CN108647729A - A kind of user's portrait acquisition methods - Google Patents
A kind of user's portrait acquisition methods Download PDFInfo
- Publication number
- CN108647729A CN108647729A CN201810455403.3A CN201810455403A CN108647729A CN 108647729 A CN108647729 A CN 108647729A CN 201810455403 A CN201810455403 A CN 201810455403A CN 108647729 A CN108647729 A CN 108647729A
- Authority
- CN
- China
- Prior art keywords
- user
- data
- characteristic value
- feature collection
- demand
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Abstract
Present invention relates particularly to a kind of user portrait acquisition methods, it can classify and analyze respectively by basic data to user and User action log corresponding with basic data, it can do and preferably establish user's disaggregated model and personal behavior model, and obtain user's portrait of different user by the two models.
Description
Technical field
The present invention relates to information classification process fields, and in particular to a kind of user's portrait acquisition methods.
Background technology
User's portrait is also known as user role, and having for target user, contact user's demand and design direction is delineated as a kind of
Effect tool, user's portrait are widely used in each field.
In scientific and technological resources supply and demand docking, the scientific and technological achievement of supplier and the Science & Technology Demands of demander are all huge, for supplying
Fang Eryan, scientific and technological resources data be it is detailed, huge, as scientific and technological achievement displaying be it is clear clearly, but as section
Skill result output is weak, this is that scientific and technological resources supply and demand docking mode determines.In most cases, scientific and technological resources supply and demand is double
Square demand can not match.It is main reason is that both sides of supply and demand possess Asymmetry information etc., on the one hand, supplier fails according to market
Demand segments the scientific and technological resources occupied, and also can not quickly learn the wish of demander;On the other hand, demander to oneself requirement description not
Enough detailed or description demand characteristic is imagined with supplier differs larger.Which results in the docking of current both sides of supply and demand scientific and technological resources extremely
It is difficult.When both sides of supply and demand complete the result of sufficient preparation, scientific and technological resources docking could be completed, this greatly reduces science and technology
Resource supply and demand docks efficiency.Even if not sufficient preparation, demander is also required to think by repeatedly searching for investigation and just can know that
The supplier's information wanted, meanwhile, in demander constantly retrieves, the retrieval type used is provided by supplier, retrieval type subdivision
Degree is inadequate, while index construct and non-compliant demander's wish, and to demander and unfriendly, such scientific and technological resources supply and demand docking is pole
Inconvenient.
, can not be corresponding with class of subscriber by scientific and technological resources since the classification to user is fuzzy, cause scientific and technological resources supply and demand double
Square dispensing is uneven, therefore the classification and accurate role positioning to supply and demand user are optimize scientific and technological resources supply and demand interconnection method first
Step,.To sum up, the classification of scientific and technological resources user and user, which draw a portrait, obtains problems demand solution.
Invention content
In order to overcome the deficiencies of existing technologies, the present invention provides a kind of user and draws a portrait acquisition methods, can preferably to
Family classify and analyze the behavior of user, and obtains user's portrait according to the result of classification and analysis.
For foregoing invention purpose, the present invention is to solve in this way:A kind of user's portrait acquisition methods, are based on multiple use
The basic data at family and User action log corresponding with basic data realize that User action log includes user resources supply and demand row
For data, which is characterized in that include the following steps:
S1, all users of extraction basic data in each data characteristic information, using corresponding characteristic information to phase
The basic data of same type carries out clustering, and obtains multiple corresponding fisrt feature collection;It is built according to all fisrt feature collection
Vertical user's disaggregated model;
The characteristic information of each data, utilizes corresponding characteristic information pair in S2, extraction user resources supply and demand behavioral data
The user resources supply and demand behavioral data of same type carries out clustering, and obtains multiple corresponding second feature collection, according to institute
There is second feature collection to establish the initial behavior model of user;According to all fisrt feature collection, second feature collection and user classification mould
Type establishes behavioral data feature set.
S3, using behavioral data feature set as the training sample of personal behavior model, to establish personal behavior model;
S4, user's portrait is obtained according to user's disaggregated model and personal behavior model.
Classify and analyze respectively by basic data to user and User action log corresponding with basic data,
It can do and preferably establish user's disaggregated model and personal behavior model, and obtain user's picture of different user by the two models
Picture.Obtain the foundation that the initial behavior model of user provides classification analysis for subsequent user behavioural analysis.Due to user behavior number
According to belonging to a kind of dynamic data increase with time, then subsequently user behavior is constantly analyzed, when optimizing behavioural characteristic collection,
Directly similar behavioral data classification storage is analyzed again, the advantage of doing so is that, redundant data is fallen in one side screening, gives
Initial behavioral data assigns label, reduces intractability;On the other hand, according to analysis result, to initial behavioural characteristic classify into
Row optimization, keeps disaggregated model more accurate.
Further, the specific forming process of fisrt feature collection is in the step S1:
S1.1 randomly selects the basic data of same type, and the data to extracting carry out clustering and obtain
Several characteristic value Mi;
S1.2 carries out stratified sampling after classifying to the basic data of same type, and gathers to the data sampled out
Alanysis obtains several characteristic value Mk;
S1.3 optimizes Mi according to the similarity of Mi and Mk, finally obtains several characteristic value M, forms fisrt feature
Collection.
Further, basic data includes the level one data for describing user role feature, description user resources for pleading
The secondary data of condition;All fisrt feature concentration in step S1 includes the third feature collection formed by level one data and by two
The fourth feature collection that level data is formed, is analyzed according to third feature set pair fourth feature collection, obtains the index between them,
User's disaggregated model is established according to the index.
Further, step S1.1 is repeated repeatedly, is repeating the characteristic value Mi obtained in the process according to sampling
Number, sampling proportion, process of cluster analysis are optimized to executing obtained characteristic value Mi for the first time.
Further, the specific method for building up of behavioral data feature set is in the step S2:
S2.1, second feature collection, third feature collection and fourth feature collection respectively include several characteristic values, each characteristic value pair
The fellow users for answering respective numbers obtain corresponding weight according to the user of the corresponding different number of characteristic value;According to different
Weight is sampled the corresponding user of each characteristic value, calculates the use sampled out respectively from second feature collection and fourth feature collection
The similarity at family obtains similar features index Q1;
S2.2, according to Q1, user's disaggregated model, to from fisrt feature concentrate each characteristic value correspond to the user to sample out into
Row similarity analysis obtains similar features index Q2;
S2.3, behavioral data feature set is established according to Q1 and Q2.
Further, before step S 4, also user's disaggregated model is optimized, the specific steps are:Acquisition is used in real time
User resources supply and demand behavioral data sort out and divide by family resource supply and demand behavioral data according to the initial behavior model of user first
Analysis obtains the behavioural characteristic of all types of user's collection, is developed according to the behavioural characteristic of all types of users and carried out to user's disaggregated model
It corrects;Secondly, user resources supply and demand behavioral data is carried out by classification analysis according to behavioural characteristic, obtains behavioural characteristic value, according to
Behavioural characteristic value classifies to user, and then is modified to user's disaggregated model.
Further, to user's disaggregated model carry out for the first time it is modified the specific steps are:It is right according to user's disaggregated model
The dynamic behaviour data of all types of users are analyzed, and behavioural characteristic evolution model is obtained;Behavioural characteristic evolution model includes spy
The variation of value indicative and respective weights, according to behavioral data feature set and behavioural characteristic evolution model to user's disaggregated model and use
User data feature set is modified.
Further, each behavioural characteristic value corresponds to the user of respective numbers, and corresponding weight is according to behavioural characteristic value pair
The user for the different number answered obtains;Layering pumping is carried out to behavioral data feature set according to weight corresponding with behavioural characteristic value
Sample, the sample obtained to sampling are analyzed, obtain the second behavioural characteristic modifying factor, utilize the second behavioural characteristic modifying factor
The weight of fisrt feature collection is corrected again.
Compared with the prior art, the beneficial effects of the present invention are:Pass through the basic data and and basic data to user
Corresponding User action log is classified and is analyzed respectively, can do and preferably establish user's disaggregated model and user behavior mould
Type, and drawn a portrait by the user of the two models acquisition different user.
Description of the drawings
Fig. 1 is the broad flow diagram of the method for the present invention.
Fig. 2 is the flow chart of the specific forming process of fisrt feature collection in step S1 of the present invention.
Fig. 3 is the flow chart of the specific method for building up of behavioral data feature set in step S2 of the present invention.
Specific implementation mode
The present invention is described in detail according to the following examples and attached drawing.
A kind of user's portrait acquisition methods, the basic data based on multiple users and user behavior corresponding with basic data
Daily record realizes that basic data includes the level one data for describing user role feature, describes the two level of user resources supply and demand situation
Data, User action log include user resources supply and demand behavioral data.Wherein level one data includes identity information (such as working year
Part, post, gender, team's essential information etc.), scientific research field, scientific achievement, potential research trends etc., secondary data includes using
Family supply and demand wish etc..For user resources supply and demand behavioral data, including time attribute data, geographical attribute data and by action
The user's operation attribute data of classification, for including search, collection, transaction, consulting etc. by the user's operation data of the classification of motion.
This method includes following steps as shown in Figure 1:
S1, the characteristic information that each data in the basic data of all users are extracted in the form of keyword, using corresponding
Characteristic information to the basic data of same type, clustering is carried out by K-means algorithms, and obtain multiple corresponding the
One feature set, wherein the K values used in K-means algorithms are by manually choosing;User's classification mould is established according to all fisrt feature collection
Type;
Scientific and technological resources supply and demand is docked in example, using identity information as the major key of data characteristics collection, including industry background, from
Industry time, post, team information etc..These identity information datas will be moved with other users basic data, such as scientific achievement, research
State etc. is combined carry out signature analysis.First, industry background keyword is extracted, and counts being averaged under the keyword industry background
Working time and standard deviation;Then, feature point is carried out to scientific achievement, research trends of user etc. according to industry background feature
Analysis;Finally, industry maturity and industry essential characteristic are obtained according to above-mentioned two steps result.By the sector essential characteristic to
Family carries out a subseries.
The characteristic information of each data, utilizes corresponding characteristic information pair in S2, extraction user resources supply and demand behavioral data
The user resources supply and demand behavioral data of same type carries out clustering, and obtains multiple corresponding second feature collection, according to institute
There is second feature collection to establish the initial behavior model of user;According to all fisrt feature collection, second feature collection and user classification mould
Type establishes behavioral data feature set.
Scientific and technological resources supply and demand is docked in example, and user resources supply and demand behavioral data includes two aspect data of supply and demand.User is
The scientific and technological resources information occupied is referred to as user resources supply and demand data, including user can disclosed instrument and equipment, technological means,
Patented method etc..User's need technical solution to be offered, instrument and equipment, actual scene solution etc. belong to user resources demand
Data.User resources supply and demand data generally include instrument data, patent data, application scheme data etc., these data need logical
It crosses general nature method of semantic differential to be handled to obtain keyword therein, further obtains supplier user's by signature analysis
Role characteristic.User resources demand data generally includes specific demand information, and instrument as required, technological means etc. are also wrapped
Include some Fuzzy Demand information.Signature analysis is carried out for specific demand information and obtains the first role feature of demander user;
For Fuzzy Demand information, such as description application scenarios or demand purpose, keyword is manually extracted from the angle of application, together
Sample obtains the second role feature of demander user by signature analysis;
S3, using behavioral data feature set as the training sample of personal behavior model, to establish personal behavior model;
S4, user's portrait is obtained according to user's disaggregated model and personal behavior model.
Classify and analyze respectively by basic data to user and User action log corresponding with basic data,
It can do and preferably establish user's disaggregated model and personal behavior model, and obtain user's picture of different user by the two models
Picture.
More preferably, the specific forming process of fisrt feature collection is in the step S1 as shown in Figure 2:
S1.1 randomly selects the basic data of same type, and the data to extracting carry out clustering and obtain
Several characteristic value Mi;
S1.2 carries out stratified sampling after classifying to the basic data of same type, and gathers to the data sampled out
Alanysis obtains several characteristic value Mk;
S1.3 optimizes Mi according to the similarity of Mi and Mk, finally obtains several characteristic value M, forms fisrt feature
Collection.
More preferably, all fisrt feature in step S1 concentrate include the third feature collection formed by level one data and by
The fourth feature collection that secondary data is formed, is analyzed according to third feature set pair fourth feature collection, obtains the rope between them
Draw, user's disaggregated model is established according to the index.
More preferably, step S1.1 is repeated repeatedly, is repeating the characteristic value Mi obtained in the process according to sampling time
Number, sampling proportion, the dispersion that clusters, the information such as evolution that cluster are optimized to executing obtained characteristic value Mi for the first time.It clusters
To divide the data aggregation zone of formation in process of cluster analysis.
More preferably, as shown in figure 3, the specific method for building up of behavioral data feature set is in the step S2:
S2.1, second feature collection, third feature collection and fourth feature collection respectively include several characteristic values, each characteristic value pair
The fellow users for answering respective numbers obtain corresponding weight according to the user of the corresponding different number of characteristic value;According to different
Weight is sampled the corresponding user of each characteristic value, calculates the use sampled out respectively from second feature collection and fourth feature collection
The similarity at family obtains similar features index Q1;
Scientific and technological resources supply and demand is docked in example, and the purpose of user behavior is to describe user resources to solve the problems, such as supply and demand
The secondary data of supply and demand situation and user resources supply and demand behavioral data are closely related, therefore the second collection that need to obtain behavioral data
The fourth feature obtained with user resources supply and demand data is docked.
S2.2, according to Q1, user's disaggregated model, to from fisrt feature concentrate each characteristic value correspond to the user to sample out into
Row similarity analysis obtains similar features index Q2;
S2.3, behavioral data feature set is established according to Q1 and Q2.
More preferably, before step S 4, also user's disaggregated model is optimized, the specific steps are:Acquisition user in real time
User resources supply and demand behavioral data is carried out classification analysis by resource supply and demand behavioral data according to the initial behavior model of user first,
The behavioural characteristic of all types of user's collection is obtained, user's disaggregated model is repaiied according to the differentiation of the behavioural characteristic of all types of users
Just;Secondly, user resources supply and demand behavioral data is carried out by classification analysis according to behavioural characteristic, behavioural characteristic value is obtained, according to row
It is characterized value to classify to user, and then user's disaggregated model is modified.
Firstly, since user resources supply and demand behavioral data is a kind of dynamic data increase with time;Secondly as not building
Vertical perfect scientific and technological resources user behavior analysis model, without specific criteria for classification.Therefore, it is necessary to by supplying user resources
It asks behavioral data to classify, the particular content of different types of behavior is analyzed, to obtain user behavior characteristics value,
In analytic process, since there is user resources supply and demand behavioral data higher complexity, the great scale of construction need to use neural network
Algorithm extracts classification to content of the act, to reduce content of the act keyword, finds the main behavioural characteristic value of all types of user.
More preferably, to user's disaggregated model carry out for the first time it is modified the specific steps are:According to user's disaggregated model, to each
The dynamic behaviour data of type of user are analyzed, and behavioural characteristic evolution model is obtained;Behavioural characteristic evolution model includes feature
The variation of value and respective weights, according to behavioral data feature set and behavioural characteristic evolution model to user's disaggregated model and user
Data characteristics collection is modified.
User's dynamic behaviour data are docked in portal website by scientific and technological resources and are obtained, and browsing record, search are generally comprised
Record, transaction record, search record etc., each behavioral data all include the information such as time, place, motivation, result.Time is logical
Often refer to the duration of behavior execution;Place is often referred to behavior execution place;Motivation is often referred to the behavior of user before the behavior
Characteristic value, referred to as motivation feature, the motivation feature are obtained by behavioural characteristic before the behavior and user's current character feature
It arrives;The result is that the evaluation to user's behavior, mainly by being obtained to the analysis of the entire action process inner link of user.Such as
Certain user inputs with the relevant keyword of device information " fluorescence microscope " for searching for " fluorescence microscope " supply-demand information, portal
All sharable " fluorescence microscopes " is enumerated in website, and user browses these information, and a certain " fluorescence microscopy is chosen by consulting
Mirror " simultaneously is completed to merchandise, and so far user behavior terminates.According to user's behavior, specific analytical procedure is:1, User ID is obtained,
User data information is searched according to User ID, the characteristic information of the user is determined according to user data feature set, is believed according to feature
Breath determines class of subscriber, then obtains the behavioral data feature of category user;2, user behavior information is recorded, according to example
Have, search behavior and search content " fluorescence ", " microscope ", navigation patterns and browsing duration, browsing data volume, seek advice from behavior and
Reference content records, trading activity and transaction details;3, semantic and signature analysis is carried out to user behavior information to be somebody's turn to do
User's current behavior characteristic value, compares the behavioural characteristic of this feature value and such user, to such user characteristics model into
Row optimization, and then with the increase of the user behavior data scale of construction, the behavioural characteristic variation that can also therefrom obtain such user becomes
Gesture advanced optimizes user data model;4, it is obtained about the motivation feature of user and behavior evaluation, the part works to user
Characteristic modification mode is telltale, and motivation characteristic value and behavior evaluation are determined by user role feature and user behavior complexity
It is fixed.
More preferably, each behavioural characteristic value corresponds to the user of respective numbers, and corresponding weight is corresponded to according to behavioural characteristic value
The user of different number obtain;Stratified sampling is carried out to behavioral data feature set according to weight corresponding with behavioural characteristic value,
The sample obtained to sampling is analyzed, and obtains the second behavioural characteristic modifying factor, again using the second behavioural characteristic modifying factor
The secondary weight for correcting fisrt feature collection.
Above-mentioned weight is determined by data irrelevance, the intersection that clusters complexity, feature hierarchy degree.
Claims (8)
1. a kind of user draws a portrait acquisition methods, the basic data based on multiple users and user behavior day corresponding with basic data
Will realizes that User action log includes user resources supply and demand behavioral data, which is characterized in that is included the following steps:
S1, all users of extraction basic data in each data characteristic information, using corresponding characteristic information to mutually similar
The basic data of type carries out clustering, and obtains multiple corresponding fisrt feature collection;It is established and is used according to all fisrt feature collection
Family disaggregated model;
The characteristic information of each data in S2, extraction user resources supply and demand behavioral data, using corresponding characteristic information to identical
The user resources supply and demand behavioral data of type carries out clustering, and obtains multiple corresponding second feature collection, according to all the
Two feature sets establish the initial behavior model of user;According to all fisrt feature collection, second feature collection and user's disaggregated model come
Establish behavioral data feature set;
S3, using behavioral data feature set as the training sample of personal behavior model, to establish personal behavior model;
S4, user's portrait is obtained according to user's disaggregated model and personal behavior model.
The acquisition methods 2. a kind of user according to claim 1 draws a portrait, which is characterized in that fisrt feature in the step S1
Collection specific forming process be:
S1.1 randomly selects the basic data of same type, and the data to extracting carry out clustering obtain it is several
Characteristic value Mi;
S1.2 carries out stratified sampling after classifying to the basic data of same type, and the data to sampling out carry out cluster point
Analysis obtains several characteristic value Mk;
S1.3 optimizes Mi according to the similarity of Mi and Mk, finally obtains several characteristic value M, forms fisrt feature collection.
The acquisition methods 3. a kind of user according to claim 1 draws a portrait, which is characterized in that basic data includes that description is used
The level one data of family role characteristic, the secondary data for describing user resources supply and demand situation;All fisrt feature collection in step S1
The fourth feature collection for including the third feature collection formed by level one data and being formed by secondary data, according to third feature collection
Fourth feature collection is analyzed, the index between them is obtained, user's disaggregated model is established according to the index.
The acquisition methods 4. a kind of user according to claim 2 draws a portrait, which is characterized in that step S1.1 repeats more
It is secondary, the characteristic value Mi obtained in the process is being repeated according to frequency in sampling, sampling proportion, process of cluster analysis to holding for the first time
The characteristic value Mi that row obtains is optimized.
The acquisition methods 5. a kind of user according to claim 3 draws a portrait, which is characterized in that behavioral data in the step S2
The specific method for building up of feature set is:
S2.1, second feature collection, third feature collection and fourth feature collection respectively include several characteristic values, and each characteristic value corresponds to phase
The fellow users for answering quantity obtain corresponding weight according to the user of the corresponding different number of characteristic value;According to different weights
To each characteristic value, corresponding user is sampled, and calculates the user's to sample out respectively from second feature collection and fourth feature collection
Similarity obtains similar features index Q1;
S2.2, according to Q1, user's disaggregated model, carry out phase to concentrating each characteristic value to correspond to the user to sample out from fisrt feature
It is analyzed like degree, obtains similar features index Q2;
S2.3, behavioral data feature set is established according to Q1 and Q2.
The acquisition methods 6. a kind of user according to claim 1 draws a portrait, which is characterized in that before step S 4, also to
Family disaggregated model optimizes, the specific steps are:Acquisition user resources supply and demand behavioral data in real time, it is first, initial according to user
User resources supply and demand behavioral data is carried out classification analysis by behavior model, the behavioural characteristic of all types of user's collection is obtained, according to each
The behavioural characteristic differentiation of type of user is modified user's disaggregated model;Secondly, according to behavioural characteristic by user resources supply and demand
Behavioral data carries out classification analysis, obtains behavioural characteristic value, is classified to user according to behavioural characteristic value, and then to user point
Class model is modified.
The acquisition methods 7. a kind of user according to claim 6 draws a portrait, which is characterized in that the is carried out to user's disaggregated model
It is primary it is modified the specific steps are:According to user's disaggregated model, the dynamic behaviour data of all types of users are analyzed, are obtained
Behavioural characteristic evolution model;Behavioural characteristic evolution model includes the variation of characteristic value and respective weights, according to behavioral data feature
Collection and behavioural characteristic evolution model are modified user's disaggregated model and user data feature set.
The acquisition methods 8. a kind of user according to claim 7 draws a portrait, which is characterized in that each behavioural characteristic value corresponds to phase
The user of quantity, the user that corresponding weight is worth corresponding different number according to behavioural characteristic is answered to obtain;According to behavioural characteristic
It is worth corresponding weight and stratified sampling is carried out to behavioral data feature set, the sample obtained to sampling is analyzed, and the second row is obtained
It is characterized modifying factor, corrects the weight of fisrt feature collection again using the second behavioural characteristic modifying factor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810455403.3A CN108647729B (en) | 2018-05-11 | 2018-05-11 | User portrait acquisition method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810455403.3A CN108647729B (en) | 2018-05-11 | 2018-05-11 | User portrait acquisition method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108647729A true CN108647729A (en) | 2018-10-12 |
CN108647729B CN108647729B (en) | 2022-03-04 |
Family
ID=63755152
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810455403.3A Active CN108647729B (en) | 2018-05-11 | 2018-05-11 | User portrait acquisition method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108647729B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109785034A (en) * | 2018-11-13 | 2019-05-21 | 北京码牛科技有限公司 | User's portrait generation method, device, electronic equipment and computer-readable medium |
CN109872242A (en) * | 2019-01-30 | 2019-06-11 | 北京字节跳动网络技术有限公司 | Information-pushing method and device |
CN110097278A (en) * | 2019-04-28 | 2019-08-06 | 广东省科技基础条件平台中心 | A kind of scientific and technological resources intelligent sharing Fusion training system and application system |
CN111159763A (en) * | 2019-12-26 | 2020-05-15 | 银江股份有限公司 | System and method for analyzing portrait of law-related personnel group |
CN112507204A (en) * | 2019-09-16 | 2021-03-16 | 北京智联云海科技有限公司 | Method for automatically constructing user portrait by utilizing data analysis |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106651424A (en) * | 2016-09-28 | 2017-05-10 | 国网山东省电力公司电力科学研究院 | Electric power user figure establishment and analysis method based on big data technology |
CN106933991A (en) * | 2017-02-24 | 2017-07-07 | 陈晶 | A kind of depth analysis towards intelligent terminal and user's portrait system and method |
CN107526754A (en) * | 2016-09-26 | 2017-12-29 | 广州速鸿信息科技有限公司 | A kind of user's portrait platform method for building up based on big data |
-
2018
- 2018-05-11 CN CN201810455403.3A patent/CN108647729B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107526754A (en) * | 2016-09-26 | 2017-12-29 | 广州速鸿信息科技有限公司 | A kind of user's portrait platform method for building up based on big data |
CN106651424A (en) * | 2016-09-28 | 2017-05-10 | 国网山东省电力公司电力科学研究院 | Electric power user figure establishment and analysis method based on big data technology |
CN106933991A (en) * | 2017-02-24 | 2017-07-07 | 陈晶 | A kind of depth analysis towards intelligent terminal and user's portrait system and method |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109785034A (en) * | 2018-11-13 | 2019-05-21 | 北京码牛科技有限公司 | User's portrait generation method, device, electronic equipment and computer-readable medium |
CN109872242A (en) * | 2019-01-30 | 2019-06-11 | 北京字节跳动网络技术有限公司 | Information-pushing method and device |
CN110097278A (en) * | 2019-04-28 | 2019-08-06 | 广东省科技基础条件平台中心 | A kind of scientific and technological resources intelligent sharing Fusion training system and application system |
CN110097278B (en) * | 2019-04-28 | 2021-06-08 | 广东省科技基础条件平台中心 | Intelligent sharing and fusion training system and application system for scientific and technological resources |
CN112507204A (en) * | 2019-09-16 | 2021-03-16 | 北京智联云海科技有限公司 | Method for automatically constructing user portrait by utilizing data analysis |
CN111159763A (en) * | 2019-12-26 | 2020-05-15 | 银江股份有限公司 | System and method for analyzing portrait of law-related personnel group |
Also Published As
Publication number | Publication date |
---|---|
CN108647729B (en) | 2022-03-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106779087B (en) | A kind of general-purpose machinery learning data analysis platform | |
CN108647729A (en) | A kind of user's portrait acquisition methods | |
CN111460252B (en) | Automatic search engine method and system based on network public opinion analysis | |
CN107368700A (en) | Based on the microbial diversity interaction analysis system and method for calculating cloud platform | |
Yang et al. | Prototype-based image search reranking | |
CN107818130A (en) | The method for building up and system of a kind of search engine | |
CN110737805B (en) | Method and device for processing graph model data and terminal equipment | |
CN110046264A (en) | A kind of automatic classification method towards mobile phone document | |
Cong | Personalized recommendation of film and television culture based on an intelligent classification algorithm | |
CN111460145A (en) | Learning resource recommendation method, device and storage medium | |
CN111159559A (en) | Method for constructing recommendation engine according to user requirements and user behaviors | |
Estivill-Castro et al. | Improving binary classification of web pages using an ensemble of feature selection algorithms | |
CN116401338A (en) | Design feature extraction and attention mechanism based on data asset intelligent retrieval input and output requirements and method thereof | |
CN112506930B (en) | Data insight system based on machine learning technology | |
Bhanot et al. | Python vs. R: A Text Mining Approach for analyzing the Research Trends in Scopus Database | |
CN113408207A (en) | Data mining method based on social network analysis technology | |
Ma | The Research of Stock Predictive Model based on the Combination of CART and DBSCAN | |
CN110059316B (en) | Dynamic scientific and technological resource semantic analysis method based on data perception | |
Bochkaryov et al. | Application of the ensemble clustering algorithm in solving the problem of segmentation of users taking into account their loyalty | |
Miksatko et al. | What’s in a cluster? automatically detecting interesting interactions in student e-discussions | |
Wang et al. | Digital Management Mode of Enterprise Human Resources under the Background of Digital Transformation | |
Chen et al. | Building a training dataset for classification under a cost limitation | |
Kumar | Business intelligence with and without data mining-A comparative study | |
Liu | Construction of personalized recommendation system of university library based on SOM neural network | |
Wilhelm | Data and knowledge mining |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |