CN114078037A - Commodity recommendation method and device based on label vectorization - Google Patents
Commodity recommendation method and device based on label vectorization Download PDFInfo
- Publication number
- CN114078037A CN114078037A CN202010819450.9A CN202010819450A CN114078037A CN 114078037 A CN114078037 A CN 114078037A CN 202010819450 A CN202010819450 A CN 202010819450A CN 114078037 A CN114078037 A CN 114078037A
- Authority
- CN
- China
- Prior art keywords
- commodity
- historical
- released
- user
- recommendation text
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0623—Item investigation
- G06Q30/0625—Directed, with specific intent or strategy
- G06Q30/0627—Directed, with specific intent or strategy using item specifications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a commodity recommendation method and device based on label vectorization, wherein the method comprises the following steps: vectorizing the commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched; obtaining historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and obtaining historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors; inputting a recommended text vector of the commodity to be released and historical releasing user response information into a deep neural network obtained through training to obtain a releasing user label of the commodity to be released; and sequencing the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sequencing result.
Description
Technical Field
The invention relates to the technical field of text classification, in particular to a commodity recommendation method and device based on label vectorization.
Background
The existing method for recommending commodities generally adopts the following modes:
1. the method comprises the steps of collecting evaluation information of users on different types of commodities, carrying out portrait analysis processing on the different types of commodities according to the evaluation information of the users, carrying out portrait analysis processing on the different users according to user transaction records and social information, matching the portraits of the different types of commodities with the portraits of the different users, and recommending the commodities for the users.
2. And establishing a commodity model according to all commodity information to form a commodity vector, acquiring purchased commodity information according to a user purchase record, calculating a preference vector of the user according to the purchased commodity and the commodity attributes in the commodity model, and recommending commodities to the user from commodities which the user does not purchase based on the proximity degree of the commodity vector and the preference vector.
3. Portraying the user through the user's behavioral records (browsing, placing orders, collecting, paying, adding to a shopping cart, etc.); generating a grade of the user for the commodity according to the relation between the user and the commodity, and then generating a grade matrix of the user for the commodity according to collaborative filtering; and generating recommendation data by combining the user portrait and the scoring matrix of the user for the commodity.
However, the above recommendation of the product has the following problems:
1. the commodity requirements are matched according to the evaluation information, purchase records, behavior records and the like of the user on the commodities, so that the user is required to purchase the commodities or generate evaluation on a platform or have behaviors of browsing, ordering and the like, and the matching cannot be identified for the newly-on-line commodities or the newly-registered user;
2. the method can only identify and match the platform users to recommend the commodities, can not identify unregistered users and users who have not purchased the commodities on the platform, can not mine the requirements of potential customers, and has the advantages of single technical function and low economic benefit.
Disclosure of Invention
In view of the above problems, the present invention has been made to provide a method and apparatus for recommending goods based on tag vectorization, which overcome or at least partially solve the above problems.
According to an aspect of the present invention, there is provided a method for recommending a product based on tag vectorization, including:
vectorizing the commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched;
obtaining historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and obtaining historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors;
inputting a recommended text vector of the commodity to be released and historical releasing user response information into a deep neural network obtained through training to obtain a releasing user label of the commodity to be released;
and sequencing the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sequencing result.
According to another aspect of the present invention, there is provided a tag vectorization-based item recommendation apparatus, including:
the vectorization module is suitable for vectorizing the commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched;
the acquisition module is suitable for acquiring historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and acquiring historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors;
the label module is suitable for inputting the recommended text vector of the commodity to be released and the historical releasing user response information into the deep neural network obtained by training to obtain a label of the releasing user of the commodity to be released;
and the sorting module is suitable for sorting the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sorting result.
According to still another aspect of the present invention, there is provided an electronic apparatus including: the system comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is used for storing at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the commodity recommendation method based on the label vectorization.
According to still another aspect of the present invention, a computer storage medium is provided, where at least one executable instruction is stored in the storage medium, and the executable instruction causes a processor to perform operations corresponding to the tag vectorization based product recommendation method as described above.
According to the commodity recommendation method and device based on the label vectorization, vectorization processing is carried out on a commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched; obtaining historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and obtaining historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors; inputting a recommended text vector of the commodity to be released and historical releasing user response information into a deep neural network obtained through training to obtain a releasing user label of the commodity to be released; and sequencing the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sequencing result. The user corresponding to the commodity to be released is obtained based on the information of the historical commodity, the historical commodity releasing user and the like through the commodity recommendation text to be released and the historical commodity recommendation text vector similar to the commodity recommendation text vector to be released, the user does not need to have historical behaviors on the commodity, the commodity does not need to have released historical data and the like, and therefore the problem of cold start of the user or the commodity is solved.
The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
FIG. 1 illustrates a flow diagram of a method for tag vectorization based recommendation of goods, according to an embodiment of the present invention;
FIG. 2 illustrates a functional block diagram of a product recommendation device based on tag vectorization according to an embodiment of the present invention;
fig. 3 shows a schematic structural diagram of an electronic device according to an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
Fig. 1 shows a flowchart of a method for recommending goods based on tag vectorization according to an embodiment of the present invention. As shown in fig. 1, the method for recommending a product based on tag vectorization specifically includes the following steps:
and step S101, performing vectorization processing on the commodity recommendation text to be released to obtain a commodity recommendation text vector to be released.
When vectorizing the recommended texts of the commodities to be launched, vectorizing the recommended texts of the commodities to be launched by using a Bert network to obtain recommended text vectors of the commodities to be launched.
The Bert network can adopt a seq2seq sequence mode and an attention mechanism to carry out vectorization processing on the recommended texts of the commodities to be launched, sequentially reads each character in the recommended texts of the commodities to be launched without carrying out word segmentation on the characters, can directly carry out vectorization processing on the whole recommended texts of the commodities to be launched, and improves the accuracy of vectorization processing. The vectorization processing process can be free from the influence of special words such as stop words and language-atmosphere words possibly contained in the commodity recommendation text to be launched, a word segmentation dictionary does not need to be preset, and the method is accurate and efficient. The Bert network can be trained by directly using the text information such as the existing general corpus, Wikipedia corpus and the like, so that the Bert network can be suitable for different commodity recommendation texts, and can also be added with new commodity recommendation texts according to actual requirements, and the accuracy of vectorization processing of the Bert network can be continuously optimized by accumulating and analyzing historical commodity recommendation texts.
Further, before vectorization processing is performed on the recommended texts of the commodities to be released, text check can be performed on the recommended texts of the commodities to be released first in order to ensure that the recommended texts of the commodities to be released can be better matched with users. The text check may be set according to specific implementation situations, for example, the text check includes whether the used word or phrase is standard, meets the recommendation standard, contains non-standard phrases, and the like, and is not limited herein.
Step S102, obtaining a historical commodity recommendation text vector similar to the commodity recommendation text vector to be released, and obtaining a historical commodity releasing user label and historical releasing user response information of the historical commodity recommendation text vector.
And after the commodity recommendation text vector to be released is obtained, calculating the vector distance between the historical commodity recommendation text vector and the commodity recommendation text vector to be released according to the historical commodity recommendation text vector. The calculation of the vector distance may be performed by using a vector distance formula, and will not be described herein. And determining the similarity between the historical commodity recommendation text vector and the commodity recommendation text vector to be released according to the vector distance. Wherein 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. And obtaining historical commodity recommendation text vectors with similarity higher than a preset threshold value, and simultaneously obtaining historical commodity release user tags and historical release user response information of the historical commodity recommendation text vectors.
The commodity release user label comprises an industry label to which a user belongs, a user behavior label, a user position label, a time label and the like. The commodity releasing user label can obtain the information of the basic attribute, the consumption habit and the commodity preference of the user under the condition of the authorization permission of the user by means of the operator big data, and a user label library can be created according to the user information to record an industry label, a user behavior label, a user position label, a time label and the like of the user. After the commodity is released, the commodity releasing user label can be recorded according to the user releasing the commodity. The industry label of the user can be determined according to the occupation, the favorite and the interest of the user; the user behavior tag can be determined according to behaviors of browsing, clicking, purchasing and the like of the user on the commodity; the user position label is determined according to the corresponding position information obtained by the mobile equipment used by the user; the time tag can determine the appropriate time for commodity release according to the time for the user to conduct behavior operation on the commodity by using the mobile device, so that the situation that the user neglects to release the commodity is avoided, the commodity browsing possibility of the user is improved, and the commodity release time can be recorded by the time tag. The historical releasing user response information is response information of the user to the released commodity after the commodity is released to the user. Specifically, the historical release user response information includes behavior information of browsing, clicking, purchasing and the like of the released commodity by the user, such as browsing time, browsing stay time, browsing frequency, clicking time, purchasing times, purchasing quantity and the like. For example, in implementation, according to specific implementation conditions, a suitable commodity release user tag and historical release user response information may be selected.
The commodity releasing user label and the historical releasing user response information are determined after various information of releasing users is collected by means of operator big data after the historical commodities are released.
Step S103, inputting the recommended text vector of the commodity to be released and the historical releasing user response information into the deep neural network obtained through training, and obtaining a label of the releasing user of the commodity to be released.
The deep neural network can be trained in advance, and the training process of the deep neural network specifically comprises the following steps: the method comprises the steps of firstly obtaining historical commodity recommendation texts, and conducting vectorization processing on the historical commodity recommendation texts to obtain historical commodity recommendation text vectors. Here, when the vectorization processing is performed, the vectorization processing may be performed by using a Bert network to obtain the historical product recommendation text vector. And then inputting the historical commodity recommendation text vector and the collected historical release user response information as sample data into the deep neural network to be trained, and adjusting training parameters of the deep neural network according to the historical commodity release user label to obtain the trained deep neural network.
After the trained deep neural network is obtained, inputting the recommended text vector of the commodity to be released and the historical releasing user response information into the trained deep neural network, so as to obtain the label of the commodity to be released. Here, the commodity launching user tag to be launched is predicted in consideration of the similarity between the commodity recommendation text vector to be launched and the historical commodity recommendation text vector according to the historical launching user response information of the historical commodity recommendation text vector similar to the commodity recommendation text vector to be launched.
And step S104, sequencing the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sequencing result.
After the commodity releasing user labels to be released are obtained, calculating the weight corresponding to each label in the historical commodity releasing user labels according to the similarity between the historical commodity recommendation text vector and the commodity recommendation text vector to be released and the weight of each label in the historical commodity releasing user labels. And then, sequencing according to the weight corresponding to each label in the historical commodity releasing user labels and the weight of the commodity releasing user label to be released to obtain a user label sequence corresponding to the commodity to be released, and determining a user industry label, a user behavior label, a user position label, a time label and the like corresponding to the commodity to be released. According to the tags contained in the user tag sequence corresponding to the commodity to be released, the user corresponding to the commodity to be released can be determined by combining the user tag library, and therefore the recommendation text of the commodity to be released can be released to the corresponding user.
According to the commodity recommendation method based on the label vectorization, provided by the invention, a commodity recommendation text to be launched is subjected to vectorization processing to obtain a commodity recommendation text vector to be launched; obtaining historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and obtaining historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors; inputting a recommended text vector of the commodity to be released and historical releasing user response information into a deep neural network obtained through training to obtain a releasing user label of the commodity to be released; and sequencing the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sequencing result. The user corresponding to the commodity to be released is obtained based on the information of the historical commodity, the historical commodity releasing user and the like through the commodity recommendation text to be released and the historical commodity recommendation text vector similar to the commodity recommendation text vector to be released, the user does not need to have historical behaviors on the commodity, the commodity does not need to have released historical data and the like, and therefore the problem of cold start of the user or the commodity is solved.
Fig. 2 shows a functional block diagram of a product recommendation device based on tag vectorization according to an embodiment of the present invention. As shown in fig. 2, the product recommendation apparatus based on tag vectorization includes the following modules:
the vectorization module 210 is adapted to: vectorizing the commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched;
the obtaining module 220 is adapted to: obtaining historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and obtaining historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors;
the tag module 230 is adapted to: inputting a recommended text vector of the commodity to be released and historical releasing user response information into a deep neural network obtained through training to obtain a releasing user label of the commodity to be released;
the sorting module 240 is adapted to: and sequencing the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to the sequencing result.
Optionally, the apparatus further comprises: a training module 250.
The training module 250 is adapted to: acquiring a historical commodity recommendation text, and performing vectorization processing on the historical commodity recommendation text to obtain a historical commodity recommendation text vector; and inputting the historical commodity recommendation text vector and the historical releasing user response information as sample data into the deep neural network to be trained, and adjusting the training parameters of the deep neural network according to the historical commodity releasing user label to obtain the trained deep neural network.
Optionally, the obtaining module 220 is further adapted to: calculating a vector distance between the historical commodity recommendation text vector and the commodity recommendation text vector to be released; determining the similarity between the historical commodity recommendation text vector and the commodity recommendation text vector to be released according to the vector distance; the similarity is inversely proportional to the vector distance; and obtaining historical commodity recommendation text vectors with similarity higher than a preset threshold value, and obtaining historical commodity release user tags and historical release user response information of the historical commodity recommendation text vectors.
Optionally, the sorting module 240 is further adapted to: calculating the weight corresponding to each label in the labels of the historical commodity releasing users according to the similarity between the historical commodity recommendation text vector and the commodity recommendation text vector to be released and the weight of each label in the labels of the historical commodity releasing users; sequencing according to the weight corresponding to each label in the historical commodity releasing user labels and the weight of the commodity releasing user label to be released to obtain a user label sequence corresponding to the commodity to be released; and determining the user corresponding to the commodity to be released according to the user tag sequence corresponding to the commodity to be released.
Optionally, the apparatus further comprises: the checking module 260.
The inspection module 260 is adapted to: and performing text check on the recommended texts of the commodities to be released.
The descriptions of the modules refer to the corresponding descriptions in the method embodiments, and are not repeated herein.
The application also provides a non-volatile computer storage medium, where at least one executable instruction is stored in the computer storage medium, and the computer executable instruction may execute the tag vectorization-based commodity recommendation method in any of the above method embodiments.
Fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and the specific embodiment of the present invention does not limit the specific implementation of the electronic device.
As shown in fig. 3, the electronic device may include: a processor (processor)302, a communication Interface 304, a memory 306, and a communication bus 308.
Wherein:
the processor 302, communication interface 304, and memory 306 communicate with each other via a communication bus 308.
A communication interface 304 for communicating with network elements of other devices, such as clients or other servers.
The processor 302 is configured to execute the program 310, and may specifically execute relevant steps in the above-mentioned tag-vectorization-based product recommendation method embodiment.
In particular, program 310 may include program code comprising computer operating instructions.
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 an embodiment of the present invention. The electronic device comprises one or more processors, which can be the same type of processor, such as one or more CPUs; or may be different types of processors such as one or more CPUs and one or more ASICs.
And a memory 306 for storing a program 310. Memory 306 may comprise high-speed RAM memory and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
The program 310 may be specifically configured to enable the processor 302 to execute the product recommendation method based on tag vectorization in any of the method embodiments described above. For specific implementation of each step in the program 310, reference may be made to corresponding steps and corresponding descriptions in units in the above tag-vectorization-based product recommendation embodiment, which are not described herein again. It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described devices and modules may refer to the corresponding process descriptions in the foregoing method embodiments, and are not described herein again.
The algorithms and displays presented herein are not inherently related to any particular computer, virtual machine, or other apparatus. Various general purpose systems may also be used with the teachings herein. The required structure for constructing such a system will be apparent from the description above. Moreover, the present invention is not directed to any particular programming language. It is appreciated that a variety of programming languages may be used to implement the teachings of the present invention as described herein, and any descriptions of specific languages are provided above to disclose the best mode of the invention.
In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and furthermore they may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the claims, any of the claimed embodiments may be used in any combination.
The various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It will be appreciated by those skilled in the art that a microprocessor or Digital Signal Processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in a tag vectorization based merchandise recommendation device according to embodiments of the present invention. The present invention may also be embodied as apparatus or device programs (e.g., computer programs and computer program products) for performing a portion or all of the methods described herein. Such programs implementing the present invention may be stored on computer-readable media 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 in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.
Claims (10)
1. A commodity recommendation method based on label vectorization is characterized by comprising the following steps:
vectorizing the commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched;
obtaining a historical commodity recommendation text vector similar to a commodity recommendation text vector to be released, and obtaining a historical commodity releasing user tag and historical releasing user response information of the historical commodity recommendation text vector;
inputting the recommended text vector of the commodity to be released and historical releasing user response information into a deep neural network obtained by training to obtain a releasing user label of the commodity to be released;
and sequencing the historical commodity releasing user labels and the commodity releasing user labels to be released according to the weight, and determining users corresponding to the commodities to be released according to a sequencing result.
2. The method of claim 1, wherein the training process comprises:
acquiring a historical commodity recommendation text, and performing vectorization processing on the historical commodity recommendation text to obtain a historical commodity recommendation text vector;
and inputting the historical commodity recommendation text vector and the historical releasing user response information serving as sample data into the deep neural network to be trained, and adjusting training parameters of the deep neural network according to the historical commodity releasing user label to obtain the trained deep neural network.
3. The method according to claim 1 or 2, wherein the vectorization process is specifically: and vectorizing by using the Bert network.
4. The method according to claim 1, wherein the commodity release user tags comprise user affiliated industry tags, user behavior tags, user location tags and/or time tags; the historical release user response information comprises: browsing, clicking and/or purchasing behavior information of the goods by the user.
5. The method of claim 1, wherein the obtaining historical commodity recommendation text vectors similar to the commodity recommendation text vector to be released, and obtaining historical commodity release user tags and historical release user response information of the historical commodity recommendation text vectors further comprises:
calculating a vector distance between the historical commodity recommendation text vector and the commodity recommendation text vector to be released;
determining the similarity between the historical commodity recommendation text vector and the commodity recommendation text vector to be released according to the vector distance; the similarity is inversely proportional to the vector distance;
and obtaining historical commodity recommendation text vectors with similarity higher than a preset threshold value, and obtaining historical commodity release user tags and historical release user response information of the historical commodity recommendation text vectors.
6. The method according to claim 5, wherein the step of sorting the historical commodity release user tags and the commodity release user tags to be released according to weight, and determining the users corresponding to the commodities to be released according to the sorting result further comprises:
calculating the weight corresponding to each label in the labels of the historical commodity releasing users according to the similarity between the historical commodity recommendation text vector and the commodity recommendation text vector to be released and the weight of each label in the labels of the historical commodity releasing users;
sequencing according to the weight corresponding to each label in the historical commodity releasing user labels and the weight of the commodity releasing user label to be released to obtain a user label sequence corresponding to the commodity to be released;
and determining the user corresponding to the commodity to be released according to the user tag sequence corresponding to the commodity to be released.
7. The method of claim 1, further comprising:
and performing text check on the recommended texts of the commodities to be released.
8. An article recommendation device based on label vectorization, the device comprising:
the vectorization module is suitable for vectorizing the commodity recommendation text to be launched to obtain a commodity recommendation text vector to be launched;
the acquisition module is suitable for acquiring historical commodity recommendation text vectors similar to the commodity recommendation text vectors to be released, and acquiring historical commodity releasing user labels and historical releasing user response information of the historical commodity recommendation text vectors;
the label module is suitable for inputting the recommended text vector of the commodity to be released and the historical releasing user response information into a deep neural network obtained by training to obtain a label of the releasing user of the commodity to be released;
and the sorting module is suitable for sorting the historical commodity releasing user tags and the commodity releasing user tags to be released according to the weight, and determining users corresponding to the commodities to be released according to a sorting result.
9. An electronic device, comprising: the system comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is used for storing at least one executable instruction, and the executable instruction causes the processor to execute the operation corresponding to the commodity recommendation method based on label vectorization in any one of claims 1-7.
10. A computer storage medium having at least one executable instruction stored therein, the executable instruction causing a processor to perform operations corresponding to the tag vectorization-based item recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010819450.9A CN114078037A (en) | 2020-08-14 | 2020-08-14 | Commodity recommendation method and device based on label vectorization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010819450.9A CN114078037A (en) | 2020-08-14 | 2020-08-14 | Commodity recommendation method and device based on label vectorization |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114078037A true CN114078037A (en) | 2022-02-22 |
Family
ID=80279524
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010819450.9A Pending CN114078037A (en) | 2020-08-14 | 2020-08-14 | Commodity recommendation method and device based on label vectorization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114078037A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114638646A (en) * | 2022-03-25 | 2022-06-17 | 广州华多网络科技有限公司 | Advertisement putting recommendation method and device, equipment, medium and product thereof |
| CN114936911A (en) * | 2022-07-26 | 2022-08-23 | 成都纳宝科技有限公司 | Systematic and intelligent marketing promotion system |
-
2020
- 2020-08-14 CN CN202010819450.9A patent/CN114078037A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114638646A (en) * | 2022-03-25 | 2022-06-17 | 广州华多网络科技有限公司 | Advertisement putting recommendation method and device, equipment, medium and product thereof |
| CN114936911A (en) * | 2022-07-26 | 2022-08-23 | 成都纳宝科技有限公司 | Systematic and intelligent marketing promotion system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107833082B (en) | Commodity picture recommendation method and device | |
| CN109816134B (en) | Method and device for predicting delivery address and storage medium | |
| CN108876479B (en) | Channel attribution method and device for object entity | |
| CN110827112B (en) | Deep learning commodity recommendation method and device, computer equipment and storage medium | |
| CN110807669B (en) | Cross-platform user information management method and device | |
| CN109360089A (en) | Credit risk prediction technique and device | |
| CN109118316B (en) | Method and device for identifying authenticity of online shop | |
| CN110580489B (en) | Data object classification system, method and equipment | |
| CN111666757A (en) | Commodity comment emotional tendency analysis method, device and equipment and readable storage medium | |
| CN114078037A (en) | Commodity recommendation method and device based on label vectorization | |
| CN111681085A (en) | Commodity pushing method and device, server and readable storage medium | |
| EP3543943A1 (en) | Purchase information utilization system, purchase information utilization method, and program | |
| CN112465598A (en) | User portrait generation method, device and equipment and computer storage medium | |
| CN111582932A (en) | Inter-scene information pushing method and device, computer equipment and storage medium | |
| KR101784559B1 (en) | Method and apparatus for analyzing pattern of consumption/interest | |
| CN110362702B (en) | Picture management method and equipment | |
| JP7043453B2 (en) | Information processing equipment, systems, information processing methods, and programs | |
| CN111292153B (en) | Information recommendation method, device, system and storage medium | |
| CN115641179A (en) | Information pushing method and device and electronic equipment | |
| CN109472025B (en) | Dish name extraction method and device | |
| CN112418969A (en) | Commodity matching method and device and computer equipment | |
| CN115496420A (en) | New user quality evaluation method and device, computer equipment and storage medium | |
| CN111639989B (en) | Commodity recommendation method and readable storage medium | |
| KR101815558B1 (en) | Consumption pattern analysis and marketing system and method for the same | |
| CN109993592A (en) | Information-pushing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |