CN110706055A - Commodity information pushing method and device, storage medium and computer equipment - Google Patents
Commodity information pushing method and device, storage medium and computer equipment Download PDFInfo
- Publication number
- CN110706055A CN110706055A CN201910780785.1A CN201910780785A CN110706055A CN 110706055 A CN110706055 A CN 110706055A CN 201910780785 A CN201910780785 A CN 201910780785A CN 110706055 A CN110706055 A CN 110706055A
- Authority
- CN
- China
- Prior art keywords
- commodity
- information
- prediction model
- introduction text
- keyword
- 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
Abstract
The application discloses a commodity information pushing method and device, a storage medium and computer equipment, relates to the technical field of image recognition, and can improve the image recognition accuracy. The method comprises the following steps: constructing a data candidate set according to the acquired commodity demand text information; training the initialized commodity information prediction model by using the constructed data candidate set to obtain a trained commodity information prediction model; inputting commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information; and displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information respectively. The application is suitable for the accurate pushing of the mobile internet and the O2O mode E-commerce commodities.
Description
Technical Field
The present application relates to the field of information processing technologies, and in particular, to a method and an apparatus for pushing commodity information, a storage medium, and a computer device.
Background
With the rapid development of the mobile internet and the O2O model, the commodity categories of the e-commerce are more and more abundant, the data information volume is increased, and data mining becomes a main way for the e-commerce to improve the benefits. The user and the commodities are combined through data mining, different commodities can be pushed for different users, different requirements of the different users on the commodities are met, and shopping experience of the users is improved.
The existing method for recommending commodities according to user information mainly comprises two methods: 1) recommending corresponding commodities for the user based on the historical purchasing behavior information of the user; 2) and recommending the corresponding commodities for the user based on the quantifiable attribute information of the commodities. Recommending corresponding commodities for the user based on the historical purchasing behavior information of the user, wherein the problem of cold start exists; the method is not suitable for the situation of massive commodities, and the two methods cannot realize an accurate pushing function according to the real-time requirements of the user.
Disclosure of Invention
In view of this, the present application provides a method and an apparatus for pushing commodity information, a storage medium, and a computer device, and mainly aims to solve the technical problem that the existing method for pushing commodity information cannot accurately push commodity information according to the real-time requirement of a user.
According to an aspect of the present application, there is provided a commodity information pushing method, including:
constructing a data candidate set according to the acquired commodity demand text information;
training the initialized commodity information prediction model by using the constructed data candidate set to obtain a trained commodity information prediction model;
inputting commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information;
and displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information respectively.
According to another aspect of the present application, there is provided a commodity information pushing apparatus including:
the building module is used for building a data candidate set according to the acquired commodity demand text information;
the training module is used for training the initialized commodity information prediction model by using the constructed data candidate set to obtain a trained commodity information prediction model;
the matching module is used for inputting the commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information;
and the display module is used for respectively displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information.
According to still another aspect of the present application, there is provided a storage medium having stored thereon a computer program which, when executed by a processor, implements the above-described commodity information push method.
According to still another aspect of the present application, there is provided a computer device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the merchandise information pushing method when executing the program.
Compared with the existing technical scheme of commodity information pushing, the commodity information pushing method and device, the storage medium and the computer device, the data candidate set is built according to the obtained commodity demand text information, the initialized commodity information prediction model is trained by the built data candidate set, the trained commodity information prediction model is obtained, commodity keyword information in the obtained commodity search request is conveniently input into the trained commodity information prediction model, a plurality of commodity introduction text information matched with the commodity keyword information is obtained, a plurality of commodity information corresponding to the commodity introduction text information are respectively displayed, and therefore accurate commodity pushing is achieved for real-time demands of users. Therefore, the trained commodity information prediction model can recommend commodity information with high matching degree with the commodity demand text information for the user according to the commodity demand text information input by the user, the cold start problem in the prior art does not exist, the commodity information does not need to be manually input in the network model building process, and the whole process of building the commodity information prediction model can be automatically completed.
The foregoing description is only an overview of the technical solutions of the present application, and the present application can be implemented according to the content of the description in order to make the technical means of the present application more clearly understood, and the following detailed description of the present application is given in order to make the above and other objects, features, and advantages of the present application more clearly understandable.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a schematic flowchart illustrating a commodity information pushing method according to an embodiment of the present application;
fig. 2 is a schematic flow chart illustrating another commodity information pushing method provided in the embodiment of the present application;
fig. 3 shows a schematic structural diagram of a commodity information pushing device according to an embodiment of the present application.
Detailed Description
The present application will be described in detail below with reference to the accompanying drawings in conjunction with embodiments. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
The method aims at solving the technical problems that the prior art recommends corresponding commodities for users based on historical purchasing behavior information of the users and has cold start, recommends corresponding commodities for the users based on the quantifiable attribute information of the commodities, and needs to manually input commodity information to realize commodity pushing for the users in the process of constructing a network model, so that the method is not suitable for massive commodities, and cannot realize an accurate pushing function according to the real-time requirements of the users. The embodiment provides a commodity information pushing method, which can effectively avoid the problems of cold start and insufficient automation and intelligence of network model construction in the prior art, and further improves the commodity information recommendation accuracy, as shown in fig. 1, the method includes:
101. and constructing a data candidate set according to the acquired commodity demand text information.
In this embodiment, historical purchase record information of a user is collected, commodity demand text information for different commodities is acquired as a data sample, and the acquired commodity demand text information belonging to the same commodity is labeled, so that a data candidate set for training a commodity information prediction model is obtained.
102. And training the initialized commodity information prediction model by using the constructed data candidate set to obtain the trained commodity information prediction model.
In this embodiment, an initialized commodity pushing prediction model is constructed based on a neural network model (for example, a convolutional neural network model, a cyclic neural network model, or a probabilistic neural network model), and a framework for initializing the commodity pushing prediction model specifically includes:
a first layer: the input layer is used for carrying out format processing on input information, and the input information is a data candidate set;
a second layer: the encoding layer is used for carrying out one-hot encoding on the input information subjected to format processing by taking characters as units to obtain corresponding semantic vectors;
third to sixth layers: the mapping layer is used for obtaining a plurality of corresponding output results after the semantic vectors are processed by a plurality of full-connection layers containing an activation function tanh;
a seventh layer: the matching layer is used for carrying out similarity calculation on the obtained multiple output results to obtain multiple similarity calculation results;
an eighth layer: and the Softmax layer is used for carrying out normalization processing on a plurality of similarity calculation results obtained by calculation by utilizing an activation function Softmax to obtain the similarity probability.
Training the initialized commodity pushing prediction model based on the constructed data candidate set to obtain a trained commodity information prediction model, so that commodity information prediction is realized according to a commodity search request from a user by using the trained commodity information prediction model, and commodity information meeting the commodity search requirement of the user is obtained.
103. Inputting the commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information.
In this embodiment, a plurality of pieces of commodity introduction text information matched with the commodity keyword information are obtained according to the commodity keyword information in the obtained commodity search request, the commodity introduction text information used for similarity calculation is obtained by using the first layer to the sixth layer in the trained commodity information prediction model, and the plurality of pieces of commodity introduction text information matched with the commodity keyword information are obtained by performing calculation through the seventh layer and the eighth layer according to the mapping result output by the sixth layer.
104. And displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information respectively.
In this embodiment, a plurality of pieces of commodity information corresponding to a plurality of pieces of commodity introduction text information are acquired, and the acquired plurality of pieces of commodity information are arranged in a descending order according to the similarity and displayed, so that a user can browse the plurality of pieces of commodity information.
According to the scheme, a data candidate set is constructed according to the acquired commodity requirement text information, the constructed data candidate set is used for training the initialized commodity information prediction model to obtain a trained commodity information prediction model, so that the commodity keyword information in the acquired commodity search request is input into the trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information, and a plurality of commodity information corresponding to the commodity introduction text information are displayed respectively, and therefore accurate commodity pushing is achieved for the real-time requirement of the user. Compared with the prior art, the commodity information pushing method and the commodity information pushing device have the advantages that the trained commodity information prediction model can recommend commodity information with high matching degree with the commodity demand text information for the user according to the commodity demand text information input by the user, the cold start problem in the prior art does not exist, the commodity information does not need to be manually input in the network model building process, and the whole process of building the commodity information prediction model can be automatically completed.
Further, as a refinement and an extension of the specific implementation of the above embodiment, in order to fully describe the specific implementation process of the embodiment, another commodity information pushing method is provided, as shown in fig. 2, and the method includes:
201. and constructing a data candidate set according to the acquired commodity demand text information.
To illustrate the specific implementation manner of step 201, as a preferred embodiment, the product requirement text information includes product keyword information and a plurality of product introduction text information corresponding to the product keyword information; the plurality of commodity introduction text messages corresponding to the commodity keyword information in the commodity requirement text messages for constructing the data candidate set comprise first commodity introduction text messages of commodities purchased by the user and second commodity introduction text messages of commodities not purchased by the user.
In this embodiment, the commodity requirement text information includes: the article keyword information (e.g., the input article name, the article name keyword), and a plurality of article introduction text information having the same article name or the article name keyword corresponding to each article name keyword. The method comprises the steps of marking commodity introduction text information with the same commodity name or commodity name keywords, namely marking first commodity introduction text information of commodities purchased by a user as a positive sample, marking second commodity introduction text information of commodities not purchased by the user as a negative sample, and using the second commodity introduction text information as a data candidate set.
202. And respectively carrying out distance similarity calculation on the commodity keyword information, the first commodity introduction text information and the plurality of second commodity introduction text information to obtain a plurality of similarity values.
In this embodiment, the input information is preprocessed by using the first layer (input layer) of the initialized goods information prediction model, the input information is a data candidate set, the number of characters of the goods name keyword in the goods demand text information in the data candidate set is not more than 30 characters, and the number of characters of the goods introduction text information having the same goods name or the goods name keyword is not more than 500 characters. If the number of the characters is less than the required number of the characters, adding 0 to the tail ends of the commodity name key words and the commodity introduction texts with the same commodity names or the same commodity name key words; if the number of characters is more than the required number of characters, the commodity name keyword and an excess part of the commodity introduction text having the same commodity name or the commodity name keyword are discarded directly.
The sample marking of the commodity introduction text information in the commodity demand text information is specifically that the first commodity introduction text information corresponding to a commodity purchased by a user is a positive sample, and corresponds to commodity keyword information and commodity information purchased by the user in a default mode; the second commodity introduction text information corresponding to the commodities not purchased by the user is negative samples, and corresponds to the commodity keyword information and the commodity information by default.
Correspondingly, the second layer (coding layer) of the initialized commodity information prediction model is used for carrying out one-hot coding on the preprocessed commodity demand text information by taking characters as units, namely, coding the commodity name or name key words, the first commodity introduction texts belonging to the positive samples and the second commodity introduction texts belonging to the negative samples to respectively obtain corresponding semantic vectors.
One-hot encoding refers to encoding N states using an N-bit state register, where each state has an independent register bit and only one state is guaranteed to be valid at any time. For example, the word segmentation is performed on the "special product of Qingdao" to obtain the feature words "Qingdao" and "special product", that is, the index of "Qingdao" is 1, the index of "special product" is 2, and the semantic vector of each feature word is obtained as: qingdao (10), specialty (01).
In an actual application scenario, the commodity name or name keyword, the first commodity introduction text and the plurality of second commodity introduction texts in each group of samples are subjected to one-hot coding to obtain corresponding semantic vectors, namely the semantic vector of the commodity name or name keyword is marked as Q, and the semantic vector of the first commodity introduction text belonging to the positive sample is marked as D1And a plurality of semantic vectors of second commodity introduction texts belonging to the negative samples are marked as D2-Dn。
Accordingly, by initializing the third layer to the sixth layer (mapping layer) of the product information prediction model, the obtained semantic vector is passed through a plurality of fully connected layers containing the activation function tanh, and an output result for representing the product keyword information and an output result for representing the product introduction text information are obtained.
Wherein the input data of the third layer is the output data (Q, D) of the second layer1，D2，……，Dn) The random initialization parameter matrix is W, b, and the calculation is carried out according to the input data, the random initialization parameter matrix and the activation function to obtain an output result, wherein the specific calculation formula is as follows:
the random initialization parameter matrix of the fourth layer is W1，b1The specific calculation formula is as follows:
by analogy, the fifth layer and the sixth layer are the same as the third layer and the fourth layer.
In this embodiment, by initializing the seventh layer (matching layer) of the product information prediction model, the obtained output results for representing the product keyword information are respectively subjected to distance similarity calculation with the output results for representing the product introduction text information, so as to obtain a plurality of similarity values. The specific calculation formula is as follows:
wherein R (Q, D) represents output result y for representing commodity keyword informationQAnd an output result y for representing the commodity introduction text informationDThe similarity value between, cosine (y)Q，yD) Cosine distance, i.e. semantic similarity, representing two semantic vectors.
203. And calculating to obtain the similarity probability of the commodity keyword information and the first commodity introduction text information according to the similarity values.
In this embodiment, the activation function Softmax in the eighth layer (Softmax layer) of the initialized product information prediction model is used to normalize the calculated similarity values, and the product keyword information Q and the first product introduction text information D are obtained1The likelihood probability of (c). The specific calculation formula is as follows:
wherein, P (D)1| Q) represents the commodity keyword information Q and the first commodity introduction text information D1Gamma is a weight parameter, i.e. a smoothing factor of Softmax.
204. And training the initialized commodity information prediction model according to the similarity probability to obtain the trained commodity information prediction model.
For illustrating the specific implementation of step 204, as a preferred embodiment, step 204 may specifically include: determining network parameters of a commodity information prediction model by carrying out maximum likelihood estimation on the similarity probability; and obtaining the trained commodity information prediction model according to the determined network parameters of the commodity information prediction model.
In this embodiment, the first item introduction text information D is associated with the item keyword information Q1Determines a loss function which is a log-likelihood loss function, i.e. based on all the commodity keyword information Q and the first commodity introduction text information D1Training the initialized commodity pushing prediction model by the maximum likelihood estimation of the similarity probability to obtain the trained commodity pushing prediction model so as to enable the commodity keyword information Q to be matched with the first commodity introduction text information D1The probability of (2) is maximized, the loss function is minimized, and the specific calculation formula of the loss function is as follows:
205. when a commodity search request from a user is monitored, commodity keyword information in the commodity search request is obtained.
206. And acquiring a plurality of commodity information matched with the commodity keyword information by using the trained commodity information prediction model.
207. And acquiring a plurality of commodity introduction text messages corresponding to the plurality of commodity information according to the plurality of matched commodity information.
In the embodiment, when a product search request from a user is received, product keyword information in the product search request and a plurality of product introduction text information matched with the product keywords are obtained. Respectively inputting the obtained commodity keyword information and the plurality of commodity introduction text information into a commodity keyword module (namely, a sub-network model corresponding to the commodity keyword Q) and a positive sample module (namely, first commodity introduction text information D) of the trained commodity push prediction model1Corresponding sub-network models) to obtain semantic vectors Q of commodity keywords and semantic vectors D of a plurality of commodity introduction text messages.
In an actual application scene, character processing is carried out on the obtained commodity key words and the plurality of commodity introduction text messages to obtain commodity key words and the plurality of commodity introduction text messages with unified character quantity, the unified commodity key words are coded by utilizing a commodity key word module in a commodity pushing prediction model, the unified plurality of commodity introduction text messages are coded by utilizing a positive sample module in the commodity pushing prediction model, and semantic vectors Q for representing the commodity key words and semantic vectors D of the plurality of commodity introduction text messages are obtained.
208. And respectively carrying out similarity calculation on the commodity keyword information and the obtained plurality of commodity introduction text information to obtain a plurality of commodity introduction text information matched with the commodity keyword information.
For the purpose of illustrating the specific implementation of step 208, as a preferred embodiment, step 208 may specifically include: respectively carrying out distance similarity calculation on the semantic vectors corresponding to the commodity keyword information and the semantic vectors corresponding to the obtained plurality of commodity introduction text information to obtain a plurality of similarity values; performing descending order arrangement on the plurality of similarity values obtained by calculation to obtain a descending order arrangement result; and according to a preset commodity matching value, obtaining a plurality of commodity introduction text messages matched with the commodity keyword messages according to the descending order arrangement result.
The preset commodity matching value may be represented by all commodities with a similarity exceeding 90%, or may be represented by commodities in a preset rank (for example, TOP100) in the descending order arrangement result, where the setting criterion and the setting dimension of the preset commodity matching value are not specifically limited.
209. And displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information respectively.
In this embodiment, according to a plurality of pieces of commodity introduction text information output by a trained commodity information prediction model, a plurality of pieces of commodity information respectively corresponding to the plurality of pieces of commodity introduction text information are acquired, and a commodity list is generated, that is, a plurality of commodities recommended to a user are determined.
By applying the technical scheme of the embodiment, a data candidate set is constructed according to the acquired commodity demand text information, the constructed data candidate set is used for training the initialized commodity information prediction model to obtain a trained commodity information prediction model, so that the commodity keyword information in the acquired commodity search request is input into the trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information, and a plurality of commodity information corresponding to the commodity introduction text information are displayed respectively, and therefore accurate commodity pushing is achieved for the real-time demand of a user. Compared with the prior art, the commodity information pushing method and the commodity information pushing device have the advantages that the trained commodity information prediction model can recommend commodity information with high matching degree with the commodity demand text information for the user according to the commodity demand text information input by the user, the cold start problem in the prior art does not exist, the commodity information does not need to be manually input in the network model building process, and the whole process of building the commodity information prediction model can be automatically completed.
Further, as a specific implementation of the method in fig. 1, an embodiment of the present application provides a commodity information pushing apparatus, and as shown in fig. 3, the apparatus includes: a construction module 31, a training module 32, a matching module 34, and a display module 35.
The building module 31 may be configured to build a data candidate set according to the obtained commodity demand text information;
the training module 32 may be configured to train the initialized commodity information prediction model by using the constructed data candidate set, so as to obtain a trained commodity information prediction model;
the matching module 34 may be configured to input the commodity keyword information in the obtained commodity search request into a trained commodity information prediction model, so as to obtain a plurality of commodity introduction text information matched with the commodity keyword information;
the display module 35 may be configured to display a plurality of product information corresponding to the plurality of product introduction text information, respectively.
In a specific application scenario, a monitoring module 33 is further included.
In a specific application scenario, the commodity requirement text information comprises commodity keyword information and a plurality of commodity introduction text information corresponding to the commodity keyword information; the plurality of commodity introduction text messages corresponding to the commodity keyword information in the commodity requirement text messages for constructing the data candidate set comprise first commodity introduction text messages of commodities purchased by the user and second commodity introduction text messages of commodities not purchased by the user.
In a specific application scenario, the first article introduction text information in the data candidate set is a positive sample, the second article introduction text information is a negative sample, and the training module 32 specifically includes: a first calculating unit 321, a second calculating unit 322, and a minimization training unit 323.
The first calculating unit 321 may be configured to perform distance similarity calculation on the product keyword information and the first product introduction text information and the plurality of second product introduction text information respectively to obtain a plurality of similarity values.
The second calculating unit 322 may be configured to calculate, according to the plurality of similarity values, a similarity probability between the product keyword information and the first product introduction text information.
The minimization training unit 323 may be configured to train the initialized commodity information prediction model according to the similarity probability, so as to obtain a trained commodity information prediction model.
In a specific application scenario, the minimization training unit 323 may be specifically configured to determine a network parameter of the commodity information prediction model by performing maximum likelihood estimation on the similarity probability; and obtaining the trained commodity information prediction model according to the determined network parameters of the commodity information prediction model.
In a specific application scenario, the monitoring module 33 may be configured to, when a product search request from a user is monitored, obtain product keyword information in the product search request.
In a specific application scenario, the matching module 34 specifically includes: a first acquisition unit 341, a second acquisition unit 342, a similarity calculation unit 343.
A first obtaining unit 341, configured to obtain, by using a trained product information prediction model, a plurality of pieces of product information that match the product keyword information;
a second obtaining unit 342, configured to obtain, according to the plurality of matching commodity information, a plurality of commodity introduction text information corresponding to the plurality of commodity information;
the similarity calculation unit 343 may be configured to perform similarity calculation on the product keyword information and the obtained plurality of product introduction text information, respectively, to obtain a plurality of product introduction text information that matches the product keyword information.
In a specific application scenario, the similarity calculation unit 343 may be specifically configured to perform distance similarity calculation on the semantic vectors corresponding to the commodity keyword information and the semantic vectors corresponding to the obtained plurality of commodity introduction text information, respectively, to obtain a plurality of similarity values; performing descending order arrangement on the plurality of similarity values obtained by calculation to obtain a descending order arrangement result; and according to a preset commodity matching value, obtaining a plurality of commodity introduction text messages matched with the commodity keyword messages according to the descending order arrangement result.
It should be noted that other corresponding descriptions of the functional units related to the product information pushing device provided in the embodiment of the present application may refer to the corresponding descriptions in fig. 1 and fig. 2, and are not described again here.
Based on the above methods shown in fig. 1 and fig. 2, correspondingly, the present application further provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the storage medium implements the above method for pushing the commodity information shown in fig. 1 and fig. 2.
Based on such understanding, the technical solution of the present application may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.), and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the implementation scenarios of the present application.
Based on the foregoing methods shown in fig. 1 and fig. 2 and the virtual device embodiment shown in fig. 3, to achieve the foregoing object, an embodiment of the present application further provides a computer device, which may specifically be a personal computer, a server, a network device, and the like, where the entity device includes a storage medium and a processor; a storage medium for storing a computer program; a processor for executing a computer program to implement the merchandise information pushing method shown in fig. 1 and 2.
Optionally, the computer device may further include a user interface, a network interface, a camera, Radio Frequency (RF) circuitry, sensors, audio circuitry, a WI-FI module, and so forth. The user interface may include a Display screen (Display), an input unit such as a keypad (Keyboard), etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (e.g., a bluetooth interface, WI-FI interface), etc.
It will be understood by those skilled in the art that the present embodiment provides a computer device structure that is not limited to the physical device, and may include more or less components, or some components in combination, or a different arrangement of components.
The storage medium may further include an operating system and a network communication module. An operating system is a program that manages the hardware and software resources of a computer device, supporting the operation of information handling programs, as well as other software and/or programs. The network communication module is used for realizing communication among components in the storage medium and other hardware and software in the entity device.
Through the above description of the embodiments, those skilled in the art will clearly understand that the present application can be implemented by software plus a necessary general hardware platform, and can also be implemented by hardware. Through the technical scheme of the application, compared with the existing technical scheme of commodity information pushing, the commodity information with high matching degree with the commodity demand text information can be recommended to the user through the trained commodity information prediction model according to the commodity demand text information input by the user, the cold start problem in the prior art does not exist, the commodity information does not need to be manually input in the process of building the network model, and the whole process of building the commodity information prediction model can be automatically completed.
Those skilled in the art will appreciate that the figures are merely schematic representations of one preferred implementation scenario and that the blocks or flow diagrams in the figures are not necessarily required to practice the present application. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario may be distributed in the devices in the implementation scenario according to the description of the implementation scenario, or may be located in one or more devices different from the present implementation scenario with corresponding changes. The modules of the implementation scenario may be combined into one module, or may be further split into a plurality of sub-modules.
The above application serial numbers are for description purposes only and do not represent the superiority or inferiority of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of the present application, but the present application is not limited thereto, and any variations that can be made by those skilled in the art are intended to fall within the scope of the present application.
Claims (10)
1. A commodity information pushing method is characterized by comprising the following steps:
constructing a data candidate set according to the acquired commodity demand text information;
training the initialized commodity information prediction model by using the constructed data candidate set to obtain a trained commodity information prediction model;
inputting commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information;
and displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information respectively.
2. The method according to claim 1, wherein the commodity requirement text information includes commodity keyword information and a plurality of commodity introduction text information corresponding to the commodity keyword information;
the plurality of commodity introduction text messages corresponding to the commodity keyword information in the commodity requirement text messages for constructing the data candidate set comprise first commodity introduction text messages of commodities purchased by the user and second commodity introduction text messages of commodities not purchased by the user.
3. The method according to claim 2, wherein the first commodity introduction text information in the data candidate set is a positive sample, the second commodity introduction text information is a negative sample, and the initialized commodity information prediction model is trained by using the constructed data candidate set to obtain a trained commodity information prediction model, specifically comprising:
respectively carrying out distance similarity calculation on the commodity keyword information and the first commodity introduction text information and the plurality of second commodity introduction text information to obtain a plurality of similarity values;
calculating to obtain the similarity probability of the commodity keyword information and the first commodity introduction text information according to the similarity values;
and training the initialized commodity information prediction model according to the similarity probability to obtain the trained commodity information prediction model.
4. The method according to claim 3, wherein the training of the initialized commodity information prediction model according to the similarity probability to obtain a trained commodity information prediction model specifically comprises:
determining network parameters of a commodity information prediction model by carrying out maximum likelihood estimation on the similarity probability;
and obtaining the trained commodity information prediction model according to the determined network parameters of the commodity information prediction model.
5. The method according to claim 1, wherein before obtaining a plurality of article introduction text messages matched with the article keyword information according to the article keyword information in the obtained article search request by using the trained article information prediction model, the method specifically further comprises:
when a commodity search request from a user is monitored, commodity keyword information in the commodity search request is obtained.
6. The method according to claim 1 or 2, wherein the step of inputting the commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information specifically comprises:
acquiring a plurality of commodity information matched with the commodity keyword information by using the trained commodity information prediction model;
according to the plurality of matched commodity information, acquiring a plurality of commodity introduction text information corresponding to the plurality of commodity information;
and respectively carrying out similarity calculation on the commodity keyword information and the obtained plurality of commodity introduction text information to obtain a plurality of commodity introduction text information matched with the commodity keyword information.
7. The method according to claim 5, wherein the calculating of the similarity between the product keyword information and the obtained plurality of product introduction text information to obtain a plurality of product introduction text information matched with the product keyword information specifically comprises:
respectively carrying out distance similarity calculation on the semantic vectors corresponding to the commodity keyword information and the semantic vectors corresponding to the obtained plurality of commodity introduction text information to obtain a plurality of similarity values;
performing descending order arrangement on the plurality of similarity values obtained by calculation to obtain a descending order arrangement result;
and according to a preset commodity matching value, obtaining a plurality of commodity introduction text messages matched with the commodity keyword messages according to the descending order arrangement result.
8. A commodity information pushing apparatus, comprising:
the building module is used for building a data candidate set according to the acquired commodity demand text information;
the training module is used for training the initialized commodity information prediction model by using the constructed data candidate set to obtain a trained commodity information prediction model;
the matching module is used for inputting the commodity keyword information in the obtained commodity search request into a trained commodity information prediction model to obtain a plurality of commodity introduction text information matched with the commodity keyword information;
and the display module is used for respectively displaying a plurality of commodity information corresponding to the plurality of commodity introduction text information.
9. A storage medium on which a computer program is stored, wherein the program, when executed by a processor, implements the merchandise information pushing method according to any one of claims 1 to 7.
10. A computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the merchandise information pushing method according to any one of claims 1 to 7 when executing the program.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910780785.1A CN110706055A (en) | 2019-08-22 | 2019-08-22 | Commodity information pushing method and device, storage medium and computer equipment |
| PCT/CN2019/118488 WO2021031410A1 (en) | 2019-08-22 | 2019-11-14 | Commodity information prediction model-based commodity information pushing method and apparatus, non-volatile readable storage medium, and computer device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910780785.1A CN110706055A (en) | 2019-08-22 | 2019-08-22 | Commodity information pushing method and device, storage medium and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110706055A true CN110706055A (en) | 2020-01-17 |
Family
ID=69193799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910780785.1A Pending CN110706055A (en) | 2019-08-22 | 2019-08-22 | Commodity information pushing method and device, storage medium and computer equipment |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110706055A (en) |
| WO (1) | WO2021031410A1 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111611807A (en) * | 2020-05-18 | 2020-09-01 | 北京邮电大学 | Keyword extraction method and device based on neural network and electronic equipment |
| CN111681086A (en) * | 2020-06-16 | 2020-09-18 | 上海风秩科技有限公司 | Commodity recommendation method and device, computer equipment and readable storage medium |
| CN111737418A (en) * | 2020-07-20 | 2020-10-02 | 北京每日优鲜电子商务有限公司 | Method, apparatus and storage medium for predicting relevance of search term and commodity |
| CN113450172A (en) * | 2020-03-27 | 2021-09-28 | 北京沃东天骏信息技术有限公司 | Commodity recommendation method and device |
| CN115250365A (en) * | 2021-04-28 | 2022-10-28 | 京东科技控股股份有限公司 | Commodity text generation method and device, computer equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109344314A (en) * | 2018-08-20 | 2019-02-15 | 腾讯科技（深圳）有限公司 | A kind of data processing method, device and server |
| CN109711929A (en) * | 2018-12-13 | 2019-05-03 | 中国平安财产保险股份有限公司 | Business recommended method and device based on prediction model |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| US20160125501A1 (en) * | 2014-11-04 | 2016-05-05 | Philippe Nemery | Preference-elicitation framework for real-time personalized recommendation |
| CN106295832B (en) * | 2015-05-12 | 2020-05-19 | 阿里巴巴集团控股有限公司 | Product information pushing method and device |
| CN108960945A (en) * | 2017-05-18 | 2018-12-07 | 北京京东尚科信息技术有限公司 | Method of Commodity Recommendation and device |
| CN108711110B (en) * | 2018-08-14 | 2023-06-23 | 中国平安人寿保险股份有限公司 | Insurance product recommendation method, apparatus, computer device and storage medium |
-
2019
- 2019-08-22 CN CN201910780785.1A patent/CN110706055A/en active Pending
- 2019-11-14 WO PCT/CN2019/118488 patent/WO2021031410A1/en active Application Filing
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109344314A (en) * | 2018-08-20 | 2019-02-15 | 腾讯科技（深圳）有限公司 | A kind of data processing method, device and server |
| CN109711929A (en) * | 2018-12-13 | 2019-05-03 | 中国平安财产保险股份有限公司 | Business recommended method and device based on prediction model |
Non-Patent Citations (1)
| Title |
|---|
| PO-SEN HUANG等: "Learning deep structured semantic models for web search using clickthrough data", ACM, pages 2333 - 2338 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113450172A (en) * | 2020-03-27 | 2021-09-28 | 北京沃东天骏信息技术有限公司 | Commodity recommendation method and device |
| CN111611807A (en) * | 2020-05-18 | 2020-09-01 | 北京邮电大学 | Keyword extraction method and device based on neural network and electronic equipment |
| CN111681086A (en) * | 2020-06-16 | 2020-09-18 | 上海风秩科技有限公司 | Commodity recommendation method and device, computer equipment and readable storage medium |
| CN111737418A (en) * | 2020-07-20 | 2020-10-02 | 北京每日优鲜电子商务有限公司 | Method, apparatus and storage medium for predicting relevance of search term and commodity |
| CN111737418B (en) * | 2020-07-20 | 2021-05-14 | 北京每日优鲜电子商务有限公司 | Method, apparatus and storage medium for predicting relevance of search term and commodity |
| CN115250365A (en) * | 2021-04-28 | 2022-10-28 | 京东科技控股股份有限公司 | Commodity text generation method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021031410A1 (en) | 2021-02-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110706055A (en) | Commodity information pushing method and device, storage medium and computer equipment | |
| CN106777318B (en) | Matrix decomposition cross-modal Hash retrieval method based on collaborative training | |
| CN109697629B (en) | Product data pushing method and device, storage medium and computer equipment | |
| CN111061946A (en) | Scenario content recommendation method and device, electronic equipment and storage medium | |
| CN109285023A (en) | For dissecting the method and device thereof of user's intention | |
| CN110008397B (en) | Recommendation model training method and device | |
| CN107562787B (en) | POI (point of interest) encoding method and device, POI recommendation method and electronic equipment | |
| CN111291765A (en) | Method and device for determining similar pictures | |
| CN110428295A (en) | Method of Commodity Recommendation and system | |
| CN112364204A (en) | Video searching method and device, computer equipment and storage medium | |
| CN113220734A (en) | Course recommendation method and device, computer equipment and storage medium | |
| CN110858226A (en) | Conversation management method and device | |
| CN113051480A (en) | Resource pushing method and device, electronic equipment and storage medium | |
| CN110083774A (en) | Using determination method, apparatus, computer equipment and the storage medium of recommendation list | |
| CN111209351B (en) | Object relation prediction method, object recommendation method, object relation prediction device, object recommendation device, electronic equipment and medium | |
| CN113656699B (en) | User feature vector determining method, related equipment and medium | |
| CN113657087B (en) | Information matching method and device | |
| CN108563648B (en) | Data display method and device, storage medium and electronic device | |
| CN116578925B (en) | Behavior prediction method, device and storage medium based on feature images | |
| CN111026973B (en) | Commodity interest degree prediction method and device and electronic equipment | |
| CN112100491A (en) | Information recommendation method, device and equipment based on user data and storage medium | |
| CN115269998A (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN111507471B (en) | Model training method, device, equipment and storage medium | |
| CN112632275B (en) | Crowd clustering data processing method, device and equipment based on personal text information | |
| JP2019106099A (en) | Providing program, providing device, and providing method |
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 |