CN107665208A - User preference measure and device - Google Patents
User preference measure and device Download PDFInfo
- Publication number
- CN107665208A CN107665208A CN201610607241.1A CN201610607241A CN107665208A CN 107665208 A CN107665208 A CN 107665208A CN 201610607241 A CN201610607241 A CN 201610607241A CN 107665208 A CN107665208 A CN 107665208A
- Authority
- CN
- China
- Prior art keywords
- vector
- theme
- behavior
- targeted customer
- preset web
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention discloses a kind of user preference measure and device, method to include:The behavior for obtaining targeted customer accesses vector;Wherein, behavior accesses the access frequency included in vector to each preset web;The behavior of targeted customer is accessed into vector input to default topic model;Topic model includes the corresponding relation that pre-set user accesses vector and theme vector with its behavior;Include preference value of the pre-set user to each theme corresponding to preset web in theme vector;Obtain the target topic vector of topic model output;According to the preference of target topic vector metric targeted customer., being capable of conveniently and efficiently measure user preference using the embodiment of the present invention.
Description
Technical field
The present invention relates to data mining technology field, more particularly to a kind of user preference measure and device.
Background technology
With the development of network, various websites provide the mode of identical browser interface to different user, can not meet
The individual demand of user.In order to meet the individual demand of user, webmaster needs to be fully understood by the inclined of each user
It is good, and personalized service is provided the user on this basis.
Wherein, the method that tradition obtains user preference, is to go to check each net that user is accessed by artificial mode
Page, and the mark that can reflect these Webpage contents is stamped to the user.Such as:Staff views targeted customer institute
The webpage accessed has the webpage on physical culture such as basketball, football, is also related to the webpage of the automobile of various brands.If target
The webpage quantity on physical culture that user accessed is more on the webpage quantity of automobile than accessing, and now staff will be visited
The webpage asked is abstracted as physical culture and automobile, and is the first preference by sports logo, is the second preference by automobile logo.Then will
These preference informations are labeled in after user's mark, and the preference of targeted customer can be thus judged according to the mark, so as to
Further provided personalized service for the targeted customer.But the operation of this method is relatively complicated, and Websites quantity and number of users
Measure more, very labor intensive cost.
Therefore, need badly and a kind of new user preference metric scheme is provided, with conveniently and efficiently measure user preference.
The content of the invention
In view of the above problems, it is proposed that the present invention so as to provide one kind overcome above mentioned problem or at least in part solve on
State the user preference measure and device of problem.Concrete technical scheme is as follows:
In a first aspect, the embodiments of the invention provide a kind of user preference measure, methods described can include:
The behavior for obtaining targeted customer accesses vector;Wherein, the behavior is accessed in vector and included to each preset web
Access frequency;
The behavior of the targeted customer is accessed into vector input to default topic model;The topic model is to default
The behavior of user accesses vector and is trained acquisition, wherein contain pre-set user accesses vector and theme vector with its behavior
Corresponding relation;Include preference value of the pre-set user to each theme corresponding to preset web in the theme vector;
Obtain the target topic vector of the topic model output;
According to the preference of targeted customer described in the target topic vector metric.
Alternatively, the pre-set user included in the topic model accesses the corresponding pass of vector and theme vector with its behavior
System, it is:
The identification information of pre-set user accesses the corresponding relation of vector and theme vector with its behavior;
Before the behavior for obtaining targeted customer accesses vector, it can also include:
Obtain the identification information of targeted customer;
According to the identification information, judge whether the targeted customer is pre-set user;
If not, perform the step that the behavior for obtaining targeted customer accesses vector;
If it is, the identification information is inputted into the topic model, the target topic of the topic model output is obtained
Vector;Perform described according to the target topic vector metric the step of preference of targeted customer.
Alternatively, the default topic model obtains for structure in advance;Building the topic model in advance can include
Following steps:
Obtain the access frequency that each pre-set user accesses each preset web;
Using each access frequency, build the behavior corresponding to each pre-set user and access vector;
Vector is accessed based on the number of topics N to be built and constructed each behavior, utilizes preset themes model algorithm
Build the topic model;
Wherein, the topic model includes N number of theme, in addition to each behavior accesses the theme corresponding to vector
Vector, wherein, cover each preset web in each theme, and in the topic model record have it is described each default
Webpage weighted value shared in each theme, wherein, each preset web weight shared in each theme
It is worth incomplete same.
Alternatively, before the access frequency for obtaining each pre-set user and accessing each preset web, the side
Method can also include:
The text corresponding to each pre-set user is built, wherein, corresponding predesignated subscriber is recorded in text and is accessed
The preset web crossed;Wherein, a word in each preset web text as corresponding in text;
Correspondingly, the access frequency for obtaining each pre-set user and accessing each preset web, including:
The frequency that each preset web in the text corresponding to each pre-set user occurs is calculated, it is each to obtain
The access frequency of each preset web in text, and will not occur in the text corresponding to each pre-set user each
The access frequency of individual preset web is arranged to zero.
Alternatively, record has each preset web weighted value shared in each theme in the topic model,
Wherein, each preset web weighted value shared in each theme is incomplete same;
The target topic vector for obtaining the topic model output, can include:
The weighted value shared by each preset web that will be covered in each theme, used respectively with the target
The access frequency that preset web is corresponded in the behavior access vector at family is weighted read group total, obtains the targeted customer to institute
State the preference value of each theme;
Based on the preference value, the target topic vector is obtained.
Alternatively, described to be based on the preference value, obtaining the target topic vector can be:
(theme 1, theme 2 ..., theme N)=(X1、X2、……、XN), wherein,Described in the Xi is represented
Preference value of the targeted customer to i-th of theme;
Correspondingly, the preference of the targeted customer according to the target topic vector metric, including:
Utilize the X1To XNCorresponding value, measure preference of the targeted customer to each theme.
Alternatively, the weighted value shared by each preset web that will be covered in each theme, respectively
The calculating that access frequency with corresponding to preset web in the behavior access vector of the targeted customer is weighted read group total is public
Formula can be:
Wherein, the A represents the preset web, and the k represents k-th of theme, the p (Ai) represent the target line
To access the access frequency of i-th of preset web in vector, the p (Ai| themek) represent k-th of theme in the topic model
Weighted value shared by lower i-th of preset web, the v represent that the behavior of the targeted customer accesses v-th of default net in vector
Page, wherein, the i≤v, the k ∈ (1, N).
Alternatively, after the topic model using preset themes model algorithm structure, methods described can also wrap
Include:
Based on to each theme institute threshold value set in advance in the topic model, to corresponding to each pre-set user
Theme vector classified, to be clustered to each pre-set user.
Alternatively, the access frequency of each preset web, can be the access frequency of each default URL or default domain names
Rate.
Second aspect, the embodiments of the invention provide a kind of user preference measurement apparatus, described device can include:First
Acquiring unit, the first input block, the first construction unit, second acquisition unit and metric element;
The first acquisition unit, the behavior for obtaining targeted customer access vector;Wherein, the behavior accesses vector
In include to the access frequency of each preset web;
First input block, it is single to the described first structure for the behavior of the targeted customer to be accessed into vector input
The topic model that member is built in advance;The topic model is that the behavior access vector to pre-set user is trained acquisition, its
In contain the corresponding relation that pre-set user and its behavior access vector and theme vector;Include in the theme vector default
Preference value of the user to each theme corresponding to preset web;
The second acquisition unit, for obtaining the target topic vector of the topic model output;
The metric element, the preference for the targeted customer according to the target topic vector metric.
Alternatively, the pre-set user included in the topic model of the first construction unit structure accesses vector with its behavior
And the corresponding relation of theme vector, it is:
The identification information of pre-set user accesses the corresponding relation of vector and theme vector with its behavior;
Described device can also include:3rd acquiring unit, judging unit and the second input block;
3rd acquiring unit, for before accessing vector in the behavior for obtaining targeted customer, obtaining targeted customer's
Identification information;
The judging unit, for according to the identification information, judging whether the targeted customer is pre-set user;
If not, trigger the first acquisition unit;
If it is, the identification information is inputted into the topic model by second input block, the master is obtained
Inscribe the target topic vector of model output;Trigger the metric element.
Alternatively, first construction unit can include:First acquisition module, the first structure module and the second structure mould
Block;
First acquisition module, the access frequency of each preset web is accessed for obtaining each pre-set user;
The first structure module, for utilizing each access frequency, builds the row corresponding to each pre-set user
To access vector;
The second structure module, for accessing vector based on the number of topics N to be built and constructed each behavior,
The topic model is built using preset themes model algorithm;
Wherein, the topic model includes N number of theme, in addition to each behavior accesses the theme corresponding to vector
Vector, wherein, cover each preset web in each theme, and in the topic model record have it is described each default
Webpage weighted value shared in each theme, wherein, each preset web weight shared in each theme
It is worth incomplete same.
Alternatively, described device can also include:Second construction unit;
Second construction unit, for described each using each pre-set user access of first acquisition module acquisition
Before the access frequency of individual preset web, the text corresponding to each pre-set user is built, wherein, recorded in text corresponding
The preset web that predesignated subscriber accessed;Wherein, a word in each preset web text as corresponding in text;
Correspondingly, first acquisition module, can include:Calculating sub module;
The calculating sub module, go out for calculating each preset web in the text corresponding to each pre-set user
Existing frequency, to obtain the access frequency of each preset web in each text, and by corresponding to each pre-set user
Text in the access frequency of each preset web that did not occurred be arranged to zero.
Alternatively, record has each preset web each in the topic model constructed by first construction unit
Shared weighted value in theme, wherein, each preset web incomplete phase of weighted value shared in each theme
Together;
The second acquisition unit, it can include:It is single that preference value computation subunit and target topic vector obtain son
Member;
The preference value computation subunit, for each preset web that will be covered in each theme
Shared weighted value, it is weighted respectively with corresponding to the access frequency of preset web in the behavior access vector of the targeted customer
Read group total, obtain preference value of the targeted customer to each theme;
Target topic vector obtains subelement, for based on the preference value, obtain the target topic to
Amount.
Alternatively, the target topic vector, which obtains the target topic vector that subelement obtains, to be:
(theme 1, theme 2 ..., theme N)=(X1、X2、……、XN), wherein,Described in the Xi is represented
Preference value of the targeted customer to i-th of theme;
Correspondingly, the metric element, can include:Measure subelement;
The measurement subelement, for utilizing the X1To XNCorresponding value, the targeted customer is measured to each theme
Preference.
Alternatively, the preference value computation subunit calculates preference journey of the targeted customer to each theme
The calculation formula that angle value is utilized can be:
Wherein, the A represents the preset web, and the k represents k-th of theme, the p (Ai) represent the target line
It is described to access the access frequency of i-th of preset web in vectorp(Ai| themek) represent k-th of theme in the topic model
Weighted value shared by lower i-th of preset web, the v represent that the behavior of the targeted customer accesses v-th of default net in vector
Page, wherein, the i≤v, the k ∈ (1, N).
Alternatively, described device can also include:Taxon;
The taxon, the theme mould is being built using preset themes model algorithm for the described second structure module
After type, based on to each theme institute threshold value set in advance in the topic model, to corresponding to each pre-set user
Theme vector is classified, to be clustered to each pre-set user.
Alternatively, the access frequency of each preset web, it is the access frequency of each default URL or default domain names.
By above-mentioned technical proposal, user preference measure and device provided by the invention, the row of targeted customer is obtained
To access vector;The behavior is accessed into vector input into the topic model built in advance, so as to what is built in advance from this
The target topic vector corresponding to behavior access vector is obtained in topic model;So as to pass through institute in the target topic vector
The targeted customer of record the preference of metric objective user, reduces metric objective user's to the preference value of each theme
The cumbersome degree of preference, improve the speed of metric objective user preference.
Described above is only the general introduction of technical solution of the present invention, in order to better understand the technological means of the present invention,
And can be practiced according to the content of specification, and in order to allow above and other objects of the present invention, feature and advantage can
Become apparent, below especially exemplified by the embodiment of the present invention.
Brief description of the drawings
By reading the detailed description of hereafter preferred embodiment, it is various other the advantages of and benefit it is common for this area
Technical staff will be clear understanding.Accompanying drawing is only used for showing the purpose of preferred embodiment, and is not considered as to the present invention
Limitation.And in whole accompanying drawing, identical part is denoted by the same reference numerals.In the accompanying drawings:
Fig. 1 shows a kind of user preference measure schematic diagram of the embodiment of the present invention;
Fig. 2 shows another user preference measure schematic diagram of the embodiment of the present invention;
Fig. 3 shows a kind of flow of the topic model of structure in advance in the user preference measure of the embodiment of the present invention
Figure;
Fig. 4 shows another structure topic model flow in advance in the user preference measure of the embodiment of the present invention
Figure;
Fig. 5 shows another user preference measure schematic diagram of the embodiment of the present invention;
Fig. 6 shows a kind of user preference measurement apparatus structured flowchart of the embodiment of the present invention.
Embodiment
The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although the disclosure is shown in accompanying drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
Limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
Completely it is communicated to those skilled in the art.
In order to solve prior art problem, the embodiments of the invention provide a kind of user preference measure and device.
The user preference measure provided first below the embodiment of the present invention is introduced.
It is understood that realizing the functional software for the user preference measure that the embodiment of the present invention is provided can be
It is arranged at software special in terminal, or the feature card being arranged in the existing software in terminal, this is all reasonable
's.Wherein, the terminal can be computer or server apparatus.
It is worth noting that, the embodiment of the present invention can be entered by the topic model built in advance to the preference of any netizen
Row measurement.Wherein, the topic model built in advance is obtained by being trained by the behavior to numerous netizens access webpage.
The behavior of the access webpage specifically refers to the times or frequency that numerous netizens are accessed each preset web, i.e., involved in the present invention
Topic model be based on the access times to each preset web or frequency composition building of corpus, rather than based on pair
The building of corpus of the web page contents composition of webpage.
As shown in figure 1, the user preference measure that the embodiment of the present invention is provided, may include steps of:
S101:The behavior for obtaining targeted customer accesses vector;Wherein, the behavior accesses in vector and included to each default net
The access frequency of page;
Wherein, when needing to carry out preference measurement to targeted customer, first obtaining can reflect the targeted customer to each pre-
If the behavior of the access frequency of webpage accesses vector.Wherein it is possible to using the targeted customer acquired to each default net
The access times of page, calculate the access frequency to each preset web, are used so as to build the target using each access frequency
The behavior at family accesses vector.
It is emphasized that any two behavior accesses the preset web corresponding to the access frequency of same position in vector
It is identical.Such as behavior accesses vectorial a=(V1、V2、……、VN), behavior accesses vectorial b=(M1、M2、……、MN), then V1 and M1
Corresponding same preset web, VNAnd MNSame preset web is equally corresponded to, wherein, V and M are access frequency.Wherein, may be used
The preset web corresponding to each position in behavior access vector is preset, and each behavior accesses any two position in vector
Corresponding preset web is put to differ.
It is worth noting that, the targeted customer can be one or more netizens, you can to obtain the behavior of a netizen
Vector is accessed, the behavior that can also obtain multiple netizens simultaneously accesses vector.In addition, netizen's quantity corresponding to the targeted customer
It can be determined according to specific measurement demand, netizen's quantity corresponding to targeted customer not limited herein.Wherein, the netizen can be with
It is interpreted as the network user that web page access behavior be present.
It is understood that by the form of expression for accessing behavior and forming behavior and accessing vector of preset web, it is really
Web page access behavior to the targeted customer is abstracted as a kind of description of characterization, that is, accesses vector and embody the targeted customer institute
Which preset web was accessed, and the feature of the access frequency of each preset web accessed.
S102:The behavior of the targeted customer is accessed into vector input to default topic model;The topic model is to pre-
If the behavior of user accesses vector and is trained acquisition, wherein contain pre-set user and its behavior access vector and theme to
The corresponding relation of amount;Include preference value of the pre-set user to each theme corresponding to preset web in the theme vector;
Wherein, access vector by the behavior to pre-set user to be trained, obtain topic model.The training in advance obtains
Topic model include:Pre-set user accesses the corresponding relation of vector and theme vector with its behavior.In each theme vector
Include preference value of the pre-set user to each theme, wherein, include all preset webs under each theme,
The simply weighted under different themes shared by each preset web.It is understood that the pre-set user refer to it is default
The substantial amounts of network user.
It is understood that in the prior art, topic model is a kind of popular common language of natural language processing field
Adopted model, and sentence is converted into the vector representation form of word by the model using word as base unit, and to the vector of many sentences
The corpus of composition is modeled, that is, corpus is modeled according to topic model algorithm, obtains topic model, and institute
Theme corresponding to each word occurs comprising some themes, under each theme in obtained topic model frequency and each sentence to
Amount.
And the topic model of advance structure involved in the present invention is using each preset web as base unit, by default use
The access behavior to each preset web at family is converted to as the vector representation form constructed by preset web access frequency, at once
To access vector.It is modeled by the corpus that numerous behaviors are accessed with vector composition, obtains present subject matter model.
S103:Obtain the target topic vector of topic model output;
It is emphasized that the present invention is not improved in itself for topic model algorithm, that is to say, that institute of the present invention
The computational methods for the structure topic model being related to are using topic model algorithm in the prior art;And due to theme in the prior art
Model belongs to the classical model of academia, the i.e. Computational frame with comparative maturity, specific topic model algorithm is not entered herein
Row is described in detail.And the theme vector corresponding to training vector is obtained using the topic model built in advance, equally there is maturation
Algorithm, the algorithm for obtaining the theme vector corresponding to input vector is not described in detail herein.
Wherein, the behavior of targeted customer is accessed into vector input into default topic model, using existing algorithm meter
Calculate the target topic vector corresponding to behavior access vector.
S104:According to the preference of the target topic vector metric targeted customer.
Wherein, by including pre-set user in theme vector to the preference value for each theme for training to obtain,
Therefore the preference of metric objective user, can be avoided according to numerical value corresponding in target topic vector by artificial mode
Remove to check the content of pages for each webpage that the targeted customer is accessed, and stamped to the targeted customer and can reflect these pages
The troublesome operation of the mark of content, realize the preference of conveniently and efficiently metric objective user.
In embodiments of the present invention, the behavior for obtaining targeted customer accesses vector;Vector will be accessed the behavior to input to pre-
In the topic model first built, accessed so as to obtain the behavior in the topic model that is built in advance from this corresponding to vector
Target topic vector;So as to the preference by the targeted customer recorded in the target topic vector to each theme
Value, the preference of metric objective user, reduces the cumbersome degree of the preference of metric objective user, improves metric objective user preference
Speed.
Another user preference measure of the embodiment of the present invention is illustrated with reference to Fig. 2.
As shown in Fig. 2 the pre-set user included in the topic model accesses pair of vector and theme vector with its behavior
It should be related to, be:When the identification information of pre-set user accesses vectorial and theme vector corresponding relation with its behavior, in step S101
Before, step can also be included:
S105:Obtain the identification information of targeted customer;
S106:According to the identification information, judge whether the targeted customer is pre-set user;
If not, the step of performing S101;
If it is, perform step S107;After performing step S107, step S104 is performed;
Wherein, step S107 is:The identification information is inputted into the topic model, obtains the target master of topic model output
Topic vector.
It is understood that in this kind of implementation, the identification information of targeted customer can be first obtained, so as to root
According to identification information judgment, whether the targeted customer is pre-set user.Due to having contained the identification information of pre-set user in model
The corresponding relation of vector and theme vector is accessed with its behavior.So when targeted customer is pre-set user, will directly can be somebody's turn to do
The identification information input model of targeted customer, model can be according to the identification information passes corresponding with its theme vector of pre-set user
System directly exports target topic vector.
It is worth noting that, the identification information of targeted customer can be the user name of the targeted customer, or other energy
Enough uniquely determine numbering of the targeted customer etc..Wherein, acquired identification information can be the identification information of a netizen,
Can be the identification information of multiple netizens, what this was possible to.
A kind of construction step of the topic model of structure in advance in the embodiment of the present invention is introduced with reference to Fig. 3.
It may include steps of as shown in figure 3, building the topic model in advance:
S1:Obtain the access frequency that each pre-set user accesses each preset web;
S2:Using each access frequency, build the behavior corresponding to each pre-set user and access vector;
S3:Vector is accessed based on the number of topics N to be built and constructed each behavior, calculated using preset themes model
Method builds the topic model;
Wherein, the topic model includes N number of theme, in addition to each behavior access theme corresponding to vector to
Amount, wherein, each preset web is covered in each theme, and record has each preset web each in the topic model
Shared weighted value in individual theme, wherein, each preset web weighted value shared in each theme is incomplete same.
It is understood that in this kind of implementation, it is necessary in advance before theme vector is obtained using topic model
Train corresponding topic model.Wherein, vector is accessed as instruction by the use of the behavior of the access frequency structure by each preset web
Practice data training topic model.After the number of topics N to be built is specified, you can train the topic model of N number of theme.Its
In, each preset web is covered in each theme, and each preset web under each theme corresponds to certain weighted value.
Wherein, it is used as training data by the use of the behavior access vector of the access frequency structure by each preset web, you can
Realize the characterization description to user so that need not be default to this according to the probability that the web page contents (word) in preset web occur
Webpage is identified, and to realize the mark for accessing pre-set user webpage, improves the speed for characterizing description pre-set user.
It is emphasized that each pre-set user can be those skilled in the art has net according to determined by preset rules
The subnetwork user of access to web page behavior or the all-network user that web page access behavior be present.And this is each pre-
Obtained if being extracted in the database that the preset web that user accessed there can be netizen's web page access behavior from record, certainly not
It is confined to this.
It is worth noting that, the number of topics N to be built is those skilled in the art's root in structure topic model in advance
According to number of topics determined by specific need, the numerical value corresponding to the N is not specifically limited herein.
In addition, preset themes model algorithm, which can be LDA, (Latent Dirichlet Allocation, implies Di Like
Thunder is distributed) algorithm corresponding to topic model or LSA (Latent Semantic Space, latent semantic space) it is main
Algorithm corresponding to inscribing model, is not limited thereto certainly.
The construction step of another topic model of structure in advance in the embodiment of the present invention is introduced with reference to Fig. 4.
As shown in figure 4, in step S1:, should before obtaining the access frequency that each pre-set user accesses each preset web
Method can also comprise the following steps:
S4:The text corresponding to each pre-set user is built, wherein, corresponding predesignated subscriber is recorded in text and is visited
The preset web asked;Wherein, a word in each preset web text as corresponding in text;
Correspondingly, step S1 corresponds to following steps:
S11:The frequency that each preset web in the text corresponding to each pre-set user occurs is calculated, it is each to obtain
The access frequency of each preset web in individual text, and will not occur in the text corresponding to each pre-set user each
The access frequency of individual preset web is arranged to zero.
It is understood that in this kind of implementation, all preset webs that each pre-set user can be accessed
Preserve into a text.And regard each preset web in each text as a virtual word.Calculate each in the text
The frequency that virtual word occurs, that is, obtain the access frequency for each preset web that the pre-set user corresponding to the text accessed
Rate.Wherein, the access frequency for each preset web that the pre-set user corresponding to the text has not visited is arranged to zero.
That is, during topic model is trained, vector, this reality are accessed in order to obtain the behavior of each pre-set user
Apply all preset webs that example first accessed each pre-set user to preserve into a text, then according to each pre-set user
The history access information recorded in text, the frequency that each preset web recorded in each text occurs is calculated, so as to obtain
The behavior for obtaining the pre-set user accesses vector.
In addition, the text of each pre-set user can be stored in topic model, so can easily inquire about default
The history access information of user.
Another user preference measure of the embodiment of the present invention is illustrated with reference to Fig. 5.
Shown in Fig. 5, record have each preset web shared by each theme in the topic model built in advance
During weighted value, wherein, each preset web weighted value shared in each theme is incomplete same;
Correspondingly, step S103:The target topic vector of acquisition topic model output, can include:
S1031:By the weighted value shared by each preset web covered in each theme, used respectively with the target
The access frequency that preset web is corresponded in the behavior access vector at family is weighted read group total, and it is each to this to obtain the targeted customer
The preference value of individual theme;
S1032:Based on the preference value, target topic vector is obtained.
It is understood that in this kind of implementation, each default net that will can be covered in each theme
Weighted value shared by page, the access frequency for accessing corresponding preset web in vector with the behavior of the targeted customer respectively are added
Read group total is weighed, so as to obtain preference value of the targeted customer to each theme.Utilize be calculated it is each partially
Good degree value, build target topic vector.
It is emphasized that the side of the theme vector corresponding to the targeted customer is calculated using the topic model built in advance
Method is not limited thereto.
Alternatively, the preference value should be based on, obtaining the target topic vector can be:
(theme 1, theme 2 ..., theme N)=(X1、X2、……、XN), wherein,The Xi represents the target
Preference value of the user to i-th of theme;
Correspondingly, this is according to the preference of the target topic vector metric targeted customer, including:
Utilize the X1To XNCorresponding value, measure preference of the targeted customer to each theme.
It is understood that in this kind of implementation, each value that can be in theme vector, metric objective user
To the preference of each theme.
Such as:Assuming that 3 themes of the topic model training in advance built in advance, and the theme corresponding to targeted customer to
Measure as (theme 1, theme 2, theme 3)=(0.6,0.3,0.1), that is to say, that the targeted customer is in the webpage of 1 type that is the theme
Access behavior it is more, next to that the webpage of the type of theme 2, therefore the targeted customer can be measured out according to the theme vector most
Like theme 1, theme 2 takes second place, followed by theme 3.
It is emphasized that it is assumed above merely exemplary, it should not form the restriction to the embodiment of the present invention.
Alternatively, the weighted value shared by by each preset web covered in each theme, respectively with the mesh
The behavior of mark user, which accesses, to be corresponded to the access frequency of preset web and is weighted the calculation formula of read group total and can be in vector:
Wherein, the A represents the preset web, and the k represents k-th of theme, the p (Ai) represent that the goal behavior accesses vector
In i-th of preset web access frequency, shouldp(Ai| themek) represent in the topic model i-th of default net under k-th of theme
Weighted value shared by page, the v represent that the behavior of the targeted customer accesses v-th of preset web in vector, wherein, the i≤v, the k
∈ (1, N).
Such as:2 themes are included in topic model, be the theme 1 and theme 2 respectively, and 2 default nets are included under each theme
Page is preset web 1 and preset web 2.It is 1 in the weight that 1 time preset web of theme 1 accounts for, the weight that preset web 2 accounts for is 0.
It is 0 in the weight that 2 times preset webs of theme 1 account for, the weight that preset web 2 accounts for is 1.
The behavior access vector that targeted customer is calculated is (preset web 1, preset web 2)=(0.7,0.3).Then count
Calculation obtains the targeted customer and is to the preference value of theme 1:0.7*1+0.3*0=0.7;The targeted customer is calculated to master
Topic 2 preference value be:0.7*0+0.3*1=0.3.That is, the target topic vector of the resulting targeted customer
For (0.7,0.3).
It is emphasized that the example above is merely illustrative, the restriction to the embodiment of the present invention should not be formed.
It is understood that in this kind of implementation, can utilize above-mentioned calculation formula calculate goal behavior access to
The preference value of the corresponding each theme of amount.Vector is accessed so as to build the goal behavior according to each preference value
Corresponding target topic vector, does not limit to this kind of calculation certainly.
Alternatively, a kind of embodiment as the present invention, the topic model is built at this using preset themes model algorithm
Afterwards, this method also includes:
Based on to each theme institute threshold value set in advance in the topic model, to the master corresponding to each pre-set user
Topic vector is classified, to be clustered to each pre-set user.
It is understood that in this kind of implementation, after training obtains topic model, the topic model includes more
Individual theme vector, each theme vector can be classified according to each theme threshold value set in advance, so as to reality
Now each pre-set user is clustered.And the theme vector classification results can be stored in trained topic model,
So as to according to after to a certain pre-set user measure user preference, you can with its of cluster where obtaining a certain pre-set user
The preference of his pre-set user, improve preference measurement results.
It is of course also possible to the theme vector classification results are stored to default memory, for those skilled in the art
The theme vector classification results can be easily extracted, and then the user in same cluster is provided on preference theme degree
High web page interlinkage is recommended, and to meet the individual demand of user, is not limited thereto certainly.
For example, all users of the preference value corresponding to theme in theme vector 1 more than 0.5 can be divided
For one kind, it is not limited thereto certainly.
It is emphasized that this can be the value of a determination to each theme threshold value set in advance, can also correspond to
One span, and the threshold value set in advance can be set by those skilled in the art according to limitation demand, herein
It is not detailed.
Alternatively, the access frequency of each preset web, it is the access frequency of each default URL or default domain names.
It is understood that in this kind of implementation, when each behavior of structure present subject matter model accesses vector
When being made up of each default URL access frequency, it can be visited as each behavior constructed by the access frequency of each default URL
Ask that vector carries out theme training, and then obtain the theme vector corresponding to each behavior access vector.
When each behavior for building present subject matter model, which accesses vector, to be made up of the access frequency of each default domain name,
Vector can be accessed as each behavior constructed by the access frequency of each default domain name and carry out theme training, and then be somebody's turn to do
Each behavior accesses the theme vector corresponding to vector.This kind of mode, due to thousands of individual URL can be included under same domain name,
That is, the quantity of domain name will be much smaller than URL quantity, therefore it can greatly reduce the amount of calculation of theme training.
It is emphasized that URL and domain name belong to existing concept, URL and domain name are not specifically described herein.
The user preference measurement apparatus provided below the embodiment of the present invention is introduced.
Corresponding to above method embodiment, as shown in fig. 6, the embodiment of the present invention additionally provides a kind of user preference measurement dress
Put, the device can include:First acquisition unit 601, the first input block 602, the first construction unit 603, second obtain single
Member 604 and metric element 605;
The first acquisition unit 601, the behavior for obtaining targeted customer access vector;Wherein, the behavior accesses vector
In include to the access frequency of each preset web;
First input block 602, for the behavior of the targeted customer to be accessed into vector input to first construction unit
The topic model built in advance;The topic model is that the behavior access vector to pre-set user is trained acquisition, wherein wrapping
The corresponding relation that pre-set user accesses vector and theme vector with its behavior is contained;Include pre-set user pair in the theme vector
The preference value of each theme corresponding to preset web;
The second acquisition unit 604, for obtaining the target topic vector of topic model output;
The metric element 605, for the preference according to the target topic vector metric targeted customer.
In embodiments of the present invention, the behavior for obtaining targeted customer accesses vector;Vector will be accessed the behavior to input to pre-
In the topic model first built, accessed so as to obtain the behavior in the topic model that is built in advance from this corresponding to vector
Target topic vector;So as to the preference by the targeted customer recorded in the target topic vector to each theme
Value, the preference of metric objective user, reduces the cumbersome degree of the preference of metric objective user, improves metric objective user preference
Speed.
Alternatively, included in the topic model built as a kind of embodiment of the invention, first construction unit 603
Pre-set user accesses the corresponding relation of vector and theme vector with its behavior, is:The identification information of pre-set user is visited with its behavior
Ask the corresponding relation of vector and theme vector;
The device can also include:3rd acquiring unit, judging unit and the second input block;
3rd acquiring unit, for before accessing vector in the behavior for obtaining targeted customer, obtaining the mark of targeted customer
Know information;
The judging unit, for according to the identification information, judging whether the targeted customer is pre-set user;
If not, trigger the first acquisition unit 601;
If it is, the identification information is inputted into the topic model by second input block, it is defeated to obtain the topic model
The target topic vector gone out;Trigger the metric element 605.
Alternatively, a kind of embodiment as the present invention, first construction unit 603 include:First acquisition module, first
Build module and the second structure module;
First acquisition module, the access frequency of each preset web is accessed for obtaining each pre-set user;
The first structure module, for utilizing each access frequency, build the behavior corresponding to each pre-set user and visit
Ask vector;
The second structure module, for accessing vector, profit based on the number of topics N to be built and constructed each behavior
The topic model is built with preset themes model algorithm;
Wherein, the topic model includes N number of theme, in addition to each behavior access theme corresponding to vector to
Amount, wherein, each preset web is covered in each theme, and record has each preset web each in the topic model
Shared weighted value in individual theme, wherein, each preset web weighted value shared in each theme is incomplete same.
Alternatively, a kind of embodiment as the present invention, the device can also include:Second construction unit;
Second construction unit, for access this each default obtaining each pre-set user using first acquisition module
Before the access frequency of webpage, the text corresponding to each pre-set user is built, wherein, corresponding predetermined use is recorded in text
The preset web that family accessed;Wherein, a word in each preset web text as corresponding in text;
Correspondingly, first acquisition module, including:Calculating sub module;
The calculating sub module, for calculating the appearance of each preset web in the text corresponding to each pre-set user
Frequency, to obtain the access frequency of each preset web in each text, and the text corresponding to by each pre-set user
In the access frequency of each preset web that did not occurred be arranged to zero.
Alternatively, as a kind of embodiment of the invention, recorded in the topic model constructed by first construction unit 603
There is each preset web weighted value shared in each theme, wherein, each preset web institute in each theme
The weighted value accounted for is incomplete same;
The second acquisition unit 604, including:Preference value computation subunit and target topic vector obtain subelement;
The preference value computation subunit, for by shared by each preset web covered in each theme
Weighted value, it is weighted summation meter with corresponding to the access frequency of preset web in the behavior access vector of the targeted customer respectively
Calculate, obtain preference value of the targeted customer to each theme;
Target topic vector obtains subelement, for based on the preference value, obtaining target topic vector.
Alternatively, target topic vector, which obtains the target topic vector that subelement obtains, to be:
(theme 1, theme 2 ..., theme N)=(X1、X2、……、XN), wherein,The Xi represents the target
Preference value of the user to i-th of theme;
Correspondingly, the metric element 605, including:Measure subelement;
The measurement subelement, for utilizing the X1To XNCorresponding value, the targeted customer is measured to the inclined of each theme
It is good.
Alternatively, the preference value computation subunit calculates preference value institute of the targeted customer to each theme
The calculation formula utilized is:
Wherein, the A represents the preset web, and the k represents k-th of theme, the p (Ai) represent that the goal behavior accesses vector
In i-th of preset web access frequency, the p (Ai| themek) represent in the topic model i-th of default net under k-th of theme
Weighted value shared by page, the v represent that the behavior of the targeted customer accesses v-th of preset web in vector, wherein, the i≤v, the k
∈ (1, N).
Alternatively, a kind of embodiment as the present invention, the device also include:Taxon;
Taxon, for the second structure module after the topic model is built using preset themes model algorithm, base
Each theme institute threshold value set in advance, is carried out to the theme vector corresponding to each pre-set user in the topic model
Classification, to be clustered to each pre-set user.
Alternatively, a kind of embodiment as the present invention, the access frequency of each preset web, is each default URL
Or the access frequency of default domain name.
The user preference measurement apparatus includes processor and memory, and the above-mentioned input of first acquisition unit 601, first is single
First 602, first construction unit 603, second acquisition unit 604 and metric element 605 etc. are stored in storage as program unit
In device, corresponding function is realized by the said procedure unit of computing device storage in memory.
Kernel is included in processor, is gone in memory to transfer corresponding program unit by kernel.Kernel can set one
Or more, by adjusting kernel parameter come conveniently and efficiently measure user preference.
Memory may include computer-readable medium in volatile memory, random access memory (RAM) and/
Or the form such as Nonvolatile memory, such as read-only storage (ROM) or flash memory (flash RAM), memory includes at least one deposit
Store up chip.
Present invention also provides a kind of computer program product, when being performed on data processing equipment, is adapted for carrying out just
The program code of beginningization there are as below methods step:
Step 1:The behavior for obtaining targeted customer accesses vector;Wherein, the behavior, which accesses, includes in vector to each default
The access frequency of webpage;
Step 2:The behavior of the targeted customer is accessed into vector input to default topic model;The topic model is pair
The behavior of pre-set user accesses vector and is trained acquisition, wherein contain pre-set user accesses vector and theme with its behavior
The corresponding relation of vector;Include preference of the pre-set user to each theme corresponding to preset web in the theme vector
Value;
Step 3:Obtain the target topic vector of topic model output;
Step 4:According to the preference of the target topic vector metric targeted customer.
It should be understood by those skilled in the art that, embodiments herein can be provided as method, system or computer program
Product.Therefore, the application can use the reality in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Apply the form of example.Moreover, the application can use the computer for wherein including computer usable program code in one or more
The computer program production that usable storage medium is implemented on (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of product.
The application is with reference to the flow according to the method for the embodiment of the present application, equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that can be by every first-class in computer program instructions implementation process figure and/or block diagram
Journey and/or the flow in square frame and flow chart and/or block diagram and/or the combination of square frame.These computer programs can be provided
The processors of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce
A raw machine so that produced by the instruction of computer or the computing device of other programmable data processing devices for real
The device for the function of being specified in present one flow of flow chart or one square frame of multiple flows and/or block diagram or multiple square frames.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory, which produces, to be included referring to
Make the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one square frame of block diagram or
The function of being specified in multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing devices so that counted
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented processing, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in individual square frame or multiple square frames.
In a typical configuration, computing device includes one or more processors (CPU), input/output interface, net
Network interface and internal memory.
Memory may include computer-readable medium in volatile memory, random access memory (RAM) and/
Or the form such as Nonvolatile memory, such as read-only storage (ROM) or flash memory (flash RAM).Memory is computer-readable Jie
The example of matter.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer-readable instruction, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase transition internal memory (PRAM), static RAM (SRAM), moved
State random access memory (DRAM), other kinds of random access memory (RAM), read-only storage (ROM), electric erasable
Programmable read only memory (EEPROM), fast flash memory bank or other memory techniques, read-only optical disc read-only storage (CD-ROM),
Digital versatile disc (DVD) or other optical storages, magnetic cassette tape, the storage of tape magnetic rigid disk or other magnetic storage apparatus
Or any other non-transmission medium, the information that can be accessed by a computing device available for storage.Define, calculate according to herein
Machine computer-readable recording medium does not include temporary computer readable media (transitory media), such as data-signal and carrier wave of modulation.
Embodiments herein is these are only, is not limited to the application.To those skilled in the art,
The application can have various modifications and variations.All any modifications made within spirit herein and principle, equivalent substitution,
Improve etc., it should be included within the scope of claims hereof.
Claims (10)
- A kind of 1. user preference measure, it is characterised in that including:The behavior for obtaining targeted customer accesses vector;Wherein, the behavior is accessed in vector comprising the visit to each preset web Ask frequency;The behavior of the targeted customer is accessed into vector input to default topic model;The topic model includes pre-set user The corresponding relation of vector and theme vector is accessed with its behavior;Include pre-set user in the theme vector to preset web pair The preference value for each theme answered;Obtain the target topic vector of the topic model output;According to the preference of targeted customer described in the target topic vector metric.
- 2. according to the method for claim 1, it is characterised in that access vector in the behavior of the acquisition targeted customer Before, in addition to:Obtain the identification information of targeted customer;According to the identification information, judge whether the targeted customer is pre-set user;If not, perform the step that the behavior for obtaining targeted customer accesses vector;If it is, the identification information is inputted into the topic model, the target topic vector of the topic model output is obtained; Perform described according to the target topic vector metric the step of preference of targeted customer.
- 3. according to the method for claim 1, it is characterised in that the default topic model obtains for structure in advance;The topic model is built in advance to comprise the following steps:Obtain the access frequency that each pre-set user accesses each preset web;Using each access frequency, build the behavior corresponding to each pre-set user and access vector;Vector is accessed based on the number of topics N to be built and constructed each behavior, built using preset themes model algorithm The topic model;Wherein, the topic model includes N number of theme, and with each behavior access theme corresponding to vector to Amount;Wherein, each preset web is covered in each theme, and record has each default net in the topic model Page weighted value shared in each theme, wherein, each preset web weighted value shared in each theme It is incomplete same.
- 4. according to the method for claim 3, it is characterised in that accessed in each pre-set user of acquisition described each pre- If before the access frequency of webpage, methods described also includes:The text corresponding to each pre-set user is built, wherein, record what corresponding predesignated subscriber accessed in text Preset web;Wherein, a word in each preset web text as corresponding in text;Correspondingly, the access frequency for obtaining each pre-set user and accessing each preset web, including:The frequency that each preset web in the text corresponding to each pre-set user occurs is calculated, to obtain each text In each preset web access frequency, it is and each pre- by what is do not occurred in the text corresponding to each pre-set user If the access frequency of webpage is arranged to zero.
- 5. according to the method for claim 3, it is characterised in that the target topic for obtaining topic model output to Amount, including:The weighted value shared by each preset web that will be covered in each theme, respectively with the targeted customer's The access frequency that preset web is corresponded in behavior access vector is weighted read group total, obtains the targeted customer to described each The preference value of individual theme;Based on the preference value, the target topic vector is obtained.
- 6. according to the method for claim 5, it is characterised in that it is described to be based on the preference value, obtain the target Theme vector is:(theme 1, theme 2 ..., theme N)=(X1、X2、……、XN), wherein,The Xi represents the target Preference value of the user to i-th of theme;Correspondingly, the preference of the targeted customer according to the target topic vector metric, including:Utilize the X1To XNCorresponding value, measure preference of the targeted customer to each theme.
- 7. the method according to claim 5 or 6, it is characterised in that it is described will be covered in each theme described in Weighted value shared by each preset web, the access that preset web is corresponded in vector is accessed with the behavior of the targeted customer respectively The calculation formula that frequency is weighted read group total is:Wherein, the A represents the preset web, and the k represents k-th of theme, the p (Ai) represent that the goal behavior is visited Ask the access frequency of i-th of preset web in vector, the p (AiThemek) represent in the topic model i-th under k-th of theme Weighted value shared by individual preset web, the v represent that the behavior of the targeted customer accesses v-th of preset web in vector, its In, the i≤v, the k ∈ (1, N).
- 8. the method according to claim 3 or 4, it is characterised in that utilize preset themes model algorithm structure institute described After stating topic model, methods described also includes:Based on to each theme institute threshold value set in advance in the topic model, to the master corresponding to each pre-set user Topic vector is classified, to be clustered to each pre-set user.
- A kind of 9. user preference measurement apparatus, it is characterised in that including:First acquisition unit, the behavior for obtaining targeted customer access vector;Wherein, the behavior access in vector comprising pair The access frequency of each preset web;First input block, for the behavior of the targeted customer to be accessed into what vector input was built in advance to the first construction unit Topic model;The topic model includes the corresponding relation that pre-set user accesses vector and theme vector with its behavior;The master Include preference value of the pre-set user to each theme corresponding to preset web in topic vector;Second acquisition unit, for obtaining the target topic vector of the topic model output;Metric element, the preference for the targeted customer according to the target topic vector metric.
- 10. device according to claim 9, it is characterised in that described device also includes:3rd acquiring unit, for before accessing vector in the behavior for obtaining targeted customer, obtaining the identification information of targeted customer;Judging unit, for according to the identification information, judging whether the targeted customer is pre-set user;If not, trigger the first acquisition unit;If it is, the identification information is inputted into the topic model by the second input block, it is defeated to obtain the topic model The target topic vector gone out;Trigger the metric element.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610607241.1A CN107665208B (en) | 2016-07-28 | 2016-07-28 | User preference measurement method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610607241.1A CN107665208B (en) | 2016-07-28 | 2016-07-28 | User preference measurement method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107665208A true CN107665208A (en) | 2018-02-06 |
CN107665208B CN107665208B (en) | 2019-12-13 |
Family
ID=61115232
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610607241.1A Active CN107665208B (en) | 2016-07-28 | 2016-07-28 | User preference measurement method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107665208B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108763400A (en) * | 2018-05-22 | 2018-11-06 | 合肥工业大学 | Object partitioning method and device based on object behavior and subject matter preferences |
CN110765387A (en) * | 2019-09-09 | 2020-02-07 | 平安科技(深圳)有限公司 | User interface generation method and device, computing equipment and storage medium |
CN112398751A (en) * | 2020-10-12 | 2021-02-23 | 联通智网科技有限公司 | Flow speed control method and device, computer equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101770520A (en) * | 2010-03-05 | 2010-07-07 | 南京邮电大学 | User interest modeling method based on user browsing behavior |
CN102521248A (en) * | 2011-11-14 | 2012-06-27 | 北京亿赞普网络技术有限公司 | Network user classification method and device |
CN103514239A (en) * | 2012-11-26 | 2014-01-15 | Tcl美国研究所 | Recommendation method and system integrating user behaviors and object content |
CN104679743A (en) * | 2013-11-26 | 2015-06-03 | 阿里巴巴集团控股有限公司 | Method and device for determining preference model of user |
-
2016
- 2016-07-28 CN CN201610607241.1A patent/CN107665208B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101770520A (en) * | 2010-03-05 | 2010-07-07 | 南京邮电大学 | User interest modeling method based on user browsing behavior |
CN102521248A (en) * | 2011-11-14 | 2012-06-27 | 北京亿赞普网络技术有限公司 | Network user classification method and device |
CN103514239A (en) * | 2012-11-26 | 2014-01-15 | Tcl美国研究所 | Recommendation method and system integrating user behaviors and object content |
CN104679743A (en) * | 2013-11-26 | 2015-06-03 | 阿里巴巴集团控股有限公司 | Method and device for determining preference model of user |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108763400A (en) * | 2018-05-22 | 2018-11-06 | 合肥工业大学 | Object partitioning method and device based on object behavior and subject matter preferences |
CN108763400B (en) * | 2018-05-22 | 2021-09-14 | 合肥工业大学 | Object dividing method and device based on object behaviors and theme preferences |
CN110765387A (en) * | 2019-09-09 | 2020-02-07 | 平安科技(深圳)有限公司 | User interface generation method and device, computing equipment and storage medium |
CN112398751A (en) * | 2020-10-12 | 2021-02-23 | 联通智网科技有限公司 | Flow speed control method and device, computer equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN107665208B (en) | 2019-12-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104679743B (en) | A kind of method and device of the preference pattern of determining user | |
CN103455522B (en) | Recommendation method and system of application extension tools | |
CN103810162B (en) | The method and system of recommendation network information | |
CN104426713B (en) | The monitoring method and device of web site access effect data | |
CN106649316A (en) | Video pushing method and device | |
CN107957957A (en) | The acquisition methods and device of test case | |
CN107368519A (en) | A kind of cooperative processing method and system for agreeing with user interest change | |
CN106919611B (en) | Product information pushing method and device | |
CN110008397B (en) | Recommendation model training method and device | |
CN105095279B (en) | File recommendation method and device | |
CN106598999B (en) | Method and device for calculating text theme attribution degree | |
CN107688984A (en) | Product information method for pushing, device, storage medium and computer equipment | |
CN107577687A (en) | Image search method and device | |
CN104102819B (en) | A kind of determination method and apparatus of user's natural quality | |
CN106503025A (en) | Method and system is recommended in a kind of application | |
CN106130756B (en) | A kind of method and device of prediction access content clicking rate | |
EP3639163A1 (en) | Systems and methods for optimizing and simulating webpage ranking and traffic | |
CA3062119A1 (en) | Method and device for setting sample weight, and electronic apparatus | |
CN104965863B (en) | A kind of clustering objects method and apparatus | |
CN108830443A (en) | A kind of contract review method and device | |
CN106920119A (en) | The evaluation method and device of a kind of user's value | |
CN107665208A (en) | User preference measure and device | |
CN106649347A (en) | Interest information identification method and apparatus | |
CN109800364A (en) | Amount of access statistical method, device, equipment and storage medium based on block chain | |
CN104750828B (en) | A kind of unconscious self-learning method of conclusion deductive knowledge based on 6W rules |
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 |