CN108491529A - Information recommendation method and device - Google Patents
Information recommendation method and device Download PDFInfo
- Publication number
- CN108491529A CN108491529A CN201810266958.3A CN201810266958A CN108491529A CN 108491529 A CN108491529 A CN 108491529A CN 201810266958 A CN201810266958 A CN 201810266958A CN 108491529 A CN108491529 A CN 108491529A
- Authority
- CN
- China
- Prior art keywords
- user
- recommended
- lda
- theme
- vectors
- 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.)
- Granted
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of information recommendation method of present invention proposition and device, wherein method include:Obtain the historical behavior data of user to be recommended;Including:The information flow that user to be recommended clicked in default historical time section;The keyword in information flow is obtained, key words text is generated;Key words text is inputted into preset document subject matter and generates model LDA, obtains the corresponding LDA vectors of user to be recommended;LDA vectors include:Key words text belongs to the probability of each theme;According to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, the corresponding similar users of user to be recommended are determined;According to the historical behavior data of similar users, to user's recommendation information stream to be recommended, so as to combine historical behavior data and LDA models to obtain the corresponding LDA vectors of user to be recommended in time, avoid the use of user model, and the calculating of similar users is carried out according to the corresponding LDA vectors of each user, calculation amount is small, and calculating speed is fast, disclosure satisfy that requirement of real time.
Description
Technical field
The present invention relates to Internet technical field more particularly to a kind of information recommendation methods and device.
Background technology
Currently, the Collaborative Recommendation algorithm based on user is given mainly using the behavior for the similar users for giving user
User recommends, for example, the information flow clicked, browsed according to similar users, to give user's recommendation information stream.
Currently, determine there are mainly two types of the methods of the similar users of given user, one is obtaining user-item matrixes,
Matrix is analyzed, given user is obtained and is mapped to the vector after low-dimensional, according to the similarity between each vector, is determined given
The similar users of user, however this method needs the explicit feedback information based on user, can not be applied to implicit feedback information
Scene.Another method is the similarity calculated based on known users model between given user and other users.Wherein, it uses
Family model includes the attributes such as interest word, category of interest.However in this method, user model is the history click data based on user
Generate, formation speed is slow, and hysteresis quality is big, and based on user model carry out similarity calculation when, need to be traversed for user model and obtain
The attribute at family is taken, calculating speed is slow, it is difficult to meet requirement of real time.
Invention content
The present invention is directed to solve at least some of the technical problems in related technologies.
For this purpose, first purpose of the present invention is to propose a kind of information recommendation method, for solving phase in the prior art
It is slow like user's calculating speed, it is computationally intensive, it is difficult to the problem of meeting requirement of real time.
Second object of the present invention is to propose a kind of information recommending apparatus.
Third object of the present invention is to propose another information recommending apparatus.
Fourth object of the present invention is to propose a kind of non-transitorycomputer readable storage medium.
The 5th purpose of the present invention is to propose a kind of computer program product.
In order to achieve the above object, first aspect present invention embodiment proposes a kind of information recommendation method, including:
Obtain the historical behavior data of user to be recommended;The historical behavior data include:The user to be recommended exists
The information flow clicked in default historical time section;
The keyword in described information stream is obtained, key words text is generated;
The key words text is inputted into preset document subject matter and generates model LDA, the user to be recommended is obtained and corresponds to
LDA vector;The LDA vectors include:The key words text belongs to the probability of each theme;
According to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, institute is determined
State the corresponding similar users of user to be recommended;
According to the historical behavior data of the similar users, to user's recommendation information stream to be recommended.
Further, described corresponding according to the corresponding LDA vectors of user to be recommended and each candidate user
LDA vectors, determine the corresponding similar users of the user to be recommended, including:
According to the corresponding LDA vectors of the user to be recommended, at least one theme belonging to the user to be recommended is determined
Group;
For each theme group, according to each in the corresponding LDA vectors of user to be recommended and the theme group
The corresponding LDA vectors of a candidate user, calculate in the user to be recommended and the theme group between each candidate user
Similarity;
The candidate user that corresponding similarity is met to preset similarity threshold is determined as the user to be recommended and corresponds to
Similar users.
Further, each theme group is provided with corresponding similarity threshold.
Further, described to be directed to each theme group, according to the corresponding LDA vectors of the user to be recommended, Yi Jisuo
State the corresponding LDA vectors of each candidate user in theme group, calculate the user to be recommended with it is each in the theme group
Before similarity between candidate user, further include:
The corresponding LDA vectors of the user to be recommended are added in affiliated at least one theme group;
For each theme group belonging to the user to be recommended, the theme group is divided, is obtained at least
Two subgroups;
It is corresponding, it is described to be directed to each theme group, according to corresponding LDA vectors of the user to be recommended and described
The corresponding LDA vectors of each candidate user, calculate the user to be recommended and each time in the theme group in theme group
The similarity between family is selected, including:
For each theme group, acquisition includes the first subgroup of the corresponding LDA vectors of the user to be recommended;
It is corresponded to according to each candidate user in the corresponding LDA vectors of user to be recommended and first subgroup
LDA vectors, calculate the similarity between each candidate user in the user to be recommended and first subgroup.
Further, described according to the corresponding LDA vectors of the user to be recommended, it determines belonging to the user to be recommended
At least one theme group, including:
Obtain the theme that corresponding probability in the corresponding LDA vectors of the user to be recommended is more than preset probability threshold value;
It will be determined as the theme group belonging to the user to be recommended with the matched theme group of the theme.
Further, the keyword include in following information any one or it is a variety of:Described information stream is corresponding
Keyword in label, the corresponding search term of described information stream and described information flow content.
Further, the method further includes:
Obtain training sample;The training sample includes:Multiple key words texts and corresponding LDA vectors;
Initial LDA models are trained according to the training sample, obtain the preset LDA models.
Further, the historical behavior data according to the similar users, to user's recommendation information to be recommended
Stream, including:
The historical behavior data of the similar users are compared with the historical behavior data of the user to be recommended, really
The information flow to be recommended that do not clicked by the user to be recommended in the historical behavior data of the fixed similar users;
The information flow to be recommended is recommended into the user to be recommended.
The information recommendation method of the embodiment of the present invention, by the historical behavior data for obtaining user to be recommended;Historical behavior
Data include:The information flow that user to be recommended clicked in default historical time section;The keyword in information flow is obtained, it is raw
At key words text;Key words text is inputted into preset document subject matter and generates model LDA, it is corresponding to obtain user to be recommended
LDA vectors;LDA vectors include:Key words text belongs to the probability of each theme;According to the corresponding LDA of user to be recommended to
Amount and the corresponding LDA vectors of each candidate user, determine the corresponding similar users of user to be recommended;According to similar users
Historical behavior data, to user's recommendation information stream to be recommended, so as to combine historical behavior data and LDA models to obtain in time
The corresponding LDA vectors of user to be recommended are taken, avoid the use of user model, and phase is carried out according to the corresponding LDA vectors of each user
Like the calculating of user, calculation amount is small, and calculating speed is fast, disclosure satisfy that requirement of real time.
In order to achieve the above object, second aspect of the present invention embodiment proposes a kind of information recommending apparatus, including:
Acquisition module, the historical behavior data for obtaining user to be recommended;The historical behavior data include:It is described
The information flow that user to be recommended clicked in default historical time section;
Generation module generates key words text for obtaining the keyword in described information stream;
Input module generates model LDA for the key words text to be inputted preset document subject matter, obtains described wait for
The corresponding LDA vectors of recommended user;The LDA vectors include:The key words text belongs to the probability of each theme;
Determining module, for corresponding according to the corresponding LDA vectors of user to be recommended and each candidate user
LDA vectors, determine the corresponding similar users of the user to be recommended;
Recommending module, for the historical behavior data according to the similar users, to user's recommendation information to be recommended
Stream.
Further, the determining module includes:
Determination unit, for according to the corresponding LDA vectors of the user to be recommended, determining belonging to the user to be recommended
At least one theme group;
Computing unit, for being directed to each theme group, according to the corresponding LDA vectors of the user to be recommended, Yi Jisuo
State the corresponding LDA vectors of each candidate user in theme group, calculate the user to be recommended with it is each in the theme group
Similarity between candidate user;
The determination unit is additionally operable to meeting corresponding similarity into the candidate user of preset similarity threshold, determines
For the corresponding similar users of the user to be recommended.
Further, each theme group is provided with corresponding similarity threshold.
Further, the determining module further includes:Adding device and division unit;
The adding device, for the corresponding LDA vectors of the user to be recommended to be added to affiliated at least one master
It inscribes in group;
The division unit, each theme group for being directed to belonging to the user to be recommended, to the theme group
It is divided, obtains at least two subgroups;
Corresponding, the computing unit is specifically used for,
For each theme group, acquisition includes the first subgroup of the corresponding LDA vectors of the user to be recommended;
It is corresponded to according to each candidate user in the corresponding LDA vectors of user to be recommended and first subgroup
LDA vectors, calculate the similarity between each candidate user in the user to be recommended and first subgroup.
Further, the determination unit is specifically used for,
Obtain the theme that corresponding probability in the corresponding LDA vectors of the user to be recommended is more than preset probability threshold value;
It will be determined as the theme group belonging to the user to be recommended with the matched theme group of the theme.
Further, the keyword include in following information any one or it is a variety of:Described information stream is corresponding
Keyword in label, the corresponding search term of described information stream and described information flow content.
Further, the device further includes:Training module;
The acquisition module is additionally operable to obtain training sample;The training sample includes:Multiple key words texts, with
And corresponding LDA vectors;
The training module obtains described default for being trained to initial LDA models according to the training sample
LDA models.
Further, the recommending module is specifically used for,
The historical behavior data of the similar users are compared with the historical behavior data of the user to be recommended, really
The information flow to be recommended that do not clicked by the user to be recommended in the historical behavior data of the fixed similar users;
The information flow to be recommended is recommended into the user to be recommended.
The information recommending apparatus of the embodiment of the present invention, by the historical behavior data for obtaining user to be recommended;Historical behavior
Data include:The information flow that user to be recommended clicked in default historical time section;The keyword in information flow is obtained, it is raw
At key words text;Key words text is inputted into preset document subject matter and generates model LDA, it is corresponding to obtain user to be recommended
LDA vectors;LDA vectors include:Key words text belongs to the probability of each theme;According to the corresponding LDA of user to be recommended to
Amount and the corresponding LDA vectors of each candidate user, determine the corresponding similar users of user to be recommended;According to similar users
Historical behavior data, to user's recommendation information stream to be recommended, so as to combine historical behavior data and LDA models to obtain in time
The corresponding LDA vectors of user to be recommended are taken, avoid the use of user model, and phase is carried out according to the corresponding LDA vectors of each user
Like the calculating of user, calculation amount is small, and calculating speed is fast, disclosure satisfy that requirement of real time.
In order to achieve the above object, third aspect present invention embodiment proposes another information recommending apparatus, including:Storage
Device, processor and storage are on a memory and the computer program that can run on a processor, which is characterized in that the processor
Information recommendation method as described above is realized when executing described program.
To achieve the goals above, fourth aspect present invention embodiment proposes a kind of computer-readable storage of non-transitory
Medium is stored thereon with computer program, which realizes information recommendation method as described above when being executed by processor.
To achieve the goals above, fifth aspect present invention embodiment proposes a kind of computer program product, when described
When instruction processing unit in computer program product executes, a kind of information recommendation method is executed, the method includes:
Obtain the historical behavior data of user to be recommended;The historical behavior data include:The user to be recommended exists
The information flow clicked in default historical time section;
The keyword in described information stream is obtained, key words text is generated;
The key words text is inputted into preset document subject matter and generates model LDA, the user to be recommended is obtained and corresponds to
LDA vector;The LDA vectors include:The key words text belongs to the probability of each theme;
According to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, institute is determined
State the corresponding similar users of user to be recommended;
According to the historical behavior data of the similar users, to user's recommendation information stream to be recommended.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partly become from the following description
Obviously, or practice through the invention is recognized.
Description of the drawings
Above-mentioned and/or additional aspect and advantage of the invention will become from the following description of the accompanying drawings of embodiments
Obviously and it is readily appreciated that, wherein:
Fig. 1 is a kind of flow diagram of information recommendation method provided in an embodiment of the present invention;
Fig. 2 is the flow diagram of another information recommendation method provided in an embodiment of the present invention;
Fig. 3 is a kind of structural schematic diagram of information recommending apparatus provided in an embodiment of the present invention;
Fig. 4 is the structural schematic diagram of another information recommending apparatus provided in an embodiment of the present invention;
Fig. 5 is the structural schematic diagram of another information recommending apparatus provided in an embodiment of the present invention;
Fig. 6 is the structural schematic diagram of another information recommending apparatus provided in an embodiment of the present invention;
Fig. 7 is the structural schematic diagram of another information recommending apparatus provided in an embodiment of the present invention.
Specific implementation mode
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to for explaining the present invention, and is not considered as limiting the invention.
Below with reference to the accompanying drawings the information recommendation method and device of the embodiment of the present invention are described.
Fig. 1 is a kind of flow diagram of information recommendation method provided in an embodiment of the present invention.As shown in Figure 1, the information
Recommendation method includes the following steps:
S101, the historical behavior data for obtaining user to be recommended;Historical behavior data include:User to be recommended is default
The information flow clicked in historical time section.
The executive agent of information recommendation method provided by the invention is information recommending apparatus, and information recommending apparatus specifically can be with
For the software etc. installed on hardware device, such as terminal device, background server, server cluster etc. or hardware device.This
In embodiment, user to be recommended refers to the user for clicking and consulting various information flows.Information flow for example Baidu feed streams, news,
The user of information etc..Historical behavior data refer to that user to be recommended consults behavior to the click in various information.For example, going through
May include in history behavioral data:Information flow that user to be recommended clicked in default historical time section, to the point of information flow
Hit the time, consult time span, information flow particular content etc..In the present embodiment, the historical behavior data of user to be recommended can
It is acquired with the terminal device by display information flow and is reported to information recommending apparatus.Can also include user in historical behavior data
Mark, so that information recommending apparatus distinguishes different users according to the mark of user.
In the present embodiment, when recommended user consults information flow using multiple terminal devices, historical behavior data can be
The combination with the relevant historical behavior data of user to be recommended that multiple terminal devices report.Wherein, the mark of user for example may be used
Think, the account of user on Baidu's feed streams etc. can be with the mark of unique mark user.
Wherein, when default historical time section is before it can be current time in 1 hour, in one day, in one month
Between section.It should be noted that in the present embodiment, information recommending apparatus can periodically obtain the historical behavior of user to be recommended
Data determine similar users according to historical behavior data, and then to user's recommendation information stream to be recommended.In addition, information recommendation fills
Setting can also be when the historical behavior data of user to be recommended meets preset condition, to user's recommendation information stream.Preset condition example
Such as can be that the quantity for the information flow that user to be recommended consults within a certain period of time meets certain threshold value or use to be recommended
The time that information flow is consulted at family meets certain threshold value etc..Wherein, preset condition can be set according to actual needs.
Keyword in S102, acquisition information flow, generates key words text.
In the present embodiment, the keyword in information flow may include in following information any one or it is a variety of:Information
Flow the keyword in the corresponding search term of corresponding label, information flow and information flow content.Wherein, the corresponding mark of information flow
It signs such as can be the classification type of information flow, field.Wherein field such as sport, economy, biology, mathematics, education, amusement
Deng.Keyword in information flow content, the word obtained after operations such as to be segmented, being filtered to information flow content.
In the present embodiment, may include in key words text:Each key in the information flow that user to be recommended clicked
Word, alternatively, each keyword in the information flow that user to be recommended clicked and its word frequency.
S103, key words text is inputted into preset document subject matter generation model LDA, it is corresponding obtains user to be recommended
LDA vectors;LDA vectors include:Key words text belongs to the probability of each theme.
In the present embodiment, it is one three that document subject matter, which generates model (Latent Dirichlet Allocation, LDA),
Layer bayesian probability model, including word, theme and document three-decker.So-called document subject matter generates model, that is to say, that one
Each word of article is by " with some theme of certain probability selection, and with certain probability selection from this theme
Such a process of a word " obtains.Therefore, LDA models can be used for identifying extensive document sets (document
Collection the subject information) or in corpus (corpus) hidden.
In the present embodiment, vectorial expression, such as onehot vectors etc. may be used in each keyword in key words text, will
LDA models are inputted by the key words text that the corresponding vector of each keyword forms, obtain the LDA vectors of LDA models output.
In the present embodiment, the number of dimensions of LDA vectors is equal to the total quantity of theme;Each dimension of LDA vectors can divide
It Dui Ying not a theme;The value of each dimension of LDA vectors represents the probability that key words text belongs to corresponding theme.With theme packet
It includes:For theme 1, theme 2, theme 3, theme 4 and theme 5, the number of dimensions of LDA vectors is first dimension of 5, LDA vectors
The value of degree indicates that key words text belongs to the probability of theme 1;The value of second dimension of LDA vectors indicates that key words text belongs to
The probability of theme 2;The value of the third dimension of LDA vectors indicates that key words text belongs to the probability of theme 3;The of LDA vectors
The value of four dimensions indicates that key words text belongs to the probability of theme 4;The value of 5th dimension of LDA vectors indicates keyword text
Originally belong to the probability of theme 5.
Further, before step 103, the method can also include:Obtain training sample;It is wrapped in training sample
It includes:Multiple key words texts and corresponding LDA vectors;Initial LDA models are trained according to training sample, are obtained
Preset LDA models.Wherein, the corresponding LDA vectors of key words text can be according to relative users to each master in training sample
The hobby etc. of the information flow of topic is determined.
S104, it is waited for according to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, determination
The corresponding similar users of recommended user.
In the present embodiment, candidate user, which refers to, to be clicked in all users for consulting information flow, in addition to user to be recommended
Other users.In the present embodiment, before step 104, the method can also include:Periodically acquire each candidate user
Historical behavior data, obtain the corresponding LDA vectors of each candidate users with 103 with reference to step 102 and store, to calculate phase
It is transferred when like user.
Specifically, information recommending apparatus can be according to the corresponding LDA vectors of user to be recommended and each candidate user pair
The LDA vectors answered, calculate the distance between the corresponding LDA vectors of user to be recommended LDA vectors corresponding with candidate user, according to
Distance calculates the similarity between user to be recommended and candidate user, and corresponding similarity is met preset similarity threshold
Candidate user is determined as similar users.
S105, the historical behavior data according to similar users, to user's recommendation information stream to be recommended.
Specifically, the process that information recommending apparatus executes step 105 is specifically as follows, by the historical behavior number of similar users
It is compared, is determined in the historical behavior data of similar users not by use to be recommended according to the historical behavior data with user to be recommended
The information flow to be recommended that family was clicked;Information flow to be recommended is recommended into user to be recommended.
The information recommendation method of the embodiment of the present invention, by the historical behavior data for obtaining user to be recommended;Historical behavior
Data include:The information flow that user to be recommended clicked in default historical time section;The keyword in information flow is obtained, it is raw
At key words text;Key words text is inputted into preset document subject matter and generates model LDA, it is corresponding to obtain user to be recommended
LDA vectors;LDA vectors include:Key words text belongs to the probability of each theme;According to the corresponding LDA of user to be recommended to
Amount and the corresponding LDA vectors of each candidate user, determine the corresponding similar users of user to be recommended;According to similar users
Historical behavior data, to user's recommendation information stream to be recommended, so as to combine historical behavior data and LDA models to obtain in time
The corresponding LDA vectors of user to be recommended are taken, avoid the use of user model, and phase is carried out according to the corresponding LDA vectors of each user
Like the calculating of user, calculation amount is small, and calculating speed is fast, disclosure satisfy that requirement of real time.
Fig. 2 is the flow diagram of another information recommendation method provided in an embodiment of the present invention, as shown in Fig. 2, in Fig. 1
On the basis of illustrated embodiment, step 104 can specifically include following steps:
S1041, according to user to be recommended corresponding LDA vector, determine at least one theme group belonging to user to be recommended
Group.
Wherein, the process of information recommending apparatus execution step 1041 is specifically as follows, and obtains the corresponding LDA of user to be recommended
Corresponding probability is more than the theme of preset probability threshold value in vector;It will be determined as to be recommended with the matched theme group of theme
Theme group belonging to user.
In the present embodiment, each theme corresponds to a theme group, includes in each theme group:The value of respective dimensions
It is vectorial more than the LDA of preset probability threshold value.For example, being said so that first dimension of LDA vectors corresponds to the first theme as an example
It is bright, if the value of first dimension of the corresponding LDA vectors of candidate user is more than preset probability threshold value, candidate user is corresponded to
LDA vectors be added in the corresponding theme group of the first theme.
In addition, it is necessary to explanation, since the temperature of different themes is different, can be different masters in the present embodiment
Different probability threshold values is arranged in topic.For example, being illustrated so that first dimension of LDA vectors corresponds to the first theme as an example, if waiting
Select the value of first dimension of the corresponding LDA vectors in family to be more than the probability threshold value of the first dimension, then it is candidate user is corresponding
LDA vectors are added in the corresponding theme group of the first theme.
S1042, it is directed to each theme group, according to each in the corresponding LDA vectors of user to be recommended and theme group
The corresponding LDA vectors of candidate user, calculate the similarity between each candidate user in user to be recommended and theme group.
S1043, the candidate user that corresponding similarity is met to preset similarity threshold, are determined as user couple to be recommended
The similar users answered.
In the present embodiment, for each theme group, each candidate use in getting user to be recommended and theme group
After similarity between family, can by between each candidate user in user to be recommended and theme group similarity with it is default
Similarity threshold be compared, by corresponding similarity be more than preset similarity threshold candidate user, determination be the theme
The corresponding similar users of user to be recommended in group.Similar users in each theme group are combined, are obtained to be recommended
The corresponding similar users of user.
Can be the setting pair of each theme group in the present embodiment in addition, since different themes has different temperatures
The similarity threshold answered.The corresponding similarity threshold of each theme group may be the same or different.
In the present embodiment, by first determining at least one theme group belonging to user to be recommended, user to be recommended is calculated
It is candidate in other theme groups without calculating with the similarity between each candidate user in affiliated at least one theme group
Similarity between user and user to be recommended improves similar to reduce the calculation amount in similar users determination process
User's constant speed degree really, so as to user's recommendation information stream to be recommended, improve the advisory speed of information flow in time and push away
Recommend efficiency.
Further, for the calculation amount being further reduced in similar users determination process, the determination of similar users is improved
Speed, on the basis of embodiment shown in Fig. 2, before step 1042, the method can also include:By user couple to be recommended
The LDA vectors answered are added in affiliated at least one theme group;It is right for each theme group belonging to user to be recommended
Theme group is divided, and at least two subgroups are obtained.Wherein, the mode divided to theme group can be random draws
Point, or divided according to the value of dimension corresponding with the theme in each LDA vectors.
Corresponding, step 1042 is specifically as follows, and for each theme group, acquisition includes that user to be recommended is corresponding
First subgroup of LDA vectors;According to each candidate user in the corresponding LDA vectors of user to be recommended and the first subgroup
Corresponding LDA vectors, calculate the similarity between each candidate user in user to be recommended and first subgroup.
In the present embodiment, in the case that the quantity of LDA vectors is excessive in theme group, user to be recommended can be first determined
Affiliated at least one theme group, to theme, group divides, and determination includes the son of the corresponding LDA vectors of user to be recommended
Group calculates the similarity between each candidate user in user to be recommended and the subgroup, without calculating other subgroups
Similarity between middle candidate user and user to be recommended further reduces the calculation amount in similar users determination process, carries
High similar users constant speed degree really, so as to user's recommendation information stream to be recommended, improve the recommendation of information flow in time
Speed and recommendation efficiency.
Fig. 3 is a kind of structural schematic diagram of information recommending apparatus provided in an embodiment of the present invention.As shown in figure 3, including:It obtains
Modulus block 31, generation module 32, input module 33, determining module 34 and recommending module 35.
Wherein, acquisition module 31, the historical behavior data for obtaining user to be recommended;It is wrapped in the historical behavior data
It includes:The information flow that the user to be recommended clicked in default historical time section;
Generation module 32 generates key words text for obtaining the keyword in described information stream;
Input module 33 generates model LDA for the key words text to be inputted preset document subject matter, obtains described
The corresponding LDA vectors of user to be recommended;The LDA vectors include:The key words text belongs to the probability of each theme;
Determining module 34, for corresponding according to the corresponding LDA vectors of user to be recommended and each candidate user
LDA vectors, determine the corresponding similar users of the user to be recommended;
Recommending module 35, for the historical behavior data according to the similar users, to user's recommendation to be recommended
Breath stream.
Information recommending apparatus provided by the invention is specifically as follows hardware device, such as terminal device, background server, clothes
The software etc. installed on business device cluster etc. or hardware device.In the present embodiment, it is each that user to be recommended refers to that click is consulted
The user of kind information flow.User of the information flow such as Baidu feed streams, news, information.Historical behavior data, which refer to, to be waited pushing away
It recommends user and behavior is consulted to the click in various information.For example, may include in historical behavior data:User to be recommended is default
The information flow clicked in historical time section, to information flow click the time, consult time span, information flow particular content
Deng.In the present embodiment, the historical behavior data of user to be recommended can be acquired and be reported to by the terminal device of display information flow
Information recommending apparatus.The mark that can also include user in historical behavior data, so that information recommending apparatus is according to the mark of user
Know to distinguish different users.
In the present embodiment, when recommended user consults information flow using multiple terminal devices, historical behavior data can be
The combination with the relevant historical behavior data of user to be recommended that multiple terminal devices report.Wherein, the mark of user for example may be used
Think, the account of user on Baidu's feed streams etc. can be with the mark of unique mark user.
Wherein, when default historical time section is before it can be current time in 1 hour, in one day, in one month
Between section.It should be noted that in the present embodiment, information recommending apparatus can periodically obtain the historical behavior of user to be recommended
Data determine similar users according to historical behavior data, and then to user's recommendation information stream to be recommended.In addition, information recommendation fills
Setting can also be when the historical behavior data of user to be recommended meets preset condition, to user's recommendation information stream.Preset condition example
Such as can be that the quantity for the information flow that user to be recommended consults within a certain period of time meets certain threshold value or use to be recommended
The time that information flow is consulted at family meets certain threshold value etc..Wherein, preset condition can be set according to actual needs.
In the present embodiment, the keyword in information flow may include in following information any one or it is a variety of:Information
Flow the keyword in the corresponding search term of corresponding label, information flow and information flow content.Wherein, the corresponding mark of information flow
It signs such as can be the classification type of information flow, field.Wherein field such as sport, economy, biology, mathematics, education, amusement
Deng.Keyword in information flow content, the word obtained after operations such as to be segmented, being filtered to information flow content.
In the present embodiment, may include in key words text:Each key in the information flow that user to be recommended clicked
Word, alternatively, each keyword in the information flow that user to be recommended clicked and its word frequency.
In the present embodiment, the number of dimensions of LDA vectors is equal to the total quantity of theme;Each dimension of LDA vectors can divide
It Dui Ying not a theme;The value of each dimension of LDA vectors represents the probability that key words text belongs to corresponding theme.With theme packet
It includes:For theme 1, theme 2, theme 3, theme 4 and theme 5, the number of dimensions of LDA vectors is first dimension of 5, LDA vectors
The value of degree indicates that key words text belongs to the probability of theme 1;The value of second dimension of LDA vectors indicates that key words text belongs to
The probability of theme 2;The value of the third dimension of LDA vectors indicates that key words text belongs to the probability of theme 3;The of LDA vectors
The value of four dimensions indicates that key words text belongs to the probability of theme 4;The value of 5th dimension of LDA vectors indicates keyword text
Originally belong to the probability of theme 5.
In the present embodiment, candidate user, which refers to, to be clicked in all users for consulting information flow, in addition to user to be recommended
Other users.In the present embodiment, information recommending apparatus can periodically acquire the historical behavior data of each candidate user, hold
The function of row generation module 32 and input module 33, to obtain the corresponding LDA vectors of each candidate user and store, to calculate
It is transferred when similar users.
Specifically, information recommending apparatus can be according to the corresponding LDA vectors of user to be recommended and each candidate user pair
The LDA vectors answered, calculate the distance between the corresponding LDA vectors of user to be recommended LDA vectors corresponding with candidate user, according to
Distance calculates the similarity between user to be recommended and candidate user, and corresponding similarity is met preset similarity threshold
Candidate user is determined as similar users.
Further, recommending module 35 specifically can be used for, by the historical behavior data of similar users and user to be recommended
Historical behavior data be compared, determine in the historical behavior data of similar users and do not waited pushing away by what user to be recommended clicked
Recommend information flow;Information flow to be recommended is recommended into user to be recommended.
Further, in conjunction with reference to figure 4, on the basis of embodiment shown in Fig. 3, the device can also include:Instruction
Practice module 36;
Wherein, the acquisition module 31 is additionally operable to obtain training sample;The training sample includes:Multiple keywords
Text and corresponding LDA vectors;
The training module 36 obtains described pre- for being trained to initial LDA models according to the training sample
If LDA models.
The information recommending apparatus of the embodiment of the present invention, by the historical behavior data for obtaining user to be recommended;Historical behavior
Data include:The information flow that user to be recommended clicked in default historical time section;The keyword in information flow is obtained, it is raw
At key words text;Key words text is inputted into preset document subject matter and generates model LDA, it is corresponding to obtain user to be recommended
LDA vectors;LDA vectors include:Key words text belongs to the probability of each theme;According to the corresponding LDA of user to be recommended to
Amount and the corresponding LDA vectors of each candidate user, determine the corresponding similar users of user to be recommended;According to similar users
Historical behavior data, to user's recommendation information stream to be recommended, so as to combine historical behavior data and LDA models to obtain in time
The corresponding LDA vectors of user to be recommended are taken, avoid the use of user model, and phase is carried out according to the corresponding LDA vectors of each user
Like the calculating of user, calculation amount is small, and calculating speed is fast, disclosure satisfy that requirement of real time.
Fig. 5 is the structural schematic diagram of another information recommending apparatus provided in an embodiment of the present invention.As shown in figure 5, in Fig. 3
On the basis of illustrated embodiment, determining module 34 can specifically include:Determination unit 341 and computing unit 342.
Wherein it is determined that unit 341, for according to the corresponding LDA vectors of the user to be recommended, determining the use to be recommended
At least one theme group belonging to family;
Computing unit 342, for being directed to each theme group, according to the corresponding LDA vectors of the user to be recommended, and
The corresponding LDA vectors of each candidate user in the theme group, calculate the user to be recommended with it is each in the theme group
Similarity between a candidate user;
The determination unit 343 is additionally operable to meeting corresponding similarity into the candidate user of preset similarity threshold, really
It is set to the corresponding similar users of the user to be recommended.
Wherein it is determined that specifically to can be used for obtaining corresponding probability in the corresponding LDA vectors of user to be recommended big for unit 341
In the theme of preset probability threshold value;It will be determined as the theme group belonging to user to be recommended with the matched theme group of theme.
In the present embodiment, each theme corresponds to a theme group, includes in each theme group:The value of respective dimensions
It is vectorial more than the LDA of preset probability threshold value.For example, being said so that first dimension of LDA vectors corresponds to the first theme as an example
It is bright, if the value of first dimension of the corresponding LDA vectors of candidate user is more than preset probability threshold value, candidate user is corresponded to
LDA vectors be added in the corresponding theme group of the first theme.
In addition, it is necessary to explanation, since the temperature of different themes is different, can be different masters in the present embodiment
Different probability threshold values is arranged in topic.For example, being illustrated so that first dimension of LDA vectors corresponds to the first theme as an example, if waiting
Select the value of first dimension of the corresponding LDA vectors in family to be more than the probability threshold value of the first dimension, then it is candidate user is corresponding
LDA vectors are added in the corresponding theme group of the first theme.In the present embodiment, for each theme group, waited for getting
After similarity in recommended user and theme group between each candidate user, can will user to be recommended in theme group
Similarity between each candidate user is compared with preset similarity threshold, and corresponding similarity is more than preset phase
Like the candidate user of degree threshold value, the corresponding similar users of user to be recommended in group that are the theme are determined.It will be in each theme group
Similar users be combined, obtain the corresponding similar users of user to be recommended.
Can be the setting pair of each theme group in the present embodiment in addition, since different themes has different temperatures
The similarity threshold answered.The corresponding similarity threshold of each theme group may be the same or different.
In the present embodiment, by first determining at least one theme group belonging to user to be recommended, user to be recommended is calculated
It is candidate in other theme groups without calculating with the similarity between each candidate user in affiliated at least one theme group
Similarity between user and user to be recommended improves similar to reduce the calculation amount in similar users determination process
User's constant speed degree really, so as to user's recommendation information stream to be recommended, improve the advisory speed of information flow in time and push away
Recommend efficiency.
Further, in conjunction with reference to figure 6, on the basis of embodiment shown in Fig. 5, determining module 34 can also include:Add
Add unit 343 and division unit 344.
Wherein, the adding device 343, for belonging to the corresponding LDA vectors of the user to be recommended are added to extremely
In a few theme group;
The division unit 344, each theme group for being directed to belonging to the user to be recommended, to the theme group
Group is divided, and at least two subgroups are obtained;
Corresponding, the computing unit 342 is specifically used for,
For each theme group, acquisition includes the first subgroup of the corresponding LDA vectors of the user to be recommended;
It is corresponded to according to each candidate user in the corresponding LDA vectors of user to be recommended and first subgroup
LDA vectors, calculate the similarity between each candidate user in the user to be recommended and first subgroup.
In the present embodiment, in the case that the quantity of LDA vectors is excessive in theme group, user to be recommended can be first determined
Affiliated at least one theme group, to theme, group divides, and determination includes the son of the corresponding LDA vectors of user to be recommended
Group calculates the similarity between each candidate user in user to be recommended and the subgroup, without calculating other subgroups
Similarity between middle candidate user and user to be recommended further reduces the calculation amount in similar users determination process, carries
High similar users constant speed degree really, so as to user's recommendation information stream to be recommended, improve the recommendation of information flow in time
Speed and recommendation efficiency.
Fig. 7 is the structural schematic diagram of another information recommending apparatus provided in an embodiment of the present invention.The information recommending apparatus
Including:
Memory 1001, processor 1002 and it is stored in the calculating that can be run on memory 1001 and on processor 1002
Machine program.
Processor 1002 realizes the information recommendation method provided in above-described embodiment when executing described program.
Further, information recommending apparatus further includes:
Communication interface 1003, for the communication between memory 1001 and processor 1002.
Memory 1001, for storing the computer program that can be run on processor 1002.
Memory 1001 may include high-speed RAM memory, it is also possible to further include nonvolatile memory (non-
Volatile memory), a for example, at least magnetic disk storage.
Processor 1002 realizes the information recommendation method described in above-described embodiment when for executing described program.
If memory 1001, processor 1002 and the independent realization of communication interface 1003, communication interface 1003, memory
1001 and processor 1002 can be connected with each other by bus and complete mutual communication.The bus can be industrial standard
Architecture (Industry Standard Architecture, referred to as ISA) bus, external equipment interconnection
(Peripheral Component, referred to as PCI) bus or extended industry-standard architecture (Extended Industry
Standard Architecture, referred to as EISA) bus etc..The bus can be divided into address bus, data/address bus, control
Bus processed etc..For ease of indicating, only indicated with a thick line in Fig. 7, it is not intended that an only bus or a type of
Bus.
Optionally, in specific implementation, if memory 1001, processor 1002 and communication interface 1003, are integrated in one
It is realized on block chip, then memory 1001, processor 1002 and communication interface 1003 can be completed mutual by internal interface
Communication.
Processor 1002 may be a central processing unit (Central Processing Unit, referred to as CPU), or
Person is specific integrated circuit (Application Specific Integrated Circuit, referred to as ASIC) or quilt
It is configured to implement one or more integrated circuits of the embodiment of the present invention.
The present invention also provides a kind of non-transitorycomputer readable storage mediums, are stored thereon with computer program, the journey
Information recommendation method as described above is realized when sequence is executed by processor.
The present invention also provides a kind of computer program products, when the instruction processing unit in the computer program product executes
When, a kind of information recommendation method is executed, the method includes:
Obtain the historical behavior data of user to be recommended;The historical behavior data include:The user to be recommended exists
The information flow clicked in default historical time section;
The keyword in described information stream is obtained, key words text is generated;
The key words text is inputted into preset document subject matter and generates model LDA, the user to be recommended is obtained and corresponds to
LDA vector;The LDA vectors include:The key words text belongs to the probability of each theme;
According to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, institute is determined
State the corresponding similar users of user to be recommended;
According to the historical behavior data of the similar users, to user's recommendation information stream to be recommended.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
It must be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be in office
It can be combined in any suitable manner in one or more embodiments or example.In addition, without conflicting with each other, the skill of this field
Art personnel can tie the feature of different embodiments or examples described in this specification and different embodiments or examples
It closes and combines.
In addition, term " first ", " second " are used for description purposes only, it is not understood to indicate or imply relative importance
Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or
Implicitly include at least one this feature.In the description of the present invention, the meaning of " plurality " is at least two, such as two, three
It is a etc., unless otherwise specifically defined.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes
It is one or more for realizing custom logic function or process the step of executable instruction code module, segment or portion
Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discuss suitable
Sequence, include according to involved function by it is basic simultaneously in the way of or in the opposite order, to execute function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered use
In the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, for
Instruction execution system, device or equipment (system of such as computer based system including processor or other can be held from instruction
The instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or set
It is standby and use.For the purpose of this specification, " computer-readable medium " can any can be included, store, communicating, propagating or passing
Defeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipment
It sets.The more specific example (non-exhaustive list) of computer-readable medium includes following:Electricity with one or more wiring
Interconnecting piece (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory
(ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable optic disk is read-only deposits
Reservoir (CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other are suitable
Medium, because can be for example by carrying out optical scanner to paper or other media, then into edlin, interpretation or when necessary with it
His suitable method is handled electronically to obtain described program, is then stored in computer storage.
It should be appreciated that each section of the present invention can be realized with hardware, software, firmware or combination thereof.Above-mentioned
In embodiment, software that multiple steps or method can in memory and by suitable instruction execution system be executed with storage
Or firmware is realized.Such as, if realized in another embodiment with hardware, following skill well known in the art can be used
Any one of art or their combination are realized:With for data-signal realize logic function logic gates from
Logic circuit is dissipated, the application-specific integrated circuit with suitable combinational logic gate circuit, programmable gate array (PGA), scene can compile
Journey gate array (FPGA) etc..
Those skilled in the art are appreciated that realize all or part of step that above-described embodiment method carries
Suddenly it is that relevant hardware can be instructed to complete by program, the program can be stored in a kind of computer-readable storage medium
In matter, which includes the steps that one or a combination set of embodiment of the method when being executed.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing module, it can also
That each unit physically exists alone, can also two or more units be integrated in a module.Above-mentioned integrated mould
The form that hardware had both may be used in block is realized, can also be realized in the form of software function module.The integrated module is such as
Fruit is realized in the form of software function module and when sold or used as an independent product, can also be stored in a computer
In read/write memory medium.
Storage medium mentioned above can be read-only memory, disk or CD etc..Although having been shown and retouching above
The embodiment of the present invention is stated, it is to be understood that above-described embodiment is exemplary, and should not be understood as the limit to the present invention
System, those skilled in the art can be changed above-described embodiment, change, replace and become within the scope of the invention
Type.
Claims (19)
1. a kind of information recommendation method, which is characterized in that including:
Obtain the historical behavior data of user to be recommended;The historical behavior data include:The user to be recommended is default
The information flow clicked in historical time section;
The keyword in described information stream is obtained, key words text is generated;
The key words text is inputted into preset document subject matter and generates model LDA, obtains the corresponding LDA of the user to be recommended
Vector;The LDA vectors include:The key words text belongs to the probability of each theme;
According to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, waited for described in determination
The corresponding similar users of recommended user;
According to the historical behavior data of the similar users, to user's recommendation information stream to be recommended.
2. according to the method described in claim 1, it is characterized in that, described according to the corresponding LDA vectors of the user to be recommended,
And the corresponding LDA vectors of each candidate user, determine the corresponding similar users of the user to be recommended, including:
According to the corresponding LDA vectors of the user to be recommended, at least one theme group belonging to the user to be recommended is determined;
For each theme group, according to each time in the corresponding LDA vectors of user to be recommended and the theme group
The corresponding LDA vectors in family are selected, are calculated similar between the user to be recommended and each candidate user in the theme group
Degree;
The candidate user that corresponding similarity is met to preset similarity threshold is determined as the corresponding phase of the user to be recommended
Like user.
3. according to the method described in claim 2, it is characterized in that, each theme group is provided with corresponding similarity threshold.
4. according to the method described in claim 2, it is characterized in that, described be directed to each theme group, according to described to be recommended
The corresponding LDA vectors of each candidate user, calculate described to be recommended in user corresponding LDA vectors and the theme group
Before similarity in user and the theme group between each candidate user, further include:
The corresponding LDA vectors of the user to be recommended are added in affiliated at least one theme group;
For each theme group belonging to the user to be recommended, the theme group is divided, obtains at least two
Subgroup;
It is corresponding, it is described to be directed to each theme group, according to the corresponding LDA vectors of user to be recommended and the theme
The corresponding LDA vectors of each candidate user in group calculate the user to be recommended and candidate are used with each in the theme group
Similarity between family, including:
For each theme group, acquisition includes the first subgroup of the corresponding LDA vectors of the user to be recommended;
According to the corresponding LDA of each candidate user in the corresponding LDA vectors of user to be recommended and first subgroup
Vector calculates the similarity between each candidate user in the user to be recommended and first subgroup.
5. according to the method described in claim 2, it is characterized in that, described according to the corresponding LDA vectors of the user to be recommended,
Determine at least one theme group belonging to the user to be recommended, including:
Obtain the theme that corresponding probability in the corresponding LDA vectors of the user to be recommended is more than preset probability threshold value;
It will be determined as the theme group belonging to the user to be recommended with the matched theme group of the theme.
6. according to the method described in claim 1, it is characterized in that, the keyword include in following information any one or
Person is a variety of:Key in the corresponding label of described information stream, the corresponding search term of described information stream and described information flow content
Word.
7. according to the method described in claim 1, it is characterized in that, further including:
Obtain training sample;The training sample includes:Multiple key words texts and corresponding LDA vectors;
Initial LDA models are trained according to the training sample, obtain the preset LDA models.
8. according to the method described in claim 1, it is characterized in that, the historical behavior data according to the similar users,
To user's recommendation information stream to be recommended, including:
The historical behavior data of the similar users are compared with the historical behavior data of the user to be recommended, determine institute
State the information flow to be recommended that do not clicked by the user to be recommended in the historical behavior data of similar users;
The information flow to be recommended is recommended into the user to be recommended.
9. a kind of information recommending apparatus, which is characterized in that including:
Acquisition module, the historical behavior data for obtaining user to be recommended;The historical behavior data include:It is described to wait pushing away
Recommend the information flow that user clicked in default historical time section;
Generation module generates key words text for obtaining the keyword in described information stream;
Input module generates model LDA for the key words text to be inputted preset document subject matter, obtains described to be recommended
The corresponding LDA vectors of user;The LDA vectors include:The key words text belongs to the probability of each theme;
Determining module, for according to the corresponding LDA vectors of the user to be recommended and the corresponding LDA of each candidate user to
Amount, determines the corresponding similar users of the user to be recommended;
Recommending module, for the historical behavior data according to the similar users, to user's recommendation information stream to be recommended.
10. device according to claim 9, which is characterized in that the determining module includes:
Determination unit, for according to the corresponding LDA vectors of the user to be recommended, determining belonging to the user to be recommended at least
One theme group;
Computing unit, for being directed to each theme group, according to the corresponding LDA vectors of user to be recommended and the master
The corresponding LDA vectors of each candidate user in group are inscribed, the user to be recommended and each candidate in the theme group are calculated
Similarity between user;
The determination unit is additionally operable to meeting corresponding similarity into the candidate user of preset similarity threshold, is determined as institute
State the corresponding similar users of user to be recommended.
11. device according to claim 10, which is characterized in that each theme group is provided with corresponding similarity threshold
Value.
12. device according to claim 10, which is characterized in that the determining module further includes:Adding device and division
Unit;
The adding device, for the corresponding LDA vectors of the user to be recommended to be added to affiliated at least one theme group
In group;
The division unit, for for each theme group belonging to the user to be recommended, being carried out to the theme group
It divides, obtains at least two subgroups;
Corresponding, the computing unit is specifically used for,
For each theme group, acquisition includes the first subgroup of the corresponding LDA vectors of the user to be recommended;
According to the corresponding LDA of each candidate user in the corresponding LDA vectors of user to be recommended and first subgroup
Vector calculates the similarity between each candidate user in the user to be recommended and first subgroup.
13. device according to claim 10, which is characterized in that the determination unit is specifically used for,
Obtain the theme that corresponding probability in the corresponding LDA vectors of the user to be recommended is more than preset probability threshold value;
It will be determined as the theme group belonging to the user to be recommended with the matched theme group of the theme.
14. device according to claim 9, which is characterized in that the keyword includes any one in following information
Or it is a variety of:Pass in the corresponding label of described information stream, the corresponding search term of described information stream and described information flow content
Keyword.
15. device according to claim 9, which is characterized in that further include:Training module;
The acquisition module is additionally operable to obtain training sample;The training sample includes:Multiple key words texts and right
The LDA vectors answered;
The training module obtains described preset for being trained to initial LDA models according to the training sample
LDA models.
16. device according to claim 9, which is characterized in that the recommending module is specifically used for,
The historical behavior data of the similar users are compared with the historical behavior data of the user to be recommended, determine institute
State the information flow to be recommended that do not clicked by the user to be recommended in the historical behavior data of similar users;
The information flow to be recommended is recommended into the user to be recommended.
17. a kind of information recommending apparatus, which is characterized in that including:
Memory, processor and storage are on a memory and the computer program that can run on a processor, which is characterized in that institute
It states when processor executes described program and realizes such as information recommendation method according to any one of claims 1-8.
18. a kind of non-transitorycomputer readable storage medium, is stored thereon with computer program, which is characterized in that the program
Such as information recommendation method according to any one of claims 1-8 is realized when being executed by processor.
19. a kind of computer program product executes a kind of letter when the instruction processing unit in the computer program product executes
Recommendation method is ceased, the method includes:
Obtain the historical behavior data of user to be recommended;The historical behavior data include:The user to be recommended is default
The information flow clicked in historical time section;
The keyword in described information stream is obtained, key words text is generated;
The key words text is inputted into preset document subject matter and generates model LDA, obtains the corresponding LDA of the user to be recommended
Vector;The LDA vectors include:The key words text belongs to the probability of each theme;
According to the corresponding LDA vectors of user to be recommended and the corresponding LDA vectors of each candidate user, waited for described in determination
The corresponding similar users of recommended user;
According to the historical behavior data of the similar users, to user's recommendation information stream to be recommended.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810266958.3A CN108491529B (en) | 2018-03-28 | 2018-03-28 | Information recommendation method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810266958.3A CN108491529B (en) | 2018-03-28 | 2018-03-28 | Information recommendation method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108491529A true CN108491529A (en) | 2018-09-04 |
CN108491529B CN108491529B (en) | 2021-11-16 |
Family
ID=63317022
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810266958.3A Active CN108491529B (en) | 2018-03-28 | 2018-03-28 | Information recommendation method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108491529B (en) |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109325146A (en) * | 2018-11-12 | 2019-02-12 | 平安科技(深圳)有限公司 | A kind of video recommendation method, device, storage medium and server |
CN109409928A (en) * | 2018-09-14 | 2019-03-01 | 上海晶赞融宣科技有限公司 | A kind of material recommended method, device, storage medium, terminal |
CN109740056A (en) * | 2018-12-28 | 2019-05-10 | 丹翰智能科技(上海)有限公司 | It is a kind of for provide a user customization financial information method and apparatus |
CN109885826A (en) * | 2019-01-07 | 2019-06-14 | 平安科技(深圳)有限公司 | Text term vector acquisition methods, device, computer equipment and storage medium |
CN110046304A (en) * | 2019-04-18 | 2019-07-23 | 腾讯科技(深圳)有限公司 | A kind of user's recommended method and device |
CN110083774A (en) * | 2019-05-10 | 2019-08-02 | 腾讯科技(深圳)有限公司 | Using determination method, apparatus, computer equipment and the storage medium of recommendation list |
CN110162703A (en) * | 2019-05-13 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Content recommendation method, training method, device, equipment and storage medium |
CN110472154A (en) * | 2019-08-26 | 2019-11-19 | 秒针信息技术有限公司 | A kind of resource supplying method, apparatus, electronic equipment and readable storage medium storing program for executing |
CN110634047A (en) * | 2019-09-05 | 2019-12-31 | 北京无限光场科技有限公司 | Method and device for recommending house resources, electronic equipment and storage medium |
CN110688464A (en) * | 2019-10-12 | 2020-01-14 | 苏州思必驰信息科技有限公司 | Man-machine conversation method and system |
CN111104590A (en) * | 2019-11-19 | 2020-05-05 | 泰康保险集团股份有限公司 | Information recommendation method, device, medium and electronic equipment |
CN111476621A (en) * | 2019-01-24 | 2020-07-31 | 百度在线网络技术(北京)有限公司 | User item recommendation method and device |
CN111666401A (en) * | 2020-05-29 | 2020-09-15 | 平安科技(深圳)有限公司 | Official document recommendation method and device based on graph structure, computer equipment and medium |
CN111950733A (en) * | 2019-05-15 | 2020-11-17 | 阿里巴巴集团控股有限公司 | Information flow sorting method and device and computer storage medium |
CN112241486A (en) * | 2019-07-17 | 2021-01-19 | 北京达佳互联信息技术有限公司 | Multimedia information acquisition method and device |
CN112434140A (en) * | 2020-11-10 | 2021-03-02 | 杭州博联智能科技股份有限公司 | Reply information processing method and system |
CN113204699A (en) * | 2020-02-03 | 2021-08-03 | 北京达佳互联信息技术有限公司 | Information recommendation method and device, electronic equipment and storage medium |
CN116662556A (en) * | 2023-08-02 | 2023-08-29 | 天河超级计算淮海分中心 | Text data processing method integrating user attributes |
CN111950733B (en) * | 2019-05-15 | 2024-06-11 | 阿里巴巴集团控股有限公司 | Method and device for ordering information streams and computer storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104731958A (en) * | 2015-04-03 | 2015-06-24 | 北京航空航天大学 | User-demand-oriented cloud manufacturing service recommendation method |
CN105069121A (en) * | 2015-08-12 | 2015-11-18 | 北京暴风科技股份有限公司 | Video pushing method based on video theme similarity |
JP2015219583A (en) * | 2014-05-14 | 2015-12-07 | 日本電信電話株式会社 | Topic determination device, utterance device, method, and program |
CN105824911A (en) * | 2016-03-15 | 2016-08-03 | 山东大学 | Video recommending method based on LDA user theme model |
US20160314195A1 (en) * | 2015-04-24 | 2016-10-27 | Linkedin Corporation | Detecting and combining synonymous topics |
CN106202394A (en) * | 2016-07-07 | 2016-12-07 | 腾讯科技(深圳)有限公司 | The recommendation method and system of text information |
US20170124192A1 (en) * | 2015-11-03 | 2017-05-04 | Dell Products, Lp | Systems and Methods for Website Improvement |
-
2018
- 2018-03-28 CN CN201810266958.3A patent/CN108491529B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2015219583A (en) * | 2014-05-14 | 2015-12-07 | 日本電信電話株式会社 | Topic determination device, utterance device, method, and program |
CN104731958A (en) * | 2015-04-03 | 2015-06-24 | 北京航空航天大学 | User-demand-oriented cloud manufacturing service recommendation method |
US20160314195A1 (en) * | 2015-04-24 | 2016-10-27 | Linkedin Corporation | Detecting and combining synonymous topics |
CN105069121A (en) * | 2015-08-12 | 2015-11-18 | 北京暴风科技股份有限公司 | Video pushing method based on video theme similarity |
US20170124192A1 (en) * | 2015-11-03 | 2017-05-04 | Dell Products, Lp | Systems and Methods for Website Improvement |
CN105824911A (en) * | 2016-03-15 | 2016-08-03 | 山东大学 | Video recommending method based on LDA user theme model |
CN106202394A (en) * | 2016-07-07 | 2016-12-07 | 腾讯科技(深圳)有限公司 | The recommendation method and system of text information |
Non-Patent Citations (1)
Title |
---|
石磊等: "个性化微博实时推荐模型研究", 《小型微型计算机系统》 * |
Cited By (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109409928A (en) * | 2018-09-14 | 2019-03-01 | 上海晶赞融宣科技有限公司 | A kind of material recommended method, device, storage medium, terminal |
CN109325146A (en) * | 2018-11-12 | 2019-02-12 | 平安科技(深圳)有限公司 | A kind of video recommendation method, device, storage medium and server |
CN109325146B (en) * | 2018-11-12 | 2024-05-07 | 平安科技(深圳)有限公司 | Video recommendation method and device, storage medium and server |
CN109740056A (en) * | 2018-12-28 | 2019-05-10 | 丹翰智能科技(上海)有限公司 | It is a kind of for provide a user customization financial information method and apparatus |
CN109885826A (en) * | 2019-01-07 | 2019-06-14 | 平安科技(深圳)有限公司 | Text term vector acquisition methods, device, computer equipment and storage medium |
CN111476621A (en) * | 2019-01-24 | 2020-07-31 | 百度在线网络技术(北京)有限公司 | User item recommendation method and device |
CN111476621B (en) * | 2019-01-24 | 2023-09-22 | 百度在线网络技术(北京)有限公司 | User article recommendation method and device |
CN110046304A (en) * | 2019-04-18 | 2019-07-23 | 腾讯科技(深圳)有限公司 | A kind of user's recommended method and device |
CN110046304B (en) * | 2019-04-18 | 2022-12-13 | 腾讯科技(深圳)有限公司 | User recommendation method and device |
CN110083774B (en) * | 2019-05-10 | 2023-11-03 | 腾讯科技(深圳)有限公司 | Method and device for determining application recommendation list, computer equipment and storage medium |
CN110083774A (en) * | 2019-05-10 | 2019-08-02 | 腾讯科技(深圳)有限公司 | Using determination method, apparatus, computer equipment and the storage medium of recommendation list |
CN110162703A (en) * | 2019-05-13 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Content recommendation method, training method, device, equipment and storage medium |
CN111950733B (en) * | 2019-05-15 | 2024-06-11 | 阿里巴巴集团控股有限公司 | Method and device for ordering information streams and computer storage medium |
CN111950733A (en) * | 2019-05-15 | 2020-11-17 | 阿里巴巴集团控股有限公司 | Information flow sorting method and device and computer storage medium |
CN112241486A (en) * | 2019-07-17 | 2021-01-19 | 北京达佳互联信息技术有限公司 | Multimedia information acquisition method and device |
CN110472154A (en) * | 2019-08-26 | 2019-11-19 | 秒针信息技术有限公司 | A kind of resource supplying method, apparatus, electronic equipment and readable storage medium storing program for executing |
CN110634047B (en) * | 2019-09-05 | 2022-04-22 | 北京无限光场科技有限公司 | Method and device for recommending house resources, electronic equipment and storage medium |
CN110634047A (en) * | 2019-09-05 | 2019-12-31 | 北京无限光场科技有限公司 | Method and device for recommending house resources, electronic equipment and storage medium |
CN110688464A (en) * | 2019-10-12 | 2020-01-14 | 苏州思必驰信息科技有限公司 | Man-machine conversation method and system |
CN111104590A (en) * | 2019-11-19 | 2020-05-05 | 泰康保险集团股份有限公司 | Information recommendation method, device, medium and electronic equipment |
CN113204699B (en) * | 2020-02-03 | 2023-10-31 | 北京达佳互联信息技术有限公司 | Information recommendation method and device, electronic equipment and storage medium |
CN113204699A (en) * | 2020-02-03 | 2021-08-03 | 北京达佳互联信息技术有限公司 | Information recommendation method and device, electronic equipment and storage medium |
CN111666401B (en) * | 2020-05-29 | 2023-06-30 | 平安科技(深圳)有限公司 | Document recommendation method, device, computer equipment and medium based on graph structure |
CN111666401A (en) * | 2020-05-29 | 2020-09-15 | 平安科技(深圳)有限公司 | Official document recommendation method and device based on graph structure, computer equipment and medium |
CN112434140A (en) * | 2020-11-10 | 2021-03-02 | 杭州博联智能科技股份有限公司 | Reply information processing method and system |
CN112434140B (en) * | 2020-11-10 | 2024-02-09 | 杭州博联智能科技股份有限公司 | Reply information processing method and system |
CN116662556A (en) * | 2023-08-02 | 2023-08-29 | 天河超级计算淮海分中心 | Text data processing method integrating user attributes |
CN116662556B (en) * | 2023-08-02 | 2023-10-20 | 天河超级计算淮海分中心 | Text data processing method integrating user attributes |
Also Published As
Publication number | Publication date |
---|---|
CN108491529B (en) | 2021-11-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108491529A (en) | Information recommendation method and device | |
CN104462364B (en) | Method and device is recommended in search | |
Chen et al. | Neural attentional rating regression with review-level explanations | |
Achakulvisut et al. | Science Concierge: A fast content-based recommendation system for scientific publications | |
Ren et al. | Social collaborative viewpoint regression with explainable recommendations | |
Zhao et al. | Improving user topic interest profiles by behavior factorization | |
Zhong et al. | Web service recommendation with reconstructed profile from mashup descriptions | |
CN104915734B (en) | Commodity popularity prediction method based on time sequence and system thereof | |
CN108153723A (en) | Hot spot information comment generation method, device and terminal device | |
Mozetič et al. | How to evaluate sentiment classifiers for Twitter time-ordered data? | |
CN109508373B (en) | Method and device for calculating enterprise public opinion index and computer readable storage medium | |
US11416907B2 (en) | Unbiased search and user feedback analytics | |
CN108287821A (en) | A kind of high-quality text screening technique, device and electronic equipment | |
CN109657116A (en) | A kind of public sentiment searching method, searcher, storage medium and terminal device | |
CN110032679A (en) | A method of the dynamic news based on level attention network is recommended | |
US20190065588A1 (en) | Page journey determination from fingerprint information in web event journals | |
Liu et al. | Explainable recommender systems via resolving learning representations | |
Ma et al. | MMM: multi-source multi-net micro-video recommendation with clustered hidden item representation learning | |
CN113590945B (en) | Book recommendation method and device based on user borrowing behavior-interest prediction | |
CN107784127A (en) | A kind of focus localization method and device | |
CN107748801A (en) | News recommends method, apparatus, terminal device and computer-readable recording medium | |
CN106610989B (en) | Search keyword clustering method and device | |
CN108021713B (en) | Document clustering method and device | |
CN111259213B (en) | Data visualization processing method and device | |
Muthivhi et al. | Multi-modal recommendation system with auxiliary information |
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 |