CN109213923B - Method and system for determining associated information of user and object - Google Patents
Method and system for determining associated information of user and object Download PDFInfo
- Publication number
- CN109213923B CN109213923B CN201710547349.0A CN201710547349A CN109213923B CN 109213923 B CN109213923 B CN 109213923B CN 201710547349 A CN201710547349 A CN 201710547349A CN 109213923 B CN109213923 B CN 109213923B
- Authority
- CN
- China
- Prior art keywords
- user
- information
- historical
- real
- target object
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application relates to the technical field of internet, in particular to a method and a system for determining associated information of a user and an object, which are used for solving the problem that the characteristic coverage rate of a mode for determining the associated information of the user and the object is low in the prior art. The method and the device for determining the relevance information of the user and the object comprise the steps of selecting the characteristic information of the object determined by the user behavior of the user from a characteristic information set, selecting the historical relevance information of the user from a historical relevance information set, wherein the characteristic information set comprises vectorized characteristic information, the historical relevance information set comprises vectorized historical relevance information, and determining the relevance information of the user and the object according to the characteristic information of the object, the historical relevance information of the user and the real-time user relevance information. The embodiment of the application corrects the real-time user associated information by using the feature information and the historical associated information, and the feature coverage rate of the feature information and the historical associated information is much larger than that of real-time data, so that the feature coverage rate of the associated information mode of the user and the object is improved.
Description
Technical Field
The present application relates to the field of internet technologies, and in particular, to a method and a system for determining association information between a user and an object.
Background
The network publishing platform is a platform for a publisher to publish own products through a network, and a user to browse the products through the network and interact through the network.
When a publisher publishes a product on a network publishing platform, the publisher configures product information for the product. And the user can interact with the website publishing platform aiming at least one product according to the product information.
Currently, a network publishing platform may recommend an object to a user according to association information of the user and the object, for example, the association information may be user preferences. User preferences are rational, inclined choices made by the user in selecting goods and services.
The real-time network behavior of the user comprises a plurality of real-time user preferences, such as style preferences, keyword preferences and the like.
At present, most systems extract some basic features of a user, such as price preference, keyword preference and the like, determine a preference value of the user for a certain object according to the basic features of a special area, and select the object to recommend the user according to the preference value.
The coverage rate of the features of the method for extracting the basic features of the user is low, for example, when the user just browses several large-code women's dresses, the large code can be extracted as the preference of the user. However, women's dresses such as "fat" and "wide" are in accordance with the user's preference, but the women's dresses containing the keywords such as "fat" and "wide" cannot be recommended to the user because the keywords such as "big code" cannot be matched.
In summary, currently, the mode feature coverage rate for determining the associated information between the user and the object is relatively low.
Disclosure of Invention
The application provides a method and a system for determining associated information of a user and an object, which are used for solving the problem that the mode characteristic coverage rate for determining the associated information of the user and the object is low in the prior art.
The method for determining the associated information of the user and the object provided by the embodiment of the application comprises the following steps:
for at least one user, selecting feature information of an object determined by user behaviors of the user from a feature information set, and selecting historical associated information of the user from a historical associated information set, wherein the feature information set comprises vectorized feature information, and the historical associated information set comprises vectorized historical associated information;
determining real-time correction information of the vectorized user according to the characteristic information of the object, the historical associated information of the user and the vectorized real-time user associated information;
and aiming at least one target object, determining the association information of the user and the target object according to the real-time user correction information.
The system for determining the associated information between the user and the object provided by the embodiment of the application comprises the following steps:
an information determination module, configured to select, for at least one user, feature information of an object determined by user behavior of the user from a feature information set, and select historical associated information of the user from a historical associated information set, where the feature information set includes vectorized feature information, and the historical associated information set includes vectorized historical associated information;
the correction module is used for determining real-time correction information of the vectorized user according to the characteristic information of the object, the historical associated information of the user and the vectorized real-time user associated information;
and the processing module is used for determining the associated information of the user and the target object according to the real-time user correction information aiming at least one target object.
The method comprises the steps of selecting characteristic information of an object determined by user behaviors of a user from a characteristic information set, selecting historical associated information of the user from a historical associated information set, wherein the characteristic information set comprises vectorized characteristic information, the historical associated information set comprises vectorized historical associated information, and determining the associated information of the user and the object according to the characteristic information of the object, the historical associated information of the user and the real-time user associated information. The characteristic information set and the historical associated information set are obtained after the neural network model is trained, so that the dimension between the characteristic information of the object and the historical associated information of the user is the same, the characteristic information and the historical associated information can be used for correcting the real-time user associated information, and the characteristic coverage rate of the characteristic information set and the historical associated information set is much larger than that of real-time data, so that the characteristic coverage rate of a mode for determining the associated information of the user and the object is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
FIG. 1 is a schematic diagram illustrating a scheme for determining association information between a user and an object according to an embodiment of the present application;
FIG. 2 is a flowchart illustrating a method for determining association information between a user and an object according to an embodiment of the present disclosure;
FIG. 3 is a flowchart illustrating a method for determining a feature information set and a historical associated information set according to an embodiment of the present application;
FIG. 4 is a flowchart illustrating a complete method for determining association information between a user and an object according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a system for determining association information between a user and an object according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the accompanying drawings, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
According to the embodiment of the application, the specific objects are different according to different application scenes. Such as if the embodiments of the present application are applied to an e-market scene, the object may be any object capable of conducting a transaction, such as a good, a service, etc.
The characteristic information of the object of the embodiment of the present application is information related to the object, such as color, material, style, and the like.
The correlation information of the user and the object in the embodiment of the application indicates the correlation between the user and the object, and parameters indicating the correlation are many, for example, the user preference information may indicate the preference of the user for one object, and the higher the preference is, the more the user is interested in the object; for example, the user tendency information may indicate the tendency degree of the user for an object, and the higher the tendency degree, the more interesting the user is for the object.
As shown in fig. 1, a scheme for determining association information between a user and an object according to an embodiment of the present application is divided into four parts:
a first part: performing off-line calculation to obtain a characteristic information set and a historical associated information set;
a second part: correction processing, which aims at correcting the real-time user associated information;
and a third part: associating, in order to determine association information of the user and the object;
the fourth part: and the purpose of the application is to perform specific application according to the determined association information of the user and the object.
Each of the following sections is described separately below.
A first part: and (4) performing off-line calculation.
This is partly by determining the feature information set and the historical associated information set from the offline data.
The historical associated information set is obtained according to historical text information; the characteristic information set is obtained according to the characteristic text information.
How the next two text messages are obtained is described below.
Firstly, historical text information:
the embodiment of the application acquires offline data from a website, wherein the offline data comprises a large number of user behaviors of users, such as behaviors of browsing objects, collecting objects, trading objects and the like.
After the offline data is acquired, an object related to the user behavior may be determined according to the user behavior, then characteristics of the object may be determined from a website, for example, key information of the object may be determined according to information such as a title of the object on the website, and the key information may be used as historical user text information.
Taking the application of the embodiment of the application to e-commerce scenes as an example, determining that the user A browses the one-piece dress from offline data, and then finding the one-piece dress from the e-commerce platform with the title of 'new style of autumn for explosion', the key information of the corresponding objects being 'autumn', 'new style' and 'one-piece dress', and taking 'autumn', 'new style' and 'one-piece dress' as historical user text information.
And if other user behaviors of the user A exist in the offline data, continuously determining the key information of the object according to the other user behaviors, and taking the union of all the key information as historical text information of the user.
Secondly, characteristic text information:
each object has feature information, and the embodiment of the application can determine corresponding feature information for each object.
For at least one object, other objects of the same type as the object can be determined, and then information such as titles on the website corresponding to the object and other objects is used for determining key information of the object, and the key information is used as characteristic text information.
When other objects of the same type as the object are determined, whether the other objects of the same type as the object are the same can be judged according to the information of the other objects, and the other objects of the same type can also be identified through an image processing technology.
Taking the application of the embodiment of the application to e-commerce scenes as an example, assuming that the object is a one-piece dress, other commodities of the same type as the one-piece dress on the e-commerce platform, such as the one-piece dress 1 and the one-piece dress 2, can be identified by using an image processing technology.
The title of the one-piece dress is 'new autumn' for burst style from the e-commerce platform, and the key information of the corresponding object is 'autumn', 'new style' and 'one-piece dress';
the title of the one-piece dress 1 is 'tipping beige one-piece dress' searched from the e-commerce platform, and key information of corresponding objects is 'beige' and 'one-piece dress';
the title of the one-piece dress 2 is searched from the e-commerce platform to be the sleeveless stripe one-piece dress, and the key information of the corresponding objects is sleeveless stripe one-piece dress, stripe one-piece dress and the sleeve-free stripe one-piece dress.
The union of these key information is used as the characteristic text information of the 'one-piece dress' object.
Training the neural network model according to the obtained historical text information and the characteristic text information, placing a user vector output by the trained neural network model in a historical associated information set as historical associated information, and placing a characteristic vector output by the trained neural network model in a characteristic information set as characteristic information.
A second part: and (6) correction processing.
The corresponding object and the real-time user association information are determined according to the user behavior of the user within the set time length. The corresponding object may be determined, for example, by user behavior in a real-time log.
Feature information of the object determined according to the user behavior is selected from the feature information set obtained from the first part.
For example, if the user Z browses the object a and purchases the object B within a set time period, the determined objects include the object a and the object B, and then the feature information of the object a and the feature information of the object B are selected from the feature information set.
And selecting the historical associated information of the user Z from the historical associated information set.
And correcting the real-time user associated information of the user Z according to the characteristic information of the object A, the characteristic information of the object B and the historical associated information of the user Z to obtain the real-time user corrected information.
And a third part: and (6) associating.
The target object can be obtained according to the object determined by the user behavior of the user.
Taking the application of the embodiment of the application to e-commerce scenes as an example, assuming that the object determined according to the user behavior of the user is a mobile phone, the target object is all mobile phones and commodities of the same type as the mobile phones, such as a tablet computer, a notebook computer and the like.
Determining the characteristic information of all target objects from the characteristic information set; and then respectively determining the associated information of the user and each target object according to the user real-time correction information obtained by the second part and the characteristic information of each target object.
For example, the target object a may determine the association information between the user and the target object a according to the real-time correction information of the user and the feature information of the target object a.
The fourth part: application is carried out.
After the associated information of the user and each target object is obtained, the associated information can be applied.
Different application scenes and specific application modes are different.
Taking the application of the embodiment of the application to e-commerce scenes as an example, it is assumed that the associated information of the user and each target object is the preference degree of the user for the target commodity.
The target commodities can be sorted according to the preference degrees of the user on all the target commodities;
and recommending the target commodity for the user according to the sequencing result.
As shown in fig. 2, a method for determining association information between a user and an object in an embodiment of the present application includes:
The method comprises the steps of selecting characteristic information of an object determined by user behaviors of a user from a characteristic information set, selecting historical associated information of the user from a historical associated information set, wherein the characteristic information set comprises vectorized characteristic information, the historical associated information set comprises vectorized historical associated information, and determining the associated information of the user and the object according to the characteristic information of the object, the historical associated information of the user and the real-time user associated information. The characteristic information set and the historical associated information set are obtained after the neural network model is trained, so that the dimension between the characteristic information of the object and the historical associated information of the user is the same, the characteristic information and the historical associated information can be used for correcting the real-time user associated information, and the characteristic coverage rate of the characteristic information set and the historical associated information set is much larger than that of real-time data, so that the characteristic coverage rate of a mode for determining the associated information of the user and the object is improved.
The feature information set of the embodiment of the application comprises a plurality of feature information;
the history associated information set of the embodiment of the application comprises a plurality of pieces of history associated information.
In implementation, the user and the history association information may be in a one-to-one correspondence; the object and feature information may be in a one-to-one correspondence.
The embodiment of the application can obtain the characteristic information set and the historical associated information set on line.
The specific mode is as follows:
regarding at least one object, taking the object and description information of the object of the same type as the object as object text information of the object; for at least one user, using description information of an object related to the user behavior in the offline data as historical user text information of the user;
training according to the object text information and the historical user text information to obtain vectorized feature information and vectorized historical associated information;
and placing the obtained characteristic information in the characteristic information set, and placing the obtained historical associated information in the historical associated information set.
According to the embodiment of the application, historical text information and characteristic text information are obtained through offline data, then the neural network model is trained through the two information, and the characteristic information and historical associated information which are finally output by the neural network model are obtained.
There are many algorithms for training the neural network model, as long as it can be ensured that vectors output after the neural network model is trained have the same dimension, addition and/or subtraction operations can be performed, and an algorithm in which the vectors obtained by addition and/or subtraction keep a certain physical meaning can be applied to the embodiments of the present application, that is, an algorithm of the nature of word-inference (word-inference) can be applied to the embodiments of the present application, such as Doc2vec algorithm.
For example, the "red" vector trained on the neural network model by Doc2vec is denoted as v1, and the "dress" vector is denoted as v2, then the resulting vector of v1+ v2 may represent a "red dress".
Taking the application of the embodiment of the application to e-commerce scenes as an example, when the characteristic text information is determined, after the offline data is acquired from the e-commerce website, the following processes are executed for each object:
aiming at the object A, searching the title of the object A from the E-commerce platform, and if the object A has a plurality of titles, searching each title; and
and searching other objects in the same type as the object A from the e-commerce platform, and searching titles of the other objects from the e-commerce platform, wherein if the other objects have a plurality of titles, each title needs to be searched.
And determining key information from all searched titles, and forming all determined key information into characteristic text information.
The key information is information capable of describing a corresponding object.
For example, the object is a one-piece dress, the title is "new autumn burst one-piece dress", and then "autumn", "new style" and "one-piece dress" are key information.
Taking the application of the embodiment of the application to e-commerce scenes as an example, when determining the historical text information, after acquiring the offline data from the e-commerce website, the following processes are executed for each user:
for the user A, the user behavior of the user A, such as browsing objects, collecting objects, trading objects and the like, is determined from the offline data.
All objects relevant to user a are determined from the user behavior.
The following procedure is performed separately for each object associated with user a:
aiming at the object A, searching the title of the object A from the E-commerce platform, and if the object A has a plurality of titles, searching each title; and
and searching other objects in the same type as the object A from the e-commerce platform, and searching titles of the other objects from the e-commerce platform, wherein if the other objects have a plurality of titles, each title needs to be searched.
And determining key information from all searched titles, and forming all determined key information into historical text information.
Inputting the obtained feature text information and historical text information into a neural network model, learning the neural network model through Doc2vec, and finally outputting a corresponding feature vector and a user vector of each user object aiming at each object.
And placing the feature vector as feature information in a feature information set, and placing the user vector as history associated information in a history associated information set.
Since the offline data also changes, a time can be set, and the feature information set and the history related information set can be updated periodically according to the latest offline data.
After the feature information set and the historical associated information set are obtained, the vectorized real-time user correction information can be determined according to the feature information of the object, the historical associated information of the user and the vectorized real-time user associated information.
Optionally, for at least one user, determining vectorized real-time user association information according to the user behavior vector of the user;
and correcting the historical associated information of the user according to the real-time user associated information and the associated correction information to obtain the vectorized real-time user correction information.
The real-time user associated information is determined according to the real-time behavior of the user, and then the historical associated information set is corrected according to the obtained characteristic information set and the real-time user associated information, so that the real-time user correction information is finally obtained.
Since the user behavior may be occurring at any time, a period of time may be set to determine the user behavior over a period of time.
Taking the application of the embodiment of the application to e-commerce scenes as an example:
suppose oiRepresenting the website commodity i, which is trained offline to obtain a feature vector of
uiThe historical preference information (namely historical associated information) obtained by offline training of the user i of the website is shown as
Suppose user uiThe real-time user associated information is Aui＝{(oi1,ti1),(oi2,ti2),.......,(oin,tin) In which (o)in,tin) Representing user uiAt tinAt that time, for the commodity oinWith user behavior (e.g., browsing, collecting, trading, etc.), the user can then modify preferences in real-time (i.e., the user can modify information in real-time)
Can be determined by the following formula:
t denotes the current time. The parameter α is used to adjust the weight of each real-time feature vector (e.g., behavior weight of browsing lines, trading, collecting, etc.), and β is used to adjust the weight of the whole feature vector, and the specific size can be determined by experience, experiment, etc. norm1|. | indicates that the vector is L1-regularized.
Optionally, for at least one target object, before determining the association information between the user and the target object according to the real-time user modification information, at least one target object is determined from all the objects according to the object determined by the user behavior of the user.
In an implementation, at least one target object may be determined from all objects according to the type of the object determined by the user behavior.
Taking the application of the embodiment of the application to e-commerce scenes as an example:
suppose uiRepresents a website user i with real-time correction information of
ojRepresenting a web commodity having a feature vector of
Optionally, for at least one target object, when determining the association information between the user and the target object according to the user real-time correction information:
and determining the associated information of the user and the target object according to the real-time correction information of the user and the characteristic information of the target object in the characteristic information set.
In implementation, the association information of the user and the target object may be determined by the following formula:
wherein, wijThe correlation information of the user i and the target object j is obtained;
correcting information for a user in real time; ojIs a target object j;
characteristic information of the target object j; (A, B) represents the inner product of vector A and vector B; | a | · | | B | | represents a vector 2-norm.
Optionally, after determining the association information between the user and each target object, the target objects may be sorted according to the association information between the user and the target objects;
recommending the target object for the user according to the ordered target object sequence.
There are, for example, four target objects, target object 1, target object 2, target object 3 and target object 4, respectively.
The associated information of the user a and the four target objects is respectively:
the association information of the user a with the target object 1 is 90;
the association information of the user a with the target object 2 is 93;
the association information of the user a with the target object 3 is 83;
the association information of the user a with the target object 4 is 20.
After sorting according to the associated information from big to small: target object 2, target object 1, target object 3 and target object 4.
The target object is recommended for the user a in this order, and the target object 2 is preferentially recommended for the user a.
As shown in fig. 3, a method for determining a feature information set and a history association information set in an embodiment of the present application includes:
and step 300, acquiring offline data from a website.
And step 306, taking the key information of the object and other objects of the same type as the object as characteristic text information.
Here, steps 301 to 303 are a method of generating the historical user text information, and steps 304 to 306 are a method of generating the feature text information. The two generation modes have no necessary time sequence relationship and can generate the two information simultaneously; or generating characteristic text information first and then generating historical user text information; or generating the historical user text information first and then generating the characteristic text information.
And 307, learning the neural network model through Doc2vec according to the historical user text information and the characteristic text information.
And 308, taking the user vector output by the neural network model after learning as historical associated information and placing the historical associated information in a historical associated information set, and taking the feature vector output by the neural network model after training as feature information and placing the feature vector in a feature information set.
As shown in fig. 4, a complete method for determining association information between a user and an object according to an embodiment of the present application includes:
Based on the same inventive concept, the embodiment of the present invention further provides a system for determining the association information between the user and the object, and because the principle of solving the problem of the system is similar to the method for determining the association information between the user and the object in the embodiment of the present application, the implementation of the system may refer to the implementation of the method, and repeated details are not repeated.
As shown in fig. 5, a system for determining association information between a user and an object according to an embodiment of the present application includes:
an information determination module 500, configured to select, for at least one user, feature information of an object determined by user behavior of the user from a feature information set, and select historical associated information of the user from a historical associated information set, where the feature information set includes vectorized feature information, and the historical associated information set includes vectorized historical associated information;
a correction module 501, configured to determine real-time correction information of the vectorized user according to the feature information of the object, the historical association information of the user, and the vectorized real-time user association information;
the processing module 502 is configured to, for at least one target object, determine association information between a user and the target object according to the user real-time modification information.
Optionally, the information determining module 500 is further configured to determine the feature information set and the historical associated information set according to the following manners:
regarding at least one object, taking the object and description information of the object of the same type as the object as object text information of the object; for at least one user, using description information of an object related to the user behavior in the offline data as historical user text information of the user;
training according to the object text information and the historical user text information to obtain vectorized feature information and vectorized historical associated information;
and placing the obtained characteristic information in the characteristic information set, and placing the obtained historical associated information in the historical associated information set.
Optionally, the modification module 501 is specifically configured to:
aiming at least one user, determining vectorized real-time user association information according to the user behavior vector of the user;
and correcting the historical associated information of the user according to the real-time user associated information and the associated correction information to obtain the vectorized real-time user correction information.
Optionally, the processing module 502 is further configured to:
and aiming at least one target object, determining at least one target object from all objects according to the objects determined by the user behavior of the user before determining the associated information of the user and the target object according to the real-time user correction information.
Optionally, the processing module 502 is specifically configured to:
and aiming at least one target object, determining the associated information of the user and the target object according to the real-time user correction information and the characteristic information of the target object in the characteristic information set.
Optionally, the processing module 502 is further configured to:
after determining the associated information of the user and the target object according to the real-time user correction information, the target object is sequenced according to the associated information of the user and the target object;
recommending the target object for the user according to the ordered target object sequence.
The modules in fig. 5 may be provided in one device, or may be provided in a plurality of devices. The functions of one module may be implemented by one device or by a plurality of devices.
The present application is described above with reference to block diagrams and/or flowchart illustrations of methods, apparatus (systems) and/or computer program products according to embodiments of the application. It will be understood that one block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, and/or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer and/or other programmable data processing apparatus, create means for implementing the functions/acts specified in the block diagrams and/or flowchart block or blocks.
Accordingly, the subject application may also be embodied in hardware and/or in software (including firmware, resident software, micro-code, etc.). Furthermore, the present application may take the form of a computer program product on a computer-usable or computer-readable storage medium having computer-usable or computer-readable program code embodied in the medium for use by or in connection with an instruction execution system. In the context of this application, a computer-usable or computer-readable medium may be any medium that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
Claims (10)
1. A method of determining association information of a user with an object, the method comprising:
for at least one user, selecting feature information of an object determined by user behaviors of the user from a feature information set, and selecting historical associated information of the user from a historical associated information set, wherein the feature information set comprises vectorized feature information, and the historical associated information set comprises vectorized historical associated information;
determining real-time correction information of the vectorized user according to the characteristic information of the object, the historical associated information of the user and the vectorized real-time user associated information;
aiming at least one target object, determining the association information of a user and the target object according to the real-time user correction information;
determining the feature information set and the historical associated information set according to the following modes:
regarding at least one object, taking the object and description information of the object of the same type as the object as object text information of the object; for at least one user, using description information of an object related to the user behavior in the offline data as historical user text information of the user;
training according to the object text information and the historical user text information to obtain vectorized feature information and vectorized historical associated information;
and placing the obtained characteristic information in the characteristic information set, and placing the obtained historical associated information in the historical associated information set.
2. The method of claim 1, wherein determining the vectored real-time user correction information based on the characteristic information of the object, the historical association information of the user, and the vectored real-time user association information comprises:
aiming at least one user, determining vectorized real-time user association information according to the user behavior vector of the user;
and correcting the historical associated information of the user according to the real-time user associated information and the associated correction information to obtain the vectorized real-time user correction information.
3. The method of claim 1, wherein for at least one target object, before determining the association information of the user with the target object according to the real-time user modification information, further comprising:
determining at least one target object from all objects according to the objects determined by the user behavior of the user.
4. The method of claim 1, wherein determining, for at least one target object, association information of a user with the target object from the user real-time revision information comprises:
and aiming at least one target object, determining the associated information of the user and the target object according to the real-time user correction information and the characteristic information of the target object in the characteristic information set.
5. The method according to any one of claims 1 to 4, wherein after determining, for at least one target object, the association information between the user and the target object according to the user real-time modification information, the method further comprises:
sequencing the target objects according to the associated information of the user and the target objects;
recommending the target object for the user according to the ordered target object sequence.
6. A system for determining association information of a user with an object, the system comprising:
an information determination module, configured to select, for at least one user, feature information of an object determined by user behavior of the user from a feature information set, and select historical associated information of the user from a historical associated information set, where the feature information set includes vectorized feature information, and the historical associated information set includes vectorized historical associated information;
the correction module is used for determining real-time correction information of the vectorized user according to the characteristic information of the object, the historical associated information of the user and the vectorized real-time user associated information;
the processing module is used for determining the association information of the user and the target object according to the real-time user correction information aiming at least one target object;
the information determination module is further configured to determine the feature information set and the history association information set according to the following manner:
regarding at least one object, taking the object and description information of the object of the same type as the object as object text information of the object; for at least one user, using description information of an object related to the user behavior in the offline data as historical user text information of the user;
training according to the object text information and the historical user text information to obtain vectorized feature information and vectorized historical associated information;
and placing the obtained characteristic information in the characteristic information set, and placing the obtained historical associated information in the historical associated information set.
7. The system of claim 6, wherein the modification module is specifically configured to:
aiming at least one user, determining vectorized real-time user association information according to the user behavior vector of the user;
and correcting the historical associated information of the user according to the real-time user associated information and the associated correction information to obtain the vectorized real-time user correction information.
8. The system of claim 6, wherein the processing module is further to:
and aiming at least one target object, determining at least one target object from all objects according to the objects determined by the user behavior of the user before determining the associated information of the user and the target object according to the real-time user correction information.
9. The system of claim 6, wherein the processing module is specifically configured to:
and aiming at least one target object, determining the associated information of the user and the target object according to the real-time user correction information and the characteristic information of the target object in the characteristic information set.
10. The system of any of claims 6 to 9, wherein the processing module is further configured to:
after determining the associated information of the user and the target object according to the real-time user correction information, the target object is sequenced according to the associated information of the user and the target object;
recommending the target object for the user according to the ordered target object sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710547349.0A CN109213923B (en) | 2017-07-06 | 2017-07-06 | Method and system for determining associated information of user and object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710547349.0A CN109213923B (en) | 2017-07-06 | 2017-07-06 | Method and system for determining associated information of user and object |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109213923A CN109213923A (en) | 2019-01-15 |
| CN109213923B true CN109213923B (en) | 2021-12-14 |
Family
ID=64992894
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710547349.0A Active CN109213923B (en) | 2017-07-06 | 2017-07-06 | Method and system for determining associated information of user and object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109213923B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102339438A (en) * | 2010-07-22 | 2012-02-01 | 阿里巴巴集团控股有限公司 | Commodity information website publishing method, system and device |
| CN102880974A (en) * | 2012-09-12 | 2013-01-16 | 江苏乐买到网络科技有限公司 | System for sending commodity group purchase information to user |
| CN102890686A (en) * | 2011-07-21 | 2013-01-23 | 腾讯科技（深圳）有限公司 | Method and system for showing commodity search result |
| CN103412937A (en) * | 2013-08-22 | 2013-11-27 | 成都数之联科技有限公司 | Searching and shopping method based on handheld terminal |
| CN104834657A (en) * | 2014-08-27 | 2015-08-12 | 腾讯科技（北京）有限公司 | User behavior analysis method and servers |
| 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 |
| CN105868299A (en) * | 2016-03-23 | 2016-08-17 | 深圳码隆科技有限公司 | Data search method and device |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4554653B2 (en) * | 2007-08-08 | 2010-09-29 | クラリオン株式会社 | Route search method, route search system, and navigation apparatus |
| US10354264B2 (en) * | 2014-03-24 | 2019-07-16 | Salesforce.Com, Inc. | Contact recommendations based on purchase history |
| US20160253325A1 (en) * | 2014-12-30 | 2016-09-01 | Socialtopias, Llc | Method and apparatus for programmatically adjusting the relative importance of content data as behavioral data changes |
| CN104821963B (en) * | 2015-04-30 | 2019-02-19 | 华为技术有限公司 | Recommendation information methods of exhibiting and device |
| CN106708938A (en) * | 2016-11-18 | 2017-05-24 | 北京大米科技有限公司 | Method and device for assisting recommendation |
-
2017
- 2017-07-06 CN CN201710547349.0A patent/CN109213923B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102339438A (en) * | 2010-07-22 | 2012-02-01 | 阿里巴巴集团控股有限公司 | Commodity information website publishing method, system and device |
| CN102890686A (en) * | 2011-07-21 | 2013-01-23 | 腾讯科技（深圳）有限公司 | Method and system for showing commodity search result |
| CN102880974A (en) * | 2012-09-12 | 2013-01-16 | 江苏乐买到网络科技有限公司 | System for sending commodity group purchase information to user |
| CN103412937A (en) * | 2013-08-22 | 2013-11-27 | 成都数之联科技有限公司 | Searching and shopping method based on handheld terminal |
| CN104834657A (en) * | 2014-08-27 | 2015-08-12 | 腾讯科技（北京）有限公司 | User behavior analysis method and servers |
| 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 |
| CN105868299A (en) * | 2016-03-23 | 2016-08-17 | 深圳码隆科技有限公司 | Data search method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109213923A (en) | 2019-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111209476B (en) | Recommendation method, model generation method, device, medium and equipment | |
| CN110909182B (en) | Multimedia resource searching method, device, computer equipment and storage medium | |
| CN110008397B (en) | Recommendation model training method and device | |
| CN109344314B (en) | Data processing method and device and server | |
| CN111506820B (en) | Recommendation model, recommendation method, recommendation device, recommendation equipment and recommendation storage medium | |
| CN109063188A (en) | A kind of entity recommended method and device | |
| CN110851699A (en) | Deep reinforcement learning-based information flow recommendation method, device, equipment and medium | |
| CN105677780A (en) | Scalable user intent mining method and system thereof | |
| CN107506495A (en) | Information-pushing method and device | |
| CN110910199A (en) | Item information sorting method and device, computer equipment and storage medium | |
| CN107766349B (en) | Method, device, equipment and client for generating text | |
| CN108307240A (en) | Video recommendation method and device | |
| CN110163703B (en) | Classification model establishing method, file pushing method and server | |
| CN109242593B (en) | Case reasoning-based clothing recommendation method and device and storage medium | |
| CN111695960A (en) | Object recommendation system, method, electronic device and storage medium | |
| CN107203558B (en) | Object recommendation method and device, and recommendation information processing method and device | |
| US10339586B1 (en) | Techniques for identifying similar products | |
| CN111159570B (en) | Information recommendation method and server | |
| CN104915860A (en) | Commodity recommendation method and device | |
| CN116894711A (en) | Commodity recommendation reason generation method and device and electronic equipment | |
| CN112100221A (en) | Information recommendation method and device, recommendation server and storage medium | |
| CN111967924A (en) | Commodity recommendation method, commodity recommendation device, computer device, and medium | |
| CN114398973B (en) | Media content tag identification method, device, equipment and storage medium | |
| CN109377284B (en) | Method and electronic equipment for pushing information | |
| CN110992127A (en) | Article recommendation 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |