CN108363821A - A kind of information-pushing method, device, terminal device and storage medium - Google Patents
A kind of information-pushing method, device, terminal device and storage medium Download PDFInfo
- Publication number
- CN108363821A CN108363821A CN201810435812.7A CN201810435812A CN108363821A CN 108363821 A CN108363821 A CN 108363821A CN 201810435812 A CN201810435812 A CN 201810435812A CN 108363821 A CN108363821 A CN 108363821A
- Authority
- CN
- China
- Prior art keywords
- user
- tags
- information
- group
- tag
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000003860 storage Methods 0.000 title claims abstract description 19
- 239000013598 vector Substances 0.000 claims abstract description 77
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 230000004931 aggregating effect Effects 0.000 claims abstract description 19
- 238000004458 analytical method Methods 0.000 claims abstract description 10
- 238000001914 filtration Methods 0.000 claims abstract description 9
- 230000006399 behavior Effects 0.000 claims description 48
- 238000004590 computer program Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 description 10
- 230000008569 process Effects 0.000 description 8
- 230000003542 behavioural effect Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 6
- 238000005406 washing Methods 0.000 description 5
- 101100545272 Caenorhabditis elegans zif-1 gene Proteins 0.000 description 4
- 230000003796 beauty Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000012163 sequencing technique Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 241000196324 Embryophyta Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
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)
Abstract
The invention discloses a kind of information-pushing method, device, terminal device and storage mediums.The method includes：Obtain the historical behavior information of user；Analysis filtering is carried out to the historical behavior information, obtains user key words；The user key words are trained by the way of term vector, determine user tag；Classified to the user tag based on K Means aggregating algorithms, obtains original user set of tags；It is browsed and is recorded according to the historical behavior information history of the user, the original user set of tags is ranked up, target user's set of tags is obtained；Based on target user's set of tags, the preference information of the user is obtained；The corresponding service label of the preference information is obtained from preset service label library, and pushes the corresponding business information of the service label to the user.Technical scheme of the present invention effectively increases the intelligent level of business message push and the popularization efficiency of business information.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of information-pushing method, device, terminal device and storages
Medium.
Background technology
With the rapid development of internet science and technology, internet has been deep into huge numbers of families, the work of many people and
Life too busy to get away internet, online have become a part for many people's study, work and life, and many people are daily
Can be done shopping using network, is social, amusement, office and lookup data etc..
The business department of numerous government unit, media and enterprise is in order to meet users' demand, often at oneself
Website on publication meet the magnanimity informations of different user demands, but this is but also the network information becomes numerous and jumbled so that user is every
It is secondary to be required for taking a significant amount of time the commodity for looking for wanting purchase or frequent using research tool search data and interested
Topic, affect the service efficiency of user.
In the prior art, the business department of some government unit, media and enterprise is according to oneself operational characteristic, to
User has pushed some business information, but the intelligent level of existing transmission service information mode is relatively low, does not have specific aim, right
For the user for not needing these business information, these business information then become a kind of covert advertisement, are unfavorable for user's
It uses, also affects the efficiency of business department's promotion business information.
Invention content
A kind of information-pushing method of offer of the embodiment of the present invention, device, terminal device and storage medium, to solve existing skill
The intelligent level of business message push is relatively low in art and business information promotes inefficient problem.
In a first aspect, the embodiment of the present invention provides a kind of information-pushing method, including：
Obtain the historical behavior information of user；
Analysis filtering is carried out to the historical behavior information, obtains user key words；
The user key words are trained by the way of term vector, determine user tag；
Classified to the user tag based on K-Means aggregating algorithms, obtains original user set of tags；
According to the historical behavior information of the user, the original user set of tags is ranked up, target user is obtained
Set of tags；
Based on target user's set of tags, the preference information of the user is obtained；
The corresponding service label of the preference information is obtained from preset service label library, and pushes institute to the user
State the corresponding business information of service label.
Second aspect, the embodiment of the present invention provide a kind of information push-delivery apparatus, including：
Historical behavior data obtaining module, the historical behavior information for obtaining user；
User key words acquisition module obtains user key words for carrying out analysis filtering to the historical behavior information；
User tag generation module is determined and is used for being trained to the user key words by the way of term vector
Family label；
Original user set of tags generation module classifies to the user tag for being based on K-Means aggregating algorithms,
Obtain original user set of tags；
Target user's set of tags generation module, for the historical behavior information according to the user, to the original user
Set of tags is ranked up, and obtains target user's set of tags；
User preference information acquisition module obtains the preference letter of the user for being based on target user's set of tags
Breath；
First business information pushing module is corresponded to for obtaining the preference information from preset service label library
Service label, and push the corresponding business information of the service label to the user.
The third aspect, the embodiment of the present invention provide a kind of terminal device, including memory, processor and are stored in described
In memory and the computer program that can run on the processor, the processor are realized when executing the computer program
The step of described information method for pushing.
Fourth aspect, the embodiment of the present invention provide a kind of computer readable storage medium, the computer-readable storage medium
Matter is stored with computer program, when the computer program is executed by processor the step of realization described information method for pushing.
Information-pushing method provided in an embodiment of the present invention has the following advantages that compared with prior art：The embodiment of the present invention
Information-pushing method, device, terminal device and the storage medium of offer, by collecting user's history behavioural information, and to the use
Family historical behavior information carries out analysis filtering, obtains user key words, is instructed to user key words by way of term vector
Practice, weed out the fuzzyyer keyword of meaning, obtain user tag, user tag is carried out using K-Means aggregating algorithms
Grouping, obtains original user set of tags, is ranked up, obtains to the original user set of tags further according to the historical behavior information of user
It pushes and uses to user based on target user's set of tags and business information in preset service label library to target user's set of tags
The business information of family preference is realized according to the method analyzed the historical behavior of user, and utilize machine learning, in real time
The interest preference of dynamic access user pushes corresponding business information for the different interest preferences of different user, to effectively
The intelligent level of business message push is improved, and then improves the popularization efficiency of different business information.
Description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the present invention
Example, for those of ordinary skill in the art, without having to pay creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the flow chart for the information-pushing method that the embodiment of the present invention 1 provides；
Fig. 2 is the implementation flow chart of step S30 in the information-pushing method that the embodiment of the present invention 1 provides；
Fig. 3 is the implementation flow chart of step S40 in the information-pushing method that the embodiment of the present invention 1 provides；
Fig. 4 is the implementation flow chart of step S50 in the information-pushing method that the embodiment of the present invention 1 provides；
Fig. 5 is the implementation flow chart that community information pushes in the information-pushing method that the embodiment of the present invention 1 provides；
Fig. 6 is to determine user group in the information-pushing method that the embodiment of the present invention 1 provides and generate group's label, Xiang Yong
Family pushes the implementation flow chart of the corresponding business information of group's label；
Fig. 7 is the schematic diagram for the information push-delivery apparatus that the embodiment of the present invention 2 provides；
Fig. 8 is the schematic diagram for the terminal device that the embodiment of the present invention 4 provides.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, the every other implementation that those of ordinary skill in the art are obtained without creative efforts
Example, shall fall within the protection scope of the present invention.
Embodiment 1
Referring to Fig. 1, Fig. 1 shows the implementation process of information-pushing method provided in this embodiment.Details are as follows：
S10：Obtain the historical behavior information of user.
Specifically, server obtains the log information of record user's history behavior from background data base, by daily record
Extraction, obtains the historical behavior information of user.
Wherein, user's history behavior refers to all operationss behavior of the user after Website login platform, the operation behavior quilt
It is recorded in the log information of server background database.
Historical behavior information includes but not limited to：Historical search record, history click record, historical viewings record.
Historical search records：User information, search time and search key.User information includes using
The essential information at family, such as name, gender, age, search time refer to the specific time for detecting search operaqtion, and retrieval is crucial
Word refers to the keyword for retrieval time inputting and being inquired.
For example, in a specific embodiment, historical search is recorded as：" (Zhang San, man, 23), 2018-01-29 20:
46:50, washing machine ", wherein " Zhang San, man, 23 " be user information, " 29 days 20 January in 2018:46:50 " be search time,
" washing machine " is search key.
History clicks record：User information, the status identifier for clicking time and clickable hyperlinks
(Identification, ID), it refers to the specific time for detecting clicking operation to click the time, and the ID of clickable hyperlinks is to give directions
Hit the ID of the object of operation, further, click historical record be also recorded for the product be clicked in search result or
It is clicked in recommendation results.
For example, in a specific embodiment, history click is recorded as：" (Zhang San, man, 23), 2018-01-29 20:
51:50,65936, S ", wherein " 65936 " are the page of corresponding one domestic washing machine in background data base specifically introduced
Hyperlink ID, " S " is the ID that the ID of the hyperlink clicked is the hyperlink obtained according to the mode of search.
Historical viewings record：User information, browsing time and browsing data, wherein browsing is user
In the browsing record of the generation of browsing product information, the page that one click is checked can record a plurality of browsing data, which is
Sampled data.
S20：Analysis filtering is carried out to historical behavior information, obtains user key words.
Specifically, believed by the page corresponding to the ID to search key, clickable hyperlinks in historical behavior information
Breath and the page info of browsing carry out key message extraction and analysis, obtain user's initial key word.
Wherein, the page info corresponding to the ID of clickable hyperlinks refers to the ID of the object of clicking operation in background data base
In corresponding page info.
For example, after user's clickable hyperlinks ID is the hyperlink of " NZ_5263 ", obtained by being inquired in background data base
Know that the page corresponding to hyperlink that hyperlink ID is " NZ_5263 " is " https://miaosha.xxxxxx.com/#
1892018 ", and then the product title of the page is extracted, product essential information etc. obtains initial user keyword：It is " exquisite luxurious
The bright embroidery that ridicules can match in excellence or beauty XXX famous brand name autumn and winter trendy Korea Spro version sweater ".
Further, the interference vocabulary in user's initial key word is rejected, obtains user key words.
Specifically, the realization method rejected to the interference vocabulary in user's initial key word is：
Word segmentation processing is carried out to initial user keyword according to pre-set dictionary, obtains multiple points of initial user keyword
Word；
Word's kinds are carried out according to the part of speech of the participle to each participle, for example, these are divided into master according to part of speech
See vocabulary and objective vocabulary；
The participle for meeting default part of speech condition is obtained as user key words.
For example, in a specific embodiment, the user's initial key word extracted is that " the exquisite luxurious bright embroidery that ridicules can be equal to
Beautiful XXX famous brand names autumn and winter trendy Korea Spro's version sweater " can be by " the exquisite luxurious bright embroidery that ridicules can match in excellence or beauty when carrying out word segmentation processing
The product title of XXX famous brand name autumn and winter trendy Korea Spro version sweater " be divided into " exquisite luxurious ", " bright to ridicule embroidery ", " can match in excellence or beauty ",
" XXX famous brand names ", " autumn and winter ", " trendy ", " Korea Spro version ", " sweater " seven key vocabularies, and " exquisite luxurious ", " bright to ridicule
Embroider ", " can match in excellence or beauty " three words be subjective vocabulary, rejectings can be given, " XXX famous brand names ", " autumn and winter ", " trendy ", " Korea Spro edition " and
" sweater " is the vocabulary for the product feature that can react user demand, then five keywords finally obtained are：" the well-known product of XXX
Board ", " autumn and winter ", " trendy ", " Korea Spro's version " and " sweater ".
S30：User key words are trained by the way of term vector, determine user tag.
In artificial intelligence, language expression refers mainly to the formalization of language or the description of mathematics, so as to table in a computer
Show language, and computer program can be allowed to automatically process.In the embodiment of the present invention signified term vector be exactly with vector form come
Indicate a keyword.
Specifically, all user key words are trained by using the mode of term vector, obtain meeting preset requirement
User tag.
In a specific embodiment, according to default corpus, the basic term vector of each user key words is built, for
Each basis term vector calculates the space length between the basis term vector and other basis term vectors, obtain the basis word to
The minimum space distance of amount, and then the basic term vector for being less than or equal to pre-set space distance threshold in minimum space distance is made
For user tag.So that when generating user tag, the user key words that user seldom pays close attention to are filtered out, and then can be more
Add and accurately determines user preference.
S40：Classified to user tag based on K-Means aggregating algorithms, obtains original user set of tags.
Specifically, user tag is polymerize using K-Means aggregating algorithms, the high user tag of the degree of polymerization is put into
The same classification obtains different classification, these classification are original user set of tags.
K-means algorithms are the clustering algorithms based on distance, the evaluation index using distance as similitude thinks two
The distance of a object is closer, and similarity is bigger.The algorithm thinks cluster by forming apart from close object, therefore
To compact and independent cluster as final goal.
S50：According to the historical behavior information of user, original user set of tags is ranked up, obtains target user's label
Group.
Specifically, by step S1 it is found that the historical behavior information of user includes the time that the historical behavior occurs, according to going through
The time sequencing that history behavioural information occurs, is ranked up the corresponding user tag of historical behavior information, and then is tied according to sequence
Fruit is ranked up original user set of tags, obtains target user's set of tags.
S60：Based on target user's set of tags, the preference information of user is obtained.
Specifically, the target user's set of tags generated according to step S50, according to the push kind of preset business information
Class, extraction includes target user's set of tags of the push type from target user's set of tags, and therefrom filters out sequence preceding
Preset quantity target user's set of tags, and then it is corresponding according to each user tag in the target user's set of tags filtered out
User key words obtain the user preference of corresponding push classification.
For example, in a specific embodiment, no more than three classes, i.e., at most preset business information push type is
Three classes user interest preference is obtained, according to the sequence of target user's set of tags, chooses first three groups target user's set of tags in order
Corresponding three groups of user key words, and obtain the preference information of active user according to three groups of user key words.
S70：The corresponding service label of preference information of user is obtained from preset service label library, and is pushed away to the user
Give the service label corresponding business information.
Specifically, according to the preference information of user, corresponding service label is chosen from preset service label library,
And then the corresponding business information of service label is obtained, it is excellent according to the user preference of each obtained push classification of step S60
The corresponding business information of user preference of first category is first pushed, the business consultation information is pushed if receiving user and closing
Request or the feedback of user is not received by preset time, then the user preference that second category is pushed to user corresponds to
Business consultation information, terminate to push when reaching default push times.
In the corresponding embodiments of Fig. 1, by collect user's history behavioural information, and to the user's history behavioural information into
Row analysis filtering, obtains user key words, by being trained to user key words, weeds out the fuzzyyer key of meaning
Word obtains user tag, is grouped to user tag using K-Means aggregating algorithms, obtains original user set of tags, then right
The original user set of tags is ranked up, and obtains target user's set of tags, is based on target user's set of tags and preset business mark
The business information in library is signed, the business information of user preference is pushed to user, is realized according to the history to user
Behavior is analyzed, and the method for utilizing machine learning, the interest preference of real-time dynamic access user, not for different user
Same interest preference pushes corresponding business information, to effectively increase the intelligent level of business message push, and then improves
The popularization efficiency of different business information.
Next, on the basis of the corresponding embodiments of Fig. 1, come to step below by a specific embodiment
Being trained to user key words by the way of term vector mentioned in S30 determines the concrete methods of realizing of user tag
It is described in detail.
Referring to Fig. 2, Fig. 2 shows the specific implementation flow of step S30 provided in an embodiment of the present invention, details are as follows：
S31：Based on default corpus, the basic term vector of each user key words is built.
Specifically, the keyword in user's history behavioural information is mapped to according to preset corpus in a vector,
These vectors are linked together, form a term vector space, each vector is the equal of a point in this space.
For example, have BMW inside certain sale of automobile Products title, run quickly the two keywords, according to preset language material
Library obtains the two keywords and is possible to classify：" automobile ", " luxury goods ", " animal ", " action " and " cuisines ".Cause
The two keywords are introduced a kind of vector expression by this：
<Automobile, luxury goods, animal, action, cuisines>
The probability that the two keywords belong to each classification, the possibility that computer arrives are calculated according to the method for statistical learning
It is：
BMW=<0.5,0.2,0.2,0.0,0.1>
Benz=<0.7,0.2,0.0,0.1,0.0>
It is to be appreciated that basic term vector represents the semanteme and can grammatically solve that one has centainly per one-dimensional value
The feature released, therefore basic term vector can be known as a key characteristics per one-dimensional.
Further, keyword term vector is built for each keyword of user, obtains basic term vector.
It should be noted that each user key words correspond to unique basic term vector, each basis term vector correspond to
Few user key words.
By being based on default corpus, the basic term vector of each user key words is built so that can not be accurate by machine
The text conversion of understanding easily identifies and carries out the term vector of operation at machine, is conducive to accurately identify user preference
S32：For each basic term vector, calculate space between the basis term vector and other basis term vectors away from
From, and minimum space distance of the minimum value as the basis term vector is chosen from space length.
Specifically, for each basic term vector, the calculation formula of use space distance calculates separately the basis term vector
With the space length between other all basic term vectors, and the minimum value of these space lengths is found out.
Basis term vector A (a are calculated according to formula (1)1,a2,...,an) and basis term vector B (b1,b2,...,bn) between
Space length L：
Wherein, n is the positive integer more than or equal to 2.
For example, in a specific embodiment, basic term vector includes G1(0.9,0.1)、G2(0.5,0.5)G3(0.8,
0.2), for G1, G is calculated separately according to formula (1)1To G2Space length be 0.5659 and G1To G3Space length be
0.1414, then G1Minimum space distance be 0.1414.
S33：The basic term vector of pre-set space distance threshold will be less than or equal in minimum space distance, is marked as user
Label.
Specifically, minimum empty to these after the minimum space distance that each basic term vector is calculated according to step S32
Between distance be compared with preset word space threshold, by minimum space distance be less than or equal to word space threshold basis
Term vector is as user tag.
By being filtered to the basic term vector for not meeting word space threshold requirement, avoid user's attention rate is low
Content be also placed in user tag, so as to more accurately determine user preference.
For example, in a specific embodiment, preset word space threshold is 0.8, and basic term vector includes H1(0.9,
0.1,0)、H2(0.8,0.1,0.1) and H3H is calculated by the formula (1) in step S32 in (0,0.1,0.9)1Minimum it is empty
Between distance be 0.4243, H2Minimum space distance be 0.4243, H3Minimum space distance be 1.1314, H1And H2Minimum
Space length is less than preset word space threshold 0.8, therefore, by H1And H2As user tag.
In the corresponding embodiments of Fig. 2, based on default corpus, the basic term vector of each user key words is built, and
For each basic term vector, calculate the space length between the basis term vector and other basis term vectors, and from space away from
With a distance from minimum space from middle selection minimum value as the basis term vector, default sky will be less than or equal in minimum space distance
Between the basic term vector of distance threshold can identify operation by the way that user key words are converted into machine as user tag
Term vector, and filter out fuzzy keyword or other lower term vectors of term vector similarity, realization pair according to preset condition
User preference accurately identifies, and effectively increases the intelligent level of business message push.
On the basis of the corresponding embodiments of Fig. 1, below by a specific embodiment come to being carried in step S40
And classified to user tag based on K-Means aggregating algorithms, obtain the concrete methods of realizing of original user set of tags into
Row is described in detail.
Referring to Fig. 3, Fig. 3 shows the specific implementation flow of step S40 provided in an embodiment of the present invention, details are as follows：
S41：From n user tag A1,A2,A3,...,AnIn randomly select m user tag as cluster centre,
In, n and m are positive integer, and m is less than or equal to n.
Specifically, n is the user tag sum of user, and m is according to the preset cluster centre number of needs, from the user's
M user tag is randomly selected in n user tag, using this m user tag as original cluster centre.
S42：For each user tag, the first distance between the user tag and current each cluster centre is calculated,
The user tag is put into minimum first in the cluster where corresponding cluster centre, obtains m interim clusters.
Specifically, it calculates the user tag using the formula (1) in step S32 for each user tag and gathers with each
Space length between class center obtains m the first distances, and minimum is obtained from the m first distance as the first distance
Value is put into apart from corresponding cluster centre in the same cluster as small first distance is done, by the user tag with minimum first,
According to this method, m interim clusters are obtained.
For example, in a specific embodiment, there are 8 user tags, preset cluster centre number is 3, with
After machine generates three cluster centres, it is poly- to three respectively that each user tag is calculated according to the formula (1) in step S32
First distance at class center is as shown in Table 1：
Table one
| Cluster centre 1 | Cluster centre 2 | Cluster centre 3 | |
| User tag 1 | 0.5 | 1.9 | 0.7 |
| User tag 2 | 2.5 | 0.2 | 0.9 |
| User tag 3 | 1.3 | 0.1 | 0.8 |
| User tag 4 | 1.6 | 0.1 | 0.7 |
| User tag 5 | 1.8 | 0.9 | 0.2 |
| User tag 6 | 0.6 | 0.8 | 1.6 |
| User tag 7 | 0.7 | 0.8 | 0.2 |
| User tag 8 | 1.1 | 0.3 | 0.9 |
According to these calculated first distances, it is easy to obtain each user tag to three cluster centres minimum first
User tag 1 for example, first distance of minimum of 1 to three cluster centres of user tag is 0.5, therefore is put into cluster by distance
In cluster where center 1, in this manner, cluster is respectively three obtained temporarily：Interim cluster 1 (user tag 1,
User tag 6), temporarily cluster 2 (user tag 2, user tag 3, user tags 4, user tag 8) and temporarily cluster 3 (users
Label 5, user tag 7).
S43：For each interim cluster, each user's mark in the mean value clustered and the interim cluster temporarily is calculated
Second distance between label and mean value chooses the corresponding user tag of minimum second distance as the new cluster clustered temporarily
Center obtains updated m interim clusters.
Specifically, the mean value clustered temporarily is calculated by formula (2)：
Wherein,For the mean value clustered temporarily, k is the number of user tag in the interim cluster, aiFor in the interim cluster
I-th of user tag, i ∈ [1, k].
Using the formula (1) in step S32, each user tag and the mean value clustered temporarily in the cluster centre are calculated
Between space length, i.e. second distance.Choose the corresponding user tag of minimum second distance as this cluster temporarily it is new
Cluster centre obtains updated m interim clusters.
It should be noted that according to the new cluster centre clustered temporarily, obtain updated m cluster temporarily it is specific
Realization process is identical with the processing procedure of step S42, and to avoid repeating, details are not described herein again.
S44：Each updated standard deviation clustered temporarily is calculated according to formula (3)：
Wherein, σ is standard deviation, and μ is user tag AiThe average value clustered temporarily at place, i ∈ [1, n].
S45：It is preset if at least there is a standard deviation in the m updated standard deviations clustered temporarily and be more than or equal to
Standard deviation threshold method, then return to step S42.
Specifically, by calculated each updated interim cluster standard deviation in step S44 and preset standard
Poor threshold value is compared, when there is interim cluster of the interim cluster standard deviation more than or equal to preset standard deviation threshold method,
Illustrate that the updated interim cluster not yet meets the similarity requirement of user tag, then return to step S42, according to step S42
Processing procedure to step S44 continues to cluster.
S46：If the m updated standard deviations clustered are respectively less than preset standard deviation threshold method temporarily, this m is updated
Interim cluster afterwards is used as original user set of tags.
Specifically, when each updated standard deviation clustered temporarily is both less than preset standard deviation threshold method, illustrate this
Updated interim cluster has been able to the similarity requirement for meeting user tag, using m updated interim clusters as original
Beginning user tag group.
In the corresponding embodiments of Fig. 3, by from randomly selected in n user tag m user tag be used as cluster in
The heart, and each user tag is calculated at a distance from this m cluster centre according to formula (1), it is poly- to find out user tag distance m
First minimum range at class center, so by the user tag and the first minimum range to cluster centre be put into it is same poly-
Class obtains m interim clusters in this approach, is then directed to each interim cluster, calculates what this was clustered temporarily according to formula (2)
Average value, and using the user tag minimum with average value distance as new cluster centre, obtain updated interim cluster, press
The updated standard deviation clustered temporarily is calculated according to formula (3), is more than if there is the updated standard deviation clustered temporarily
Or continue to cluster as stated above again then according to new cluster centre equal to preset standard difference threshold value, until it is all more
The standard deviation clustered temporarily after new terminates to cluster when being respectively less than the standard deviation of preset standard difference threshold value, and current m are faced
When cluster be used as original user set of tags.Classified to user tag by using K-Means aggregating algorithms so that divide originally
Scattered user tag can be clustered according to the similarity between user tag, realized the classification to user tag, effectively carried
The high accuracy of classification, to will be clustered in the same classification with the user of identical preference, to according to
When user tag is to user's transmission service information, the interest preference that can be directed to user accurately pushes relevant business money
News, improve the efficiency of the intelligent level and business department's promotion business information of business message push.
On the basis of the corresponding embodiments of Fig. 1, below by a specific embodiment come to being carried in step S50
And be directed to each user group, to the initial population, set of tags is ranked up, and obtains the corresponding group's label of the user group
The concrete methods of realizing of group is described in detail.
Referring to Fig. 4, Fig. 4 shows the specific implementation flow of step S50 provided in an embodiment of the present invention, details are as follows：
S51：Based on historical behavior information, the generated time of the corresponding historical behavior information of user tag is obtained.
Specifically, based on illustrating in step S10, historical behavior information includes but not limited to：Historical search record,
History clicks record and historical viewings record.Wherein historical search record includes search time, and it includes to click that history, which clicks record,
Time, historical viewings record includes the browsing time, and therefore, every user's history behavioural information is corresponding with generated time.
Further, according to the corresponding historical behavior information of user tag, you can determine the corresponding generation of the user tag
Time.
For example, in a specific embodiment, user tag is<0.6,0.15,0.25>, corresponding user key words
For " washing machine ", the corresponding user's history behavioural information of user key words " washing machine " includes that " history clicks record：(Zhang San,
Man, 23), 2018-01-29 20:51:50,65936, S ", it is readily appreciated that ground, the user tag generated time are " 2018-01-
2920:51:50”。
S52：User tag is ranked up according to generated time, obtains user tag sequence.
Specifically, after the generated time for getting each user tag, according to the sequencing of generated time, to user
Label is ranked up, and obtains user tag sequence.
For example, user tag A corresponding generated times are " 29 days 15 January in 2018:06:38 ", user tag B is corresponding
Generated time is：" 29 days 15 January in 2018:23:54 ", user tag C corresponding generated time is：" on January 26th, 2018
15:07:14 ", be according to the user tag sequence obtained after the sequencing of generated time：" user tag C, user tag A,
User tag B ".
S53：According to user tag sequence, original user set of tags is ranked up, obtains target user's set of tags.
Specifically, it according to the user tag sequence obtained in step S52, obtains and sorts in each original user set of tags
One user tag, first user tag that these are sorted in the original tag group of place are ranked up, and obtain original user
Original user set of tags is sequentially ranked up according to this, obtains target user's set of tags by the sequence of set of tags.
For example, in a specific embodiment, there are 3 groups of original user set of tags, respectively original user set of tags A
(user tag 1, user tag 2, user tag 3), original user set of tags B (user tag 4, user tag 5) and original use
Family set of tags C (user tag 6, user tag 7), wherein the sequence of user tag is " user tag 5, user tag 2, user
Label 1, user tag 7, user tag 4, user tag 6, user tag 3 ", by the sequence it is found that in original user set of tags A
The user tag of sequence first be " user tag 2 ", the user tag of the sequence first in original user set of tags B is " to use
The user tag of family label 5 ", the sequence first in original user set of tags C is " user tag 7 ", by " user tag
2 ", " user tag 5 " and " user tag 7 " is ranked up, and obtains the clooating sequence of original user set of tags, is arranged according to the sequence
Target user's set of tags that sequence obtains is：" target user's set of tags B, target user's set of tags A, target user's set of tags C ".
In the corresponding embodiments of Fig. 4, according to historical behavior information, the corresponding historical behavior information of user tag is obtained
Generated time, and user tag is ranked up according to generated time, user tag sequence is obtained, and then according to user tag sequence
Row, are ranked up original user set of tags, obtain target user's set of tags.So that the preference of the user extracted according to
The time sequencing that family is paid close attention to recently carries out priority sequence, can preferentially elect user nearest when carrying out the push of business information
The preference information of concern, improves the intelligent level of business message push.
On the basis of the corresponding embodiments of Fig. 1, step S30 refer to by the way of term vector to user key
Word is trained, and after determining user tag, can also be further determined that user group and be generated group's label, is pushed to user
The corresponding business information of group's label, as shown in figure 5, the information-pushing method further includes：
S81：According to preset client's tag library, determines different user groups and its corresponding user tag, obtain group
Body label.
Specifically, in preset client's tag library, include the population characteristic information of different user group, according to user's
The user tag of the user by the user attaching in the corresponding user group of the population characteristic, and then is put into this by population characteristic
In group's label, different user group and the corresponding group's label of each user group are obtained.
It should be noted that according to preset population characteristic information, each user belongs at least one user group.
The mode classification of different user groups can be arranged as required in preset client's tag library, for example,
In one specific implementation mode, the user group of client's tag library includes：User group A (man, 18-25 Sui), user group B (it is male,
26-45 Sui) and user group C (man, 46 years old or more), wherein " man " is sex character, and " 18-25 Sui " is age characteristics, another
In one specific implementation mode, the user group of client's tag library includes：User group D (man, training) and user group E (man, sheet
Section), wherein " training " is education degree feature.And then according to the essential information of user, determine user's owning user group, it will
The user tag of each user is put among group's label.
S82：For each user group, is classified to group's label based on K-Means aggregating algorithms, obtain the user
The corresponding initial population's set of tags of group.
Specifically, for each user group, group's label is clustered using K-Means aggregating algorithms, will be polymerize
It spends high group's label and is put into the same classification, obtain different classification, these classification are initial population's set of tags.
It should be noted that the realization process classified to group's label using K-Means aggregating algorithms and step S41
The method classified to step S46 to user tag is identical, and to avoid repeating, details are not described herein again.
S83：For each user group, the corresponding initial population's set of tags of the user group is ranked up, is somebody's turn to do
The corresponding group's set of tags of user group.
Specifically, it for the corresponding initial population's set of tags of each user group, counts in each initial population's set of tags
The number that user tag occurs, is ranked up initial population's set of tags by there is the descending sequence of total degree, is somebody's turn to do
The corresponding group's set of tags of user group so that the high content of group's totality attention rate can be pushed preferentially, and industry is improved
The efficiency of business consultation information push.
S84：Determine the targeted user population of user.
Specifically, it after the request for receiving user's access website, is referred to step S81 according to the essential information of user
Client's tag library in population characteristic information, determine the user group belonging to the user, the i.e. targeted user population of the user.
It is to be appreciated that it may be multiple that the targeted user population of user, which can be one,.
For example, in a specific embodiment, the essential information of user is " man, 22 years old, undergraduate course ", including these features
There are two user groups, respectively：User group G (man, 18-25 Sui) and user group K (man, undergraduate course), therefore, the user's
Targeted user population is：User group G and user group K.
S85：According to each user group and its corresponding group's set of tags, the corresponding target of the targeted user population is obtained
Group's set of tags.
Specifically, in step S83, its corresponding group's set of tags is generated for each user group, in determination
After the targeted user population of user, the group label corresponding to all user groups for including in targeted user population is obtained
Group, target group's set of tags as the user.
S86：Based on target group's set of tags, group's preference information is obtained.
Specifically, the target group's set of tags obtained according to step S85, according to the push kind of preset business information
Class, extraction includes target group's set of tags of the push type from target group's set of tags, and therefrom selected and sorted is preceding
Target group's set of tags of preset quantity, and then according to the corresponding use of each group label in the target group's set of tags filtered out
Family keyword obtains the user preference of corresponding push classification.
S87：The corresponding service label of group's preference information of user is obtained from preset service label library, and to the use
Family pushes the corresponding business information of the service label.
Specifically, it for the fewer new user of historical behavior information, or receives closing in step S70 and pushes the industry
Be engaged in consultation information request old user, the corresponding business information of group preference can be recommended to these users, specifically from
The corresponding service label of group's preference information of user is obtained in preset service label library, and pushes the business mark to the user
The realization process for signing corresponding business information is identical as the implementation method in step S70, no longer superfluous herein to avoid repeating
It states.
It is to be appreciated that the preference information or group that select push user according to preset pushing condition are inclined
The corresponding business consultation information of good information, the preset pushing condition can be configured according to the needs of practical application, herein
It is not limited.
In the corresponding embodiments of Fig. 5, by according to preset client's tag library, determining different user group and its right
The user tag answered obtains group's label, and is directed to each user group, based on K-Means aggregating algorithms to group's label
Classify, obtains the corresponding initial population's set of tags of the user group, and be ranked up to initial population's set of tags, obtain
The corresponding group's set of tags of the user group, it is true according to the essential information of user after receiving the access site requests of user
Determine the targeted user population of user, and then determines the targeted user population of user and corresponding targeted user population set of tags, from
And group's preference of the user is obtained, and can flexibly be selected according to the preference information or group's preference information of user different
Business consultation information, is pushed to the user, and improves the intelligent level of business message push, and then improves different business
The popularization efficiency of information.
On the basis of the corresponding embodiments of Fig. 5, below by a specific embodiment come to being carried in step S83
And be directed to each user group, the corresponding initial population's set of tags of the user group is ranked up, the user group is obtained
Corresponding group's set of tags concrete methods of realizing is described in detail.
Referring to Fig. 6, Fig. 6 shows the specific implementation flow of step S83 provided in an embodiment of the present invention, details are as follows：
S831：For each user group, the corresponding initial population's set of tags B of the user group is obtained1,B2,...,BtIn
The user tag A of each user1,A2,A3,...,ApUser tag word frequency within a preset time intervalWherein, t and p is positive integer.
Specifically, for each user group, each user in the corresponding initial population's set of tags of the user group is obtained
All user tags, prefixed time interval as needed, and to all users of each user in the prefixed time interval
Label carries out the calculating of word frequency.
Wherein, it is corresponding all to refer to that number that the corresponding keyword of the user tag occurs accounts for all user tags for word frequency
The ratio of keyword.
S832：By initial population set of tags BjIn each user tag word frequency be added, obtain group set of tags BjGroup
Body label word frequency Qbj, wherein j ∈ [1, t].
Specifically, for each initial population's set of tags, all user tags inside initial population's set of tags are counted
The sum of word frequency, as initial population's label word frequency.
S833：According to group's label word frequencySize initial population's set of tags is ranked up, obtain
To group's set of tags.
Specifically, group's label word frequency of all initial population's set of tags is ranked up according to the size of the value of word frequency,
The value of group's label word frequency is more than initial population's set of tags of preset word frequency threshold value as group's set of tags.
In the corresponding embodiments of Fig. 6, for each user group, each user in initial population's set of tags is obtained
The word frequency of keyword corresponding to all user tags, calculates group's label word frequency of initial population's set of tags, and according to
The size order of group's label word frequency, is ranked up initial population's set of tags, obtains target group's set of tags so that Mei Gequn
The content that user's attention rate of body is high is come the preferential push in front, improves the efficiency of business department's promotion business information.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Embodiment 2
Corresponding to the information-pushing method in embodiment 1, Fig. 7 shows the information-pushing method one provided with embodiment 1
One corresponding information push-delivery apparatus illustrates only and the relevant part of the embodiment of the present invention for convenience of description.
As shown in fig. 7, the information push-delivery apparatus includes：Historical behavior data obtaining module 10, user key words obtain mould
Block 20, original user set of tags generation module 40, target user's set of tags generation module 50, is used user tag generation module 30
Family preference information acquisition module 60 and business information pushing module 70.Detailed description are as follows for each function module：
Historical behavior data obtaining module 10, the historical behavior information for obtaining user；
User's key message acquisition module 20 obtains user key words for carrying out analysis filtering to historical behavior information；
User tag generation module 30 determines user for being trained to user key words by the way of term vector
Label；
Original user set of tags generation module 40 is classified to user tag for being based on K-Means aggregating algorithms, is obtained
To original user set of tags；
Target user's set of tags generation module 50, for the historical behavior information according to user, to original user set of tags
It is ranked up, obtains target user's set of tags；
User preference information acquisition module 60 obtains the preference information of user for being based on target user's set of tags；
First business information pushing module 70, it is corresponding for obtaining preference information from preset service label library
Service label, and to the corresponding business information of user's transmission service label.
Further, user tag generation module 30 includes：
Basic term vector acquiring unit 31, for based on default corpus, build the basic words of each user key words to
Amount；
Minimum space distance acquiring unit 32, for for each basic term vector, calculating the basis term vector and other
Space length between basic term vector, and from space length choose minimum value as the basis term vector minimum space away from
From；
User tag generation unit 33, for the base of pre-set space distance threshold will to be less than or equal in minimum space distance
Plinth term vector, as user tag.
Further, original user set of tags generation module 40 includes：
Cluster centre acquiring unit 41 is used for from n user tag A1,A2,A3,...,AnIn randomly select m user mark
Label are used as cluster centre, wherein n and m is positive integer, and m is less than or equal to n；
Interim cluster acquiring unit 42 calculates the user tag and current each cluster for being directed to each user tag
The user tag is put into minimum first in the cluster where corresponding cluster centre, obtained by the first distance between center
To m interim clusters；
Interim cluster updating unit 43, for for each interim cluster, calculating the mean value clustered and this facing temporarily
When cluster in each second distance between user tag and mean value, choose the corresponding user tag of minimum second distance as should
The new cluster centre clustered temporarily obtains updated m interim clusters；
Standard deviation computing unit 44, for calculating each updated standard deviation clustered according to following formula temporarily：
Wherein, σ is standard deviation, and μ is user tag AiThe updated average value clustered temporarily at place, i ∈ [1, n]；
Cluster cell 45 is recycled, if at least there is a standard deviation in the m updated standard deviations clustered temporarily
More than or equal to preset standard deviation threshold method, then return to execution and be directed to each user tag, calculate the user tag with it is current
The user tag is put into minimum first apart from poly- where corresponding cluster centre by the first distance between each cluster centre
In class, m interim the step of clustering are obtained；
Original user set of tags generation unit 46, if being respectively less than standard for the m updated standard deviations clustered temporarily
Poor threshold value, then using the m updated interim clusters as original user set of tags.
Further, target user's set of tags generation module 50 includes：
Label generated time acquiring unit 51 obtains the corresponding history row of user tag for being based on historical behavior information
For the generated time of information；
User tag sequence generating unit 52 obtains user's mark for being ranked up to user tag according to generated time
Sign sequence；
Target user's set of tags generation unit 53, for according to user tag sequence, arranging original user set of tags
Sequence obtains target user's set of tags.
Further, which further includes：
Group's label acquiring unit 81, for according to preset client's tag library, determining different user group and its right
The user tag answered obtains group's label；
Initial population's set of tags generation unit 82, for being directed to each user group, based on K-Means aggregating algorithms to this
Group's label is classified, and the corresponding initial population's set of tags of user group is obtained；
Group's set of tags generation unit 83 is ranked up initial population's set of tags, obtains for being directed to each user group
To the corresponding group's set of tags of user group；
Targeted user population determination unit 84, the targeted user population for determining user；
Target group's set of tags acquiring unit 85, for according to each user group and its corresponding target group's label
Group obtains the corresponding target group's set of tags of targeted user population；
Group's preference information acquiring unit 86 obtains group's preference information for being based on target group's set of tags；
Second business information push unit 87, for obtaining group's preference information pair from preset service label library
The service label answered, and to the corresponding business information of user's transmission service label.
Further, group's set of tags generation unit 83 includes：
User tag word frequency obtains subelement 831, for being directed to each user group, obtains initial population set of tags B1,
B2,...,BtIn each user user tag A1,A2,A3,...,ApUser tag word frequency within a preset time intervalWherein, t and p is positive integer；
Group's label word frequency obtains subelement 832, is used for initial population set of tags BjIn each user tag word
Frequency is added, and obtains initial population set of tags BjGroup label word frequency Qbj, wherein j ∈ [1, t]；
Group's set of tags generates subelement 833, for according to group's label word frequencySize to original
Beginning group's set of tags is ranked up, and obtains target group's set of tags.
Each module realizes the process of respective function in a kind of information push-delivery apparatus provided in this embodiment, specifically refers to reality
The description of example 1 is applied, details are not described herein again.
Embodiment 3
The present embodiment provides a computer readable storage medium, computer journey is stored on the computer readable storage medium
Sequence realizes information-pushing method in embodiment 1, alternatively, the computer program is located when the computer program is executed by processor
Manage the function that each module/unit in information push-delivery apparatus in embodiment 2 is realized when device executes.It is no longer superfluous here to avoid repeating
It states.
It is to be appreciated that the computer readable storage medium may include：The computer program code can be carried
Any entity or device, recording medium, USB flash disk, mobile hard disk, magnetic disc, CD, computer storage, read-only memory
(Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), electric carrier signal and
Telecommunication signal etc..
Embodiment 4
Fig. 8 is the schematic diagram for the terminal device that one embodiment of the invention provides.As shown in figure 8, the terminal of the embodiment is set
Standby 90 include：Processor 91, memory 92 and it is stored in the computer journey that can be run in memory 92 and on processor 91
Sequence 93, for example, information-pushing method program.Processor 91 realizes the information in above-described embodiment 1 when executing computer program 93
The step of method for pushing, such as step S1 shown in FIG. 1 to step S7.Alternatively, reality when processor 91 executes computer program 93
The function of each module/unit in existing above-mentioned each device embodiment, such as module 10 shown in Fig. 7 is to the function of module 70.
Illustratively, computer program 93 can be divided into one or more module/units, one or more mould
Block/unit is stored in memory 92, and is executed by processor 91, to complete the present invention.One or more module/units can
To be the series of computation machine program instruction section that can complete specific function, the instruction segment is for describing computer program 93 at end
Implementation procedure in end equipment 90.For example, computer program 93 can be divided into historical behavior data obtaining module, Yong Huguan
Key word acquisition module, user tag generation module, original user set of tags generation module, target user's set of tags generation module,
User preference information acquisition module and business information pushing module.The concrete function of each module is as described in Example 2, to keep away
Exempt to repeat, which is not described herein again.
Terminal device 90 can be the computing devices such as desktop PC, notebook, palm PC and cloud server.Eventually
End equipment 90 may include, but be not limited only to, processor 91, memory 92.It will be understood by those skilled in the art that Fig. 8 is only
The example of terminal device 90 does not constitute the restriction to terminal device 90, may include components more more or fewer than diagram, or
Person combines certain components or different components, such as terminal device 90 can also be set including input-output equipment, network insertion
Standby, bus etc..
Alleged processor 91 can be central processing unit (Central Processing Unit, CPU), can also be
Other general processors, digital signal processor (Digital Signal Processor, DSP), application-specific integrated circuit
(Application Specific Integrated Circuit, ASIC), ready-made programmable gate array (Field-
Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or transistor logic,
Discrete hardware components etc..
Memory 92 can be the internal storage unit of terminal device 90, such as the hard disk or memory of terminal device 90.It deposits
Reservoir 92 can also be the plug-in type hard disk being equipped on the External memory equipment of terminal device 90, such as terminal device 90, intelligence
Storage card intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card
(Flash Card) etc..Further, memory 92 can also both include terminal device 90 internal storage unit and also including outside
Portion's storage device.Memory 92 is used to store other programs and data needed for computer program and terminal device 90.Storage
Device 92 can be also used for temporarily storing the data that has exported or will export.
It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each work(
Can unit, module division progress for example, in practical application, can be as needed and by above-mentioned function distribution by different
Functional unit, module are completed, i.e., the internal structure of described device are divided into different functional units or module, more than completion
The all or part of function of description.
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations；Although with reference to aforementioned reality
Applying example, invention is explained in detail, it will be understood by those of ordinary skill in the art that：It still can be to aforementioned each
Technical solution recorded in embodiment is modified or equivalent replacement of some of the technical features；And these modification or
Person replaces, and the spirit and scope for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution should all wrap
Containing within protection scope of the present invention.
Claims (10)
1. a kind of information-pushing method, which is characterized in that described information method for pushing includes：
Obtain the historical behavior information of user；
Analysis filtering is carried out to the historical behavior information, obtains user key words；
The user key words are trained by the way of term vector, determine user tag；
Classified to the user tag based on K-Means aggregating algorithms, obtains original user set of tags；
According to the historical behavior information of the user, the original user set of tags is ranked up, target user's label is obtained
Group；
Based on target user's set of tags, the preference information of the user is obtained；
The corresponding service label of the preference information is obtained from preset service label library, and pushes the industry to the user
The corresponding business information of business label.
2. information-pushing method as described in claim 1, which is characterized in that it is described by the way of term vector to the user
Keyword is trained, and determines that user tag includes：
Based on default corpus, the basic term vector of each user key words is built；
For each basic term vector, the space length between the basis term vector and other basic term vectors is calculated, and
Minimum space distance of the minimum value as the basis term vector is chosen from the space length；
The basic term vector of pre-set space distance threshold will be less than or equal in minimum space distance, as user tag.
3. information-pushing method as described in claim 1, which is characterized in that the K-Means aggregating algorithms that are based on are to described
User tag is classified, and is obtained original user set of tags and is included：
From n user tag A1,A2,A3,...,AnIn randomly select m user tag as cluster centre, wherein the n and
The m is positive integer, and the m is less than or equal to the n；
For each user tag, the first distance between the user tag and current each cluster centre is calculated,
The user tag is put into minimum first in the cluster where corresponding cluster centre, obtains m interim clusters；
For each interim cluster, calculate in the mean value clustered and the interim cluster temporarily each user tag and
Second distance between the mean value chooses the corresponding user tag of minimum second distance as the new cluster clustered temporarily
Center obtains updated m interim clusters；
Each updated standard deviation clustered temporarily is calculated according to following formula：
Wherein, σ is the standard deviation, and μ is user tag AiThe updated average value clustered temporarily at place, i ∈ [1, n]；
If at least there is a standard deviation in the m updated standard deviations clustered temporarily is more than or equal to preset mark
Quasi- difference threshold value then returns described in execution for each user tag, calculates the user tag and current each cluster
The user tag is put into minimum first in the cluster where corresponding cluster centre, obtained by the first distance between center
The step of to m interim clusters；
If the m updated standard deviations clustered are respectively less than the standard deviation threshold method, after the m updates temporarily
Interim cluster be used as the original user set of tags.
4. information-pushing method as described in claim 1, which is characterized in that described to be believed according to the historical behavior of the user
Breath, is ranked up the original user set of tags, obtaining target user's set of tags includes：
Based on the historical behavior information, the generated time of the corresponding historical behavior information of user tag is obtained；
The user tag is ranked up according to the generated time, obtains user tag sequence；
According to the user tag sequence, the original user set of tags is ranked up, obtains target user's set of tags.
5. information-pushing method as described in claim 1, which is characterized in that it is described by the way of term vector to the use
Family keyword is trained, and after determining user tag, described information method for pushing further includes：
According to preset client's tag library, determines different user groups and its corresponding user tag, obtain group's label；
For each user group, is classified to group's label based on the K-Means aggregating algorithms, be somebody's turn to do
The corresponding initial population's set of tags of user group；
For each user group, initial population's set of tags is ranked up, obtains the corresponding group of the user group
Body set of tags；
Determine the targeted user population of the user；
According to each user group and its corresponding target group's set of tags, the corresponding mesh of the targeted user population is obtained
Mark group's set of tags；
Based on target group's set of tags, group's preference information is obtained；
The corresponding service label of group's preference information is obtained from preset service label library, and pushes institute to the user
State the corresponding business information of service label.
6. information-pushing method as claimed in claim 5, which is characterized in that it is described for each user group, to institute
It states initial population's set of tags to be ranked up, obtaining the corresponding group's set of tags of the user group includes：
For each user group, initial population's set of tags B is obtained1,B2,...,BtIn each user user's mark
Sign A1,A2,A3,...,ApUser tag word frequency within a preset time intervalWherein, the t and
The p is positive integer；
By initial population's set of tags BjEach of the user tag word frequency be added, obtain initial population's label
Group BjGroup's label word frequencyWherein, [1, t] j ∈；
According to group's label word frequencySize initial population's set of tags is ranked up, obtain
Group's set of tags.
7. a kind of information push-delivery apparatus, which is characterized in that described information pusher includes：
Historical behavior data obtaining module, the historical behavior information for obtaining user；
User key words acquisition module obtains user key words for carrying out analysis filtering to the historical behavior information；
User tag generation module determines that user marks for being trained to the user key words by the way of term vector
Label；
Original user set of tags generation module is classified to the user tag for being based on K-Means aggregating algorithms, is obtained
Original user set of tags；
Target user's set of tags generation module, for the historical behavior information according to the user, to the original user label
Group is ranked up, and obtains target user's set of tags；
User preference information acquisition module obtains the preference information of the user for being based on target user's set of tags；
First business information pushing module, for obtaining the corresponding industry of the preference information from preset service label library
Business label, and push the corresponding business information of the service label to the user.
8. information push-delivery apparatus as claimed in claim 7, which is characterized in that the user tag generation module includes：
Basic term vector acquiring unit, for based on default corpus, building the basic term vector of each user key words；
Minimum space distance acquiring unit, for for each basic term vector, calculating the basis term vector and other bases
Space length between plinth term vector, and minimum space of the minimum value as the basis term vector is chosen from the space length
Distance；
User tag generation unit, for the basis of pre-set space distance threshold will to be less than or equal in minimum space distance
Term vector, as user tag.
9. a kind of terminal device, including memory, processor and it is stored in the memory and can be on the processor
The computer program of operation, which is characterized in that the processor realizes such as claim 1 to 6 when executing the computer program
The step of any one described information method for pushing.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, feature to exist
In the step of any one of such as claim 1 to 6 of realization described information method for pushing when the computer program is executed by processor
Suddenly.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810435812.7A CN108363821A (en) | 2018-05-09 | 2018-05-09 | A kind of information-pushing method, device, terminal device and storage medium |
| PCT/CN2018/122729 WO2019214245A1 (en) | 2018-05-09 | 2018-12-21 | Information pushing method and apparatus, and terminal device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810435812.7A CN108363821A (en) | 2018-05-09 | 2018-05-09 | A kind of information-pushing method, device, terminal device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108363821A true CN108363821A (en) | 2018-08-03 |
Family
ID=63011935
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810435812.7A Pending CN108363821A (en) | 2018-05-09 | 2018-05-09 | A kind of information-pushing method, device, terminal device and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108363821A (en) |
| WO (1) | WO2019214245A1 (en) |
Cited By (47)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214904A (en) * | 2018-10-11 | 2019-01-15 | 平安科技（深圳）有限公司 | Acquisition methods, device, computer equipment and the storage medium of financial fraud clue |
| CN109271609A (en) * | 2018-09-14 | 2019-01-25 | 广州神马移动信息科技有限公司 | Label generating method, device, terminal device and computer storage medium |
| CN109284372A (en) * | 2018-09-03 | 2019-01-29 | 平安证券股份有限公司 | User's operation behavior analysis method, electronic device and computer readable storage medium |
| CN109299356A (en) * | 2018-08-22 | 2019-02-01 | 中国平安人寿保险股份有限公司 | Activity recommendation method, apparatus, electronic equipment and storage medium based on big data |
| CN109344253A (en) * | 2018-09-18 | 2019-02-15 | 平安科技（深圳）有限公司 | Add method, apparatus, computer equipment and the storage medium of user tag |
| CN109344232A (en) * | 2018-11-13 | 2019-02-15 | 平安科技（深圳）有限公司 | A kind of public feelings information search method and terminal device |
| CN109360658A (en) * | 2018-11-01 | 2019-02-19 | 北京航空航天大学 | A kind of the disease pattern method for digging and device of word-based vector model |
| CN109492159A (en) * | 2018-10-31 | 2019-03-19 | 深圳市轱辘汽车维修技术有限公司 | A kind of content delivery method, content push device and electronic equipment |
| CN109658129A (en) * | 2018-11-22 | 2019-04-19 | 北京奇虎科技有限公司 | A kind of generation method and device of user's portrait |
| CN109842688A (en) * | 2019-03-07 | 2019-06-04 | 北京达佳互联信息技术有限公司 | A kind of content recommendation method, device, electronic equipment and storage medium |
| CN109857854A (en) * | 2019-01-02 | 2019-06-07 | 新浪网技术(中国)有限公司 | A kind of user's commercial labels method for digging and device, server |
| CN109960763A (en) * | 2019-03-21 | 2019-07-02 | 湖南大学 | A kind of photography community personalization friend recommendation method based on user's fine granularity photography preference |
| CN110069635A (en) * | 2019-04-30 | 2019-07-30 | 秒针信息技术有限公司 | A kind of determination method and device of temperature word |
| CN110110038A (en) * | 2018-08-17 | 2019-08-09 | 平安科技（深圳）有限公司 | Traffic predicting method, device, server and storage medium |
| CN110135893A (en) * | 2019-04-16 | 2019-08-16 | 深圳壹账通智能科技有限公司 | Determination method, apparatus, computer equipment and the storage medium of potential user |
| CN110163719A (en) * | 2019-04-15 | 2019-08-23 | 深圳壹账通智能科技有限公司 | Information-pushing method, device, computer equipment and storage medium |
| CN110188273A (en) * | 2019-05-27 | 2019-08-30 | 北京字节跳动网络技术有限公司 | Notification method, device, server and the readable medium of information content |
| CN110351343A (en) * | 2019-06-21 | 2019-10-18 | 北京纵横无双科技有限公司 | A kind of accurate information method for pushing and device based on big data analysis |
| WO2019214245A1 (en) * | 2018-05-09 | 2019-11-14 | 深圳壹账通智能科技有限公司 | Information pushing method and apparatus, and terminal device and storage medium |
| CN110517114A (en) * | 2019-08-21 | 2019-11-29 | 广州云徙科技有限公司 | A kind of information-pushing method and system based on community discovery algorithm |
| CN110688475A (en) * | 2019-09-05 | 2020-01-14 | 上海异势信息科技有限公司 | Article recommendation method and system based on content subjective tendency |
| CN110688406A (en) * | 2019-09-06 | 2020-01-14 | 平安医疗健康管理股份有限公司 | Data processing method, data processing device, computer equipment and storage medium |
| CN110727859A (en) * | 2019-09-12 | 2020-01-24 | 北京十分科技有限公司 | Recommendation information pushing method and device |
| CN110765275A (en) * | 2019-10-14 | 2020-02-07 | 平安医疗健康管理股份有限公司 | Search method, search device, computer equipment and storage medium |
| CN110765328A (en) * | 2019-10-25 | 2020-02-07 | 腾讯科技（深圳）有限公司 | Data processing method, device and storage medium |
| CN110827138A (en) * | 2019-10-31 | 2020-02-21 | 北京芯盾时代科技有限公司 | Push information determining method and device |
| CN110969486A (en) * | 2019-11-29 | 2020-04-07 | 中国银行股份有限公司 | Advertisement putting method, user terminal, server, system and storage medium |
| CN111047336A (en) * | 2019-12-24 | 2020-04-21 | 太平金融科技服务（上海）有限公司 | User label pushing method, user label display method, device and computer equipment |
| WO2020082596A1 (en) * | 2018-10-23 | 2020-04-30 | 深圳壹账通智能科技有限公司 | Data processing-based automatic user profile generating method and system |
| CN111309903A (en) * | 2020-01-20 | 2020-06-19 | 北京大米未来科技有限公司 | Data processing method and device, storage medium and electronic equipment |
| CN111401935A (en) * | 2020-02-21 | 2020-07-10 | 中国平安财产保险股份有限公司 | Resource allocation method, device and storage medium |
| CN111400375A (en) * | 2020-03-19 | 2020-07-10 | 畅捷通信息技术股份有限公司 | Business opportunity mining method and device based on financial service data |
| CN111444417A (en) * | 2018-12-27 | 2020-07-24 | 北京奇虎科技有限公司 | Intelligent orientation method and device for popularizing user group and computing equipment |
| CN111723289A (en) * | 2020-06-08 | 2020-09-29 | 北京声智科技有限公司 | Information recommendation method and device |
| CN111782949A (en) * | 2020-06-30 | 2020-10-16 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN111882400A (en) * | 2020-07-31 | 2020-11-03 | 平安国际融资租赁有限公司 | Behavior recognition analysis method and device, computer equipment and readable storage medium |
| CN112150199A (en) * | 2020-09-21 | 2020-12-29 | 北京小米松果电子有限公司 | Method and device for determining vertical characteristics and storage medium |
| CN112182390A (en) * | 2020-09-29 | 2021-01-05 | 中国平安人寿保险股份有限公司 | Letter pushing method and device, computer equipment and storage medium |
| CN112419011A (en) * | 2020-12-14 | 2021-02-26 | 国泰新点软件股份有限公司 | Public resource transaction recommendation method and device and computer-readable storage medium |
| CN112613917A (en) * | 2020-12-30 | 2021-04-06 | 平安壹钱包电子商务有限公司 | Information pushing method, device and equipment based on user portrait and storage medium |
| CN112788087A (en) * | 2019-11-11 | 2021-05-11 | 中国电信股份有限公司 | Business product information pushing method and device and storage medium |
| CN113222697A (en) * | 2021-05-11 | 2021-08-06 | 湖北三赫智能科技有限公司 | Commodity information pushing method, commodity information pushing device, computer equipment and readable storage medium |
| CN113239185A (en) * | 2021-07-13 | 2021-08-10 | 深圳市创能亿科科技开发有限公司 | Method and device for making teaching courseware and computer readable storage medium |
| CN113412608A (en) * | 2019-06-24 | 2021-09-17 | 深圳市欢太科技有限公司 | Content pushing method and device, server and storage medium |
| CN113516524A (en) * | 2020-04-10 | 2021-10-19 | 北京沃东天骏信息技术有限公司 | Method and device for pushing information |
| CN113676505A (en) * | 2020-05-15 | 2021-11-19 | 财付通支付科技有限公司 | Information pushing method and device, computer equipment and storage medium |
| CN113723984A (en) * | 2021-03-03 | 2021-11-30 | 京东城市(北京)数字科技有限公司 | Method and device for acquiring crowd consumption portrait information and storage medium |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113362105B (en) * | 2021-06-01 | 2024-02-02 | 北京十一贝科技有限公司 | User tag forming method, apparatus and computer readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102253996A (en) * | 2011-07-08 | 2011-11-23 | 北京航空航天大学 | Multi-visual angle stagewise image clustering method |
| US20140247978A1 (en) * | 2013-03-04 | 2014-09-04 | Xerox Corporation | Pre-screening training data for classifiers |
| CN104636470A (en) * | 2015-02-11 | 2015-05-20 | 广州华多网络科技有限公司 | Method and device for recommending business information |
| CN106503015A (en) * | 2015-09-07 | 2017-03-15 | 国家计算机网络与信息安全管理中心 | A kind of method for building user's portrait |
| CN106611052A (en) * | 2016-12-26 | 2017-05-03 | 东软集团股份有限公司 | Text label determination method and device |
| CN107908789A (en) * | 2017-12-12 | 2018-04-13 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103164463B (en) * | 2011-12-16 | 2017-03-22 | 国际商业机器公司 | Method and device for recommending labels |
| CN104216881A (en) * | 2013-05-29 | 2014-12-17 | 腾讯科技（深圳）有限公司 | Method and device for recommending individual labels |
| CN103886090B (en) * | 2014-03-31 | 2018-01-02 | 北京搜狗科技发展有限公司 | Content recommendation method and device based on user preferences |
| CN108363821A (en) * | 2018-05-09 | 2018-08-03 | 深圳壹账通智能科技有限公司 | A kind of information-pushing method, device, terminal device and storage medium |
-
2018
- 2018-05-09 CN CN201810435812.7A patent/CN108363821A/en active Pending
- 2018-12-21 WO PCT/CN2018/122729 patent/WO2019214245A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102253996A (en) * | 2011-07-08 | 2011-11-23 | 北京航空航天大学 | Multi-visual angle stagewise image clustering method |
| US20140247978A1 (en) * | 2013-03-04 | 2014-09-04 | Xerox Corporation | Pre-screening training data for classifiers |
| CN104636470A (en) * | 2015-02-11 | 2015-05-20 | 广州华多网络科技有限公司 | Method and device for recommending business information |
| CN106503015A (en) * | 2015-09-07 | 2017-03-15 | 国家计算机网络与信息安全管理中心 | A kind of method for building user's portrait |
| CN106611052A (en) * | 2016-12-26 | 2017-05-03 | 东软集团股份有限公司 | Text label determination method and device |
| CN107908789A (en) * | 2017-12-12 | 2018-04-13 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
Cited By (60)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019214245A1 (en) * | 2018-05-09 | 2019-11-14 | 深圳壹账通智能科技有限公司 | Information pushing method and apparatus, and terminal device and storage medium |
| CN110110038A (en) * | 2018-08-17 | 2019-08-09 | 平安科技（深圳）有限公司 | Traffic predicting method, device, server and storage medium |
| CN109299356B (en) * | 2018-08-22 | 2024-02-06 | 中国平安人寿保险股份有限公司 | Activity recommendation method and device based on big data, electronic equipment and storage medium |
| CN109299356A (en) * | 2018-08-22 | 2019-02-01 | 中国平安人寿保险股份有限公司 | Activity recommendation method, apparatus, electronic equipment and storage medium based on big data |
| CN109284372A (en) * | 2018-09-03 | 2019-01-29 | 平安证券股份有限公司 | User's operation behavior analysis method, electronic device and computer readable storage medium |
| CN109284372B (en) * | 2018-09-03 | 2023-07-25 | 平安证券股份有限公司 | User operation behavior analysis method, electronic device and computer readable storage medium |
| CN109271609A (en) * | 2018-09-14 | 2019-01-25 | 广州神马移动信息科技有限公司 | Label generating method, device, terminal device and computer storage medium |
| CN109344253A (en) * | 2018-09-18 | 2019-02-15 | 平安科技（深圳）有限公司 | Add method, apparatus, computer equipment and the storage medium of user tag |
| CN109214904A (en) * | 2018-10-11 | 2019-01-15 | 平安科技（深圳）有限公司 | Acquisition methods, device, computer equipment and the storage medium of financial fraud clue |
| WO2020082596A1 (en) * | 2018-10-23 | 2020-04-30 | 深圳壹账通智能科技有限公司 | Data processing-based automatic user profile generating method and system |
| CN109492159A (en) * | 2018-10-31 | 2019-03-19 | 深圳市轱辘汽车维修技术有限公司 | A kind of content delivery method, content push device and electronic equipment |
| CN109360658A (en) * | 2018-11-01 | 2019-02-19 | 北京航空航天大学 | A kind of the disease pattern method for digging and device of word-based vector model |
| CN109360658B (en) * | 2018-11-01 | 2021-06-08 | 北京航空航天大学 | Disease pattern mining method and device based on word vector model |
| CN109344232A (en) * | 2018-11-13 | 2019-02-15 | 平安科技（深圳）有限公司 | A kind of public feelings information search method and terminal device |
| CN109344232B (en) * | 2018-11-13 | 2024-03-15 | 平安科技（深圳）有限公司 | Public opinion information retrieval method and terminal equipment |
| CN109658129A (en) * | 2018-11-22 | 2019-04-19 | 北京奇虎科技有限公司 | A kind of generation method and device of user's portrait |
| CN111444417A (en) * | 2018-12-27 | 2020-07-24 | 北京奇虎科技有限公司 | Intelligent orientation method and device for popularizing user group and computing equipment |
| CN109857854A (en) * | 2019-01-02 | 2019-06-07 | 新浪网技术(中国)有限公司 | A kind of user's commercial labels method for digging and device, server |
| CN109842688A (en) * | 2019-03-07 | 2019-06-04 | 北京达佳互联信息技术有限公司 | A kind of content recommendation method, device, electronic equipment and storage medium |
| CN109842688B (en) * | 2019-03-07 | 2022-03-01 | 北京达佳互联信息技术有限公司 | Content recommendation method and device, electronic equipment and storage medium |
| CN109960763A (en) * | 2019-03-21 | 2019-07-02 | 湖南大学 | A kind of photography community personalization friend recommendation method based on user's fine granularity photography preference |
| CN110163719A (en) * | 2019-04-15 | 2019-08-23 | 深圳壹账通智能科技有限公司 | Information-pushing method, device, computer equipment and storage medium |
| CN110135893A (en) * | 2019-04-16 | 2019-08-16 | 深圳壹账通智能科技有限公司 | Determination method, apparatus, computer equipment and the storage medium of potential user |
| CN110069635A (en) * | 2019-04-30 | 2019-07-30 | 秒针信息技术有限公司 | A kind of determination method and device of temperature word |
| CN110188273A (en) * | 2019-05-27 | 2019-08-30 | 北京字节跳动网络技术有限公司 | Notification method, device, server and the readable medium of information content |
| CN110351343A (en) * | 2019-06-21 | 2019-10-18 | 北京纵横无双科技有限公司 | A kind of accurate information method for pushing and device based on big data analysis |
| CN113412608A (en) * | 2019-06-24 | 2021-09-17 | 深圳市欢太科技有限公司 | Content pushing method and device, server and storage medium |
| CN110517114A (en) * | 2019-08-21 | 2019-11-29 | 广州云徙科技有限公司 | A kind of information-pushing method and system based on community discovery algorithm |
| CN110688475A (en) * | 2019-09-05 | 2020-01-14 | 上海异势信息科技有限公司 | Article recommendation method and system based on content subjective tendency |
| CN110688406A (en) * | 2019-09-06 | 2020-01-14 | 平安医疗健康管理股份有限公司 | Data processing method, data processing device, computer equipment and storage medium |
| CN110727859A (en) * | 2019-09-12 | 2020-01-24 | 北京十分科技有限公司 | Recommendation information pushing method and device |
| CN110765275B (en) * | 2019-10-14 | 2023-02-07 | 深圳平安医疗健康科技服务有限公司 | Search method, search device, computer equipment and storage medium |
| CN110765275A (en) * | 2019-10-14 | 2020-02-07 | 平安医疗健康管理股份有限公司 | Search method, search device, computer equipment and storage medium |
| CN110765328A (en) * | 2019-10-25 | 2020-02-07 | 腾讯科技（深圳）有限公司 | Data processing method, device and storage medium |
| CN110827138B (en) * | 2019-10-31 | 2022-07-05 | 北京芯盾时代科技有限公司 | Push information determining method and device |
| CN110827138A (en) * | 2019-10-31 | 2020-02-21 | 北京芯盾时代科技有限公司 | Push information determining method and device |
| CN112788087B (en) * | 2019-11-11 | 2023-04-07 | 中国电信股份有限公司 | Business product information pushing method and device and storage medium |
| CN112788087A (en) * | 2019-11-11 | 2021-05-11 | 中国电信股份有限公司 | Business product information pushing method and device and storage medium |
| CN110969486B (en) * | 2019-11-29 | 2024-02-27 | 中国银行股份有限公司 | Advertisement putting method, user, server, system and storage medium |
| CN110969486A (en) * | 2019-11-29 | 2020-04-07 | 中国银行股份有限公司 | Advertisement putting method, user terminal, server, system and storage medium |
| CN111047336A (en) * | 2019-12-24 | 2020-04-21 | 太平金融科技服务（上海）有限公司 | User label pushing method, user label display method, device and computer equipment |
| CN111309903A (en) * | 2020-01-20 | 2020-06-19 | 北京大米未来科技有限公司 | Data processing method and device, storage medium and electronic equipment |
| CN111401935A (en) * | 2020-02-21 | 2020-07-10 | 中国平安财产保险股份有限公司 | Resource allocation method, device and storage medium |
| CN111401935B (en) * | 2020-02-21 | 2023-04-07 | 中国平安财产保险股份有限公司 | Resource allocation method, device and storage medium |
| CN111400375A (en) * | 2020-03-19 | 2020-07-10 | 畅捷通信息技术股份有限公司 | Business opportunity mining method and device based on financial service data |
| CN113516524A (en) * | 2020-04-10 | 2021-10-19 | 北京沃东天骏信息技术有限公司 | Method and device for pushing information |
| CN113676505A (en) * | 2020-05-15 | 2021-11-19 | 财付通支付科技有限公司 | Information pushing method and device, computer equipment and storage medium |
| CN111723289B (en) * | 2020-06-08 | 2024-02-02 | 北京声智科技有限公司 | Information recommendation method and device |
| CN111723289A (en) * | 2020-06-08 | 2020-09-29 | 北京声智科技有限公司 | Information recommendation method and device |
| CN111782949A (en) * | 2020-06-30 | 2020-10-16 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN111882400A (en) * | 2020-07-31 | 2020-11-03 | 平安国际融资租赁有限公司 | Behavior recognition analysis method and device, computer equipment and readable storage medium |
| CN112150199A (en) * | 2020-09-21 | 2020-12-29 | 北京小米松果电子有限公司 | Method and device for determining vertical characteristics and storage medium |
| CN112182390A (en) * | 2020-09-29 | 2021-01-05 | 中国平安人寿保险股份有限公司 | Letter pushing method and device, computer equipment and storage medium |
| CN112182390B (en) * | 2020-09-29 | 2024-02-09 | 中国平安人寿保险股份有限公司 | Mail pushing method, device, computer equipment and storage medium |
| CN112419011A (en) * | 2020-12-14 | 2021-02-26 | 国泰新点软件股份有限公司 | Public resource transaction recommendation method and device and computer-readable storage medium |
| CN112613917A (en) * | 2020-12-30 | 2021-04-06 | 平安壹钱包电子商务有限公司 | Information pushing method, device and equipment based on user portrait and storage medium |
| CN113723984A (en) * | 2021-03-03 | 2021-11-30 | 京东城市(北京)数字科技有限公司 | Method and device for acquiring crowd consumption portrait information and storage medium |
| CN113222697A (en) * | 2021-05-11 | 2021-08-06 | 湖北三赫智能科技有限公司 | Commodity information pushing method, commodity information pushing device, computer equipment and readable storage medium |
| CN113239185B (en) * | 2021-07-13 | 2021-10-29 | 深圳市创能亿科科技开发有限公司 | Method and device for making teaching courseware and computer readable storage medium |
| CN113239185A (en) * | 2021-07-13 | 2021-08-10 | 深圳市创能亿科科技开发有限公司 | Method and device for making teaching courseware and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019214245A1 (en) | 2019-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108363821A (en) | A kind of information-pushing method, device, terminal device and storage medium | |
| CN102708131B (en) | By consumer's automatic classification in fine point | |
| CN110163647B (en) | Data processing method and device | |
| CN106844407B (en) | Tag network generation method and system based on data set correlation | |
| CN104834641B (en) | The processing method and related system of network media information | |
| US20110113095A1 (en) | System and Method For Value Significance Evaluation of Ontological Subjects of Networks and The Applications Thereof | |
| CN106919619A (en) | A kind of commercial articles clustering method, device and electronic equipment | |
| CN108460082B (en) | Recommendation method and device and electronic equipment | |
| CN105095187A (en) | Search intention identification method and device | |
| CN107391582B (en) | The information recommendation method of user preference similarity is calculated based on context ontology tree | |
| CN105225135B (en) | Potential customer identification method and device | |
| CN111489201A (en) | Method, device and storage medium for analyzing customer value | |
| CN109389321B (en) | Item list classification method and device | |
| CN112070577A (en) | Commodity recommendation method, system, equipment and medium | |
| CN114238573B (en) | Text countercheck sample-based information pushing method and device | |
| CN111696656B (en) | Doctor evaluation method and device of Internet medical platform | |
| CN112669113A (en) | Product recommendation method and device, storage medium and electronic device | |
| CN114266443A (en) | Data evaluation method and device, electronic equipment and storage medium | |
| CN110795613A (en) | Commodity searching method, device and system and electronic equipment | |
| CN111538909A (en) | Information recommendation method and device | |
| CN110852785A (en) | User grading method, device and computer readable storage medium | |
| CN116823410B (en) | Data processing method, object processing method, recommending method and computing device | |
| CN112989053A (en) | Periodical recommendation method and device | |
| CN117520503A (en) | Financial customer service dialogue generation method, device, equipment and medium based on LLM model | |
| CN114881712B (en) | Intelligent advertisement putting method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1256481 Country of ref document: HK |
|
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180803 |