronic device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0066] The memory 306 is used to store the program 310. The memory 306 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0067] Program 310 can be specifically used to enable processor 302 to execute the product recommendation method based on label vectorization in any of the above-mentioned method embodiments. The specific implementation of each step in program 310 can refer to the corresponding descriptions in the corresponding steps and units in the above-mentioned product recommendation embodiment based on label vectorization, which will not be repeated here. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described devices and modules can refer to the corresponding process description in the aforementioned method embodiment, which will not be repeated here.

[0068] The algorithm and display provided herein are not inherently related to any particular computer, virtual system or other device. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious that the structure required for constructing such systems. In addition, the present invention is not directed to any specific programming language either. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description of the above specific languages ​​is for disclosing the best mode of the present invention.

[0069] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0070] Similarly, it should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the intention that the claimed invention requires more features than those expressly recited in each claim. Rather, as reflected in the claims, inventive aspects lie in less than all of the features of the individual embodiments previously disclosed. Therefore, the claims that follow the detailed description are hereby expressly incorporated into the detailed description, with each claim itself serving as a separate embodiment of the present invention.

[0071] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0072] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims, any one of the claimed embodiments may be used in any combination.

[0073] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components of the commodity recommendation device based on tag vectorization according to an embodiment of the present invention. The present invention may also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0074] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

Claims

1. A product recommendation method based on label vectorization, It is characterized in that Methods include: The product recommendation text to be placed is vectorized to obtain a product recommendation text vector; wherein each word in the product recommendation text to be placed is read in sequence, and the product recommendation text to be placed is vectorized as a whole; Acquire a historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, and acquire a historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector; Inputting the recommended text vector of the product to be placed and the historical user response information into the trained deep neural network to obtain the user label of the product to be placed; The historical commodity placement user tags and the commodity placement user tags to be placed are sorted according to weights, and the user corresponding to the commodity to be placed is determined according to the sorting result.

2. The method according to claim 1, It is characterized in that The training process includes: Obtain historical product recommendation texts, and vectorize the historical product recommendation texts to obtain historical product recommendation text vectors; The historical product recommendation text vector and the historical product delivery user response information are used as sample data and input into the deep neural network to be trained. The training parameters of the deep neural network are adjusted according to the historical product delivery user labels to obtain the trained deep neural network.

3. The method according to claim 1 or 2, It is characterized in that The vectorization processing specifically includes: performing vectorization processing using a Bert network.

4. The method according to claim 1, It is characterized in that The commodity delivery user tag includes a user industry tag, a user behavior tag, a user location tag and / or a time tag; the historical delivery user response information includes: user browsing, clicking and / or purchasing behavior information for commodities.

5. The method according to claim 1, It is characterized in that The step of obtaining a historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, and obtaining a historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector further includes: Calculate the vector distance between the historical product recommendation text vector and the product recommendation text vector to be released; Determine the similarity between the historical product recommendation text vector and the product recommendation text vector to be placed according to the vector distance; the similarity is inversely proportional to the vector distance; The historical commodity recommendation text vectors whose similarity is higher than a preset threshold are obtained, and the historical commodity placement user labels and historical placement user response information of the historical commodity recommendation text vectors are obtained.

6. The method according to claim 5, It is characterized in that The step of sorting the historical product placement user tags and the product placement user tags to be placed according to weights, and determining the user corresponding to the product to be placed according to the sorting result further includes: According to the similarity between the historical product recommendation text vector and the product recommendation text vector to be placed, combined with the weight of each tag in the historical product placement user tag, the weight corresponding to each tag in the product placement user tag to be placed is calculated; Sorting is performed according to the weights corresponding to the tags in the historical product placement user tags and the weights of the product placement user tags to be placed, so as to obtain the user tag sequence corresponding to the product to be placed; The user corresponding to the commodity to be placed is determined according to the user tag sequence corresponding to the commodity to be placed.

7. The method according to claim 1, It is characterized in that The method further comprises: Perform text check on the product recommendation text to be placed.

8. A product recommendation device based on label vectorization, It is characterized in that The device includes: A vectorization module is adapted to perform vectorization processing on the commodity recommendation text to be placed, and obtain a vector of the commodity recommendation text to be placed; wherein each word in the commodity recommendation text to be placed is read in sequence, and the commodity recommendation text to be placed is vectorized as a whole; An acquisition module, adapted to acquire a historical commodity recommendation text vector similar to the commodity recommendation text vector to be placed, and acquire a historical commodity placement user label and historical placement user response information of the historical commodity recommendation text vector; A label module, adapted to input the product recommendation text vector to be placed and historical user response information into the trained deep neural network to obtain a user label for the product to be placed; The sorting module is adapted to sort the historical commodity placement user tags and the commodity placement user tags to be placed according to weights, and determine the user corresponding to the commodity to be placed according to the sorting result.

9. An electronic device, include: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the product recommendation method based on label vectorization as described in any one of claims 1-7.

10. A computer storage medium, wherein at least one executable instruction is stored in the storage medium, and the executable instruction enables a processor to perform operations corresponding to the product recommendation method based on label vectorization as described in any one of claims 1 to 7.

Images