CN108647729B - User portrait acquisition method - Google Patents
User portrait acquisition method Download PDFInfo
- Publication number
- CN108647729B CN108647729B CN201810455403.3A CN201810455403A CN108647729B CN 108647729 B CN108647729 B CN 108647729B CN 201810455403 A CN201810455403 A CN 201810455403A CN 108647729 B CN108647729 B CN 108647729B
- Authority
- CN
- China
- Prior art keywords
- user
- behavior
- feature
- data
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Abstract
The invention particularly relates to a user portrait acquisition method which can better establish a user classification model and a user behavior model by classifying and analyzing basic data of a user and user behavior logs corresponding to the basic data respectively, and acquire user portraits of different users through the two models.
Description
Technical Field
The invention relates to the field of information classification processing, in particular to a user portrait acquisition method.
Background
The user portrait is also called a user role and is an effective tool for delineating target users and connecting user appeal and design direction, and the user portrait is widely applied to various fields.
In the scientific and technological resource supply and demand docking, the scientific and technological achievements of the supplier and the demander are huge, and for the supplier, the scientific and technological resource data are detailed and huge, so that the scientific and technological achievement display is clear and clear, but the output of the scientific and technological achievement is weak, which is determined by the scientific and technological resource supply and demand docking mode. In most cases, the requirements of both the supply and demand sides of the scientific and technological resources cannot be matched. The main reason is that the information owned by the supplier and the demander is not equal, on one hand, the supplier cannot subdivide the occupied scientific and technological resources according to the market demand and cannot quickly know the desire of the demander; on the other hand, the demander does not describe his own requirements exhaustively or the described characteristics of the requirements differ considerably from the supplier's conception. This makes it extremely difficult to interface the scientific resources of both the supply and demand parties. When the supply and demand parties complete the result of sufficient preparation work, the scientific and technological resource butt joint can be completed, and therefore the supply and demand butt joint efficiency of the scientific and technological resources is greatly reduced. Even if insufficient preparation work is available, the demander needs to search and research for many times to know the information of the wanted supplier, meanwhile, in the continuous search of the demander, the searching formula used by the demander is provided by the supplier, the searching formula is not fine enough, meanwhile, the index construction does not comply with the will of the demander, and the demander is not friendly, so that the supply and demand docking of the scientific resources is not convenient.
The classification and accurate role positioning of the users are the first step of optimizing the scientific and technological resource supply and demand docking method because the classification of the users is fuzzy and the scientific and technological resources cannot be corresponded to the user categories, so that the allocation of the scientific and technological resource supply and demand is uneven. In summary, the problem of classification of scientific and technological resource users and user portrait acquisition is to be solved urgently.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a user portrait acquisition method which can better classify users, analyze the behaviors of the users and acquire user portraits according to the classification and analysis results.
In view of the above object, the present invention is achieved as follows: a user portrait acquisition method is realized based on basic data of a plurality of users and a user behavior log corresponding to the basic data, wherein the user behavior log comprises user resource supply and demand behavior data, and is characterized by comprising the following steps:
s1, extracting characteristic information of each datum in the basic data of all users, carrying out cluster analysis on the basic data of the same type by using the corresponding characteristic information, and obtaining a plurality of corresponding first characteristic sets; establishing a user classification model according to all the first feature sets;
s2, extracting feature information of each data in the user resource supply and demand behavior data, performing cluster analysis on the same type of user resource supply and demand behavior data by using the corresponding feature information to obtain a plurality of corresponding second feature sets, and establishing a user behavior model according to all the second feature sets; and establishing a behavior data feature set according to all the first feature set, the second feature set and the user classification model.
S3, taking the behavior data feature set as a training sample of the user behavior model, and accordingly establishing the user behavior model;
and S4, acquiring the user portrait according to the user classification model and the user behavior model.
By classifying and analyzing the basic data of the user and the user behavior logs corresponding to the basic data respectively, a user classification model and a user behavior model can be well established, and user figures of different users can be obtained through the two models. The obtained user behavior model provides a basis for classification analysis for subsequent user behavior analysis. Because the user behavior data belongs to dynamic data which is increased along with time, the user behavior is continuously analyzed subsequently, and when the behavior characteristic set is optimized, similar behavior data are directly classified and stored for analysis, the method has the advantages that on one hand, redundant data are screened out, the initial behavior data are labeled, and the processing difficulty is reduced; on the other hand, according to the analysis result, the initial behavior feature classification is optimized, so that the classification model is more accurate.
Further, the specific forming process of the first feature set in step S1 is as follows:
s1.1, randomly extracting basic data of the same type, and performing cluster analysis on the extracted data to obtain a plurality of characteristic values Mi;
s1.2, classifying the basic data of the same type, then performing layered sampling, and performing cluster analysis on the sampled data to obtain a plurality of characteristic values Mk;
s1.3, optimizing the Mi according to the similarity of the Mi and the Mk, and finally obtaining a plurality of characteristic values M to form a first characteristic set.
Further, the basic data comprises primary data describing the role characteristics of the user and secondary data describing the supply and demand conditions of the user resources; all the first feature sets in step S1 include a third feature set formed by the primary data and a fourth feature set formed by the secondary data, the fourth feature set is analyzed according to the third feature set to obtain an index between them, and a user classification model is established according to the index.
Further, step S1.1 is repeatedly executed for a plurality of times, and the eigenvalue Mi obtained in the repeated execution process is optimized according to the sampling times, the sampling proportion, and the cluster analysis process.
Further, the specific establishing method of the behavior data feature set in step S2 is as follows:
s2.1, the second feature set, the third feature set and the fourth feature set respectively comprise a plurality of feature values, each feature value corresponds to a corresponding number of similar users, and corresponding weights are obtained according to different numbers of users corresponding to the feature values; sampling the users corresponding to each characteristic value according to different weights, and calculating the similarity of the users respectively sampled from the second characteristic set and the fourth characteristic set to obtain a similar characteristic index Q1;
s2.2, according to the Q1 and the user classification model, carrying out similarity analysis on the users correspondingly sampled from each feature value in the first feature set to obtain a similar feature index Q2;
s2.3, establishing a behavior data feature set according to Q1 and Q2.
Further, before step S4, the user classification model is further optimized, and the specific steps are as follows: acquiring user resource supply and demand behavior data in real time, firstly, classifying and analyzing the user resource supply and demand behavior data according to a user behavior model to obtain behavior characteristics of each type of user set, and correcting a user classification model according to behavior characteristic evolution of each type of user; secondly, classifying and analyzing the user resource supply and demand behavior data according to the behavior characteristics to obtain behavior characteristic values, classifying the users according to the behavior characteristic values, and further correcting the user classification model.
Further, the specific steps of correcting the user classification model for the first time are as follows: analyzing the dynamic behavior data of each type of user according to the user classification model to obtain a behavior characteristic evolution model; the behavior feature evolution model comprises the change of the feature values and the corresponding weights, and the user classification model and the user data feature set are corrected according to the behavior data feature set and the behavior feature evolution model.
Furthermore, each behavior characteristic value corresponds to a corresponding number of users, and corresponding weights are obtained according to different numbers of users corresponding to the behavior characteristic values; and hierarchically sampling the behavior data feature set according to the weight corresponding to the behavior feature value, analyzing the sampled sample to obtain a second behavior feature correction factor, and correcting the weight of the first feature set again by using the second behavior feature correction factor.
Compared with the prior art, the invention has the beneficial effects that: by classifying and analyzing the basic data of the user and the user behavior logs corresponding to the basic data respectively, a user classification model and a user behavior model can be well established, and user figures of different users can be obtained through the two models.
Drawings
FIG. 1 is a principal flow diagram of the process of the present invention.
FIG. 2 is a flowchart illustrating the detailed formation process of the first feature set in step S1 according to the present invention.
Fig. 3 is a flowchart of a specific establishing method of the behavior data feature set in step S2 according to the present invention.
Detailed Description
The present invention is explained in detail based on the following examples and the accompanying drawings.
A user portrait acquisition method is realized based on basic data of a plurality of users and user behavior logs corresponding to the basic data, wherein the basic data comprises primary data describing user role characteristics and secondary data describing user resource supply and demand conditions, and the user behavior logs comprise user resource supply and demand behavior data. The primary data comprises identity information (such as working years, jobs, sexes, team basic information and the like), scientific research fields, scientific research achievements, potential research dynamics and the like, and the secondary data comprises supply and demand intentions and the like of users. The user resource supply and demand behavior data comprises time attribute data, geographic attribute data and user operation attribute data classified according to actions, and the user operation data classified according to actions comprises searching, collecting, trading, consulting and the like.
The method comprises the following steps as shown in figure 1:
s1, extracting feature information of each datum in basic data of all users in a keyword mode, carrying out clustering analysis on the basic data of the same type by using corresponding feature information through a K-means algorithm, and obtaining a plurality of corresponding first feature sets, wherein a K value used by the K-means algorithm is manually selected; establishing a user classification model according to all the first feature sets;
in the scientific and technological resource supply and demand docking example, the identity information is used as a primary key of a data feature set, and the data feature set comprises industry background, working years, jobs, team information and the like. The identity information data is combined with other user basic data, such as scientific research results, research dynamics and the like, to perform characteristic analysis. Firstly, extracting an industry background keyword, and counting the average working years and standard deviation of the keyword under the industry background; then, according to the industry background characteristics, carrying out characteristic analysis on the scientific research achievements, research dynamics and the like of the user; and finally, obtaining the industry maturity and the industry basic characteristics according to the two steps of results. The users are classified once through the basic characteristics of the industry.
S2, extracting feature information of each data in the user resource supply and demand behavior data, performing cluster analysis on the same type of user resource supply and demand behavior data by using the corresponding feature information to obtain a plurality of corresponding second feature sets, and establishing a user behavior model according to all the second feature sets; and establishing a behavior data feature set according to all the first feature set, the second feature set and the user classification model.
In the scientific and technological resource supply and demand docking example, the user resource supply and demand behavior data comprises supply and demand data. The scientific and technological resource information occupied by the user is called user resource supply and demand data, and comprises instruments, technical means, patent methods and the like which can be disclosed by the user. Technical schemes, instruments and equipment, actual scene solutions and the like which need to be provided by users belong to user resource demand data. The user resource supply and demand data generally comprises instrument data, patent data, application scheme data and the like, the data needs to be processed through a general natural semantic analysis method to obtain keywords in the data, and the role characteristics of the supplier user are further obtained through characteristic analysis. The user resource demand data generally includes specific demand information, such as required instruments, technical means, and the like, and also includes some fuzzy demand information. Carrying out characteristic analysis on the specific demand information to obtain a first character characteristic of the demand user; for fuzzy demand information, such as description of an application scene or a demand purpose, keywords are manually extracted from the perspective of an application object, and second role characteristics of a demand party user are obtained through characteristic analysis;
s3, taking the behavior data feature set as a training sample of the user behavior model, and accordingly establishing the user behavior model;
and S4, acquiring the user portrait according to the user classification model and the user behavior model.
By classifying and analyzing the basic data of the user and the user behavior logs corresponding to the basic data respectively, a user classification model and a user behavior model can be well established, and user figures of different users can be obtained through the two models.
Preferably, as shown in fig. 2, the specific forming process of the first feature set in step S1 is as follows:
s1.1, randomly extracting basic data of the same type, and performing cluster analysis on the extracted data to obtain a plurality of characteristic values Mi;
s1.2, classifying the basic data of the same type, then performing layered sampling, and performing cluster analysis on the sampled data to obtain a plurality of characteristic values Mk;
s1.3, optimizing the Mi according to the similarity of the Mi and the Mk, and finally obtaining a plurality of characteristic values M to form a first characteristic set.
Preferably, all the first feature sets in step S1 include a third feature set formed by the primary data and a fourth feature set formed by the secondary data, the fourth feature set is analyzed according to the third feature set to obtain an index between them, and a user classification model is established according to the index.
Preferably, step S1.1 is repeatedly executed for a plurality of times, and the eigenvalue Mi obtained during the repeated execution optimizes the eigenvalue Mi obtained during the first execution according to the information such as the number of samples, the sampling ratio, the cluster dispersion, the cluster evolution, and the like. Clustering is the data aggregation area formed by dividing in the process of cluster analysis.
Preferably, as shown in fig. 3, the specific establishing method of the behavioral data feature set in step S2 is as follows:
s2.1, the second feature set, the third feature set and the fourth feature set respectively comprise a plurality of feature values, each feature value corresponds to a corresponding number of similar users, and corresponding weights are obtained according to different numbers of users corresponding to the feature values; sampling the users corresponding to each characteristic value according to different weights, and calculating the similarity of the users respectively sampled from the second characteristic set and the fourth characteristic set to obtain a similar characteristic index Q1;
in the scientific and technological resource supply and demand docking example, the purpose of the user behavior is to solve the supply and demand problem, the secondary data describing the supply and demand conditions of the user resource is closely related to the supply and demand behavior data of the user resource, and therefore the second collection obtained by the behavior data needs to be docked with the fourth feature obtained by the supply and demand data of the user resource.
S2.2, according to the Q1 and the user classification model, carrying out similarity analysis on the users correspondingly sampled from each feature value in the first feature set to obtain a similar feature index Q2;
s2.3, establishing a behavior data feature set according to Q1 and Q2.
Preferably, before step S4, the user classification model is further optimized, and the specific steps are as follows: acquiring user resource supply and demand behavior data in real time, firstly, classifying and analyzing the user resource supply and demand behavior data according to a user behavior model to obtain behavior characteristics of each type of user set, and correcting a user classification model according to behavior characteristic evolution of each type of user; secondly, classifying and analyzing the user resource supply and demand behavior data according to the behavior characteristics to obtain behavior characteristic values, classifying the users according to the behavior characteristic values, and further correcting the user classification model.
Firstly, the user resource supply and demand behavior data is dynamic data which increases along with time; secondly, no clear classification standard exists because a perfect scientific and technological resource user behavior analysis model is not established. Therefore, the user resource supply and demand behavior data need to be classified, specific contents of different types of behaviors need to be analyzed, so as to obtain the user behavior characteristic values, and in the analysis process, because the user resource supply and demand behavior data have higher complexity and extremely large volume, a neural network algorithm needs to be used for extracting and classifying the behavior contents, so as to reduce behavior content keywords and find the most main behavior characteristic values of various users.
Preferably, the specific step of first correcting the user classification model is as follows: analyzing the dynamic behavior data of each type of user according to the user classification model to obtain a behavior characteristic evolution model; the behavior feature evolution model comprises the change of the feature values and the corresponding weights, and the user classification model and the user data feature set are corrected according to the behavior data feature set and the behavior feature evolution model.
The user dynamic behavior data is obtained from a scientific and technological resource docking portal website and generally comprises browsing records, searching records, transaction records, consultation records and the like, and each behavior data comprises information such as time, place, motivation, results and the like. Time generally refers to the length of time that the action is executed; location generally refers to the place where the action is performed; the motivation generally refers to a behavior characteristic value of the user before the behavior, and is called motivation characteristic, and the motivation characteristic is obtained through the behavior characteristic before the behavior and the current role characteristic of the user; the result is the evaluation of the user's behavior, mainly through the analysis of the relation in the user's whole behavior process. For example, a user inputs a keyword 'fluorescence microscope' related to instrument information for searching 'fluorescence microscope' supply and demand information, a portal lists all sharable 'fluorescence microscopes', the user browses the information, selects a certain 'fluorescence microscope' through consultation and completes a transaction, and then the user behavior is finished. According to the behavior of the user, the specific analysis steps are as follows: 1. acquiring a user ID, searching user data information according to the user ID, determining the characteristic information of the user according to a user data characteristic set, determining the user category according to the characteristic information, and then acquiring the behavior data characteristics of the user of the category; 2. recording user behavior information, wherein according to examples, searching behaviors and searching contents 'fluorescence' and 'microscope', browsing behaviors and browsing duration, browsing data volume, consultation behavior and consultation content records, and transaction behaviors and transaction detailed information; 3. semantic and characteristic analysis is carried out on the user behavior information to obtain a current behavior characteristic value of the user, the characteristic value is compared with the behavior characteristics of the user, so that the user characteristic model is optimized, and further the behavior characteristic change trend of the user can be obtained from the characteristic value along with the increase of the user behavior data volume, so that the user data model is further optimized; 4. the part of work supervises the user characteristic correction mode about the user motivation characteristic acquisition and behavior evaluation, and the motivation characteristic value and the behavior evaluation are determined by the user role characteristics and the user behavior complexity.
Preferably, each behavior characteristic value corresponds to a corresponding number of users, and the corresponding weight is obtained according to different numbers of users corresponding to the behavior characteristic values; and hierarchically sampling the behavior data feature set according to the weight corresponding to the behavior feature value, analyzing the sampled sample to obtain a second behavior feature correction factor, and correcting the weight of the first feature set again by using the second behavior feature correction factor.
The weights are determined by data bias, cluster crossing complexity, and feature level.
Claims (7)
1. A user portrait acquisition method is realized based on basic data of a plurality of users and a user behavior log corresponding to the basic data, wherein the user behavior log comprises user resource supply and demand behavior data, and is characterized by comprising the following steps:
s1, extracting characteristic information of each datum in the basic data of all users, carrying out cluster analysis on the basic data of the same type by using the corresponding characteristic information, and obtaining a plurality of corresponding first characteristic sets; establishing a user classification model according to all the first feature sets;
s2, extracting feature information of each data in the user resource supply and demand behavior data, performing cluster analysis on the same type of user resource supply and demand behavior data by using the corresponding feature information to obtain a plurality of corresponding second feature sets, and establishing a user behavior model according to all the second feature sets; establishing a behavior data feature set according to all the first feature sets, the second feature sets and the user classification model;
s3, taking the behavior data feature set as a training sample of the user behavior model, and accordingly establishing the user behavior model;
s4, acquiring a user portrait according to the user classification model and the user behavior model;
before step S4, the user classification model is further optimized, and the specific steps are as follows: acquiring user resource supply and demand behavior data in real time, firstly, classifying and analyzing the user resource supply and demand behavior data according to a user behavior model to obtain behavior characteristics of each type of user set, and correcting a user classification model according to behavior characteristic evolution of each type of user; secondly, classifying and analyzing the user resource supply and demand behavior data according to the behavior characteristics to obtain behavior characteristic values, classifying the users according to the behavior characteristic values, and further correcting the user classification model.
2. The method of claim 1, wherein the step S1 includes the steps of:
s1.1, randomly extracting basic data of the same type, and performing cluster analysis on the extracted data to obtain a plurality of characteristic values Mi;
s1.2, classifying the basic data of the same type, then performing layered sampling, and performing cluster analysis on the sampled data to obtain a plurality of characteristic values Mk;
s1.3, optimizing the Mi according to the similarity of the Mi and the Mk, and finally obtaining a plurality of characteristic values M to form a first characteristic set.
3. The method of claim 1, wherein the basic data includes primary data describing the character of the user and secondary data describing the supply and demand condition of the user resource; all the first feature sets in step S1 include a third feature set formed by the primary data and a fourth feature set formed by the secondary data, the fourth feature set is analyzed according to the third feature set to obtain an index between them, and a user classification model is established according to the index.
4. The method as claimed in claim 2, wherein step S1.1 is repeated a plurality of times, and the eigenvalue Mi obtained in the repeated execution process optimizes the eigenvalue Mi obtained in the first execution according to the sampling times, the sampling proportion and the cluster analysis process.
5. A method for capturing a user portrait according to claim 3, wherein the specific establishment method of the behavior data feature set in step S2 is as follows:
s2.1, the second feature set, the third feature set and the fourth feature set respectively comprise a plurality of feature values, each feature value corresponds to a corresponding number of similar users, and corresponding weights are obtained according to different numbers of users corresponding to the feature values; sampling the users corresponding to each characteristic value according to different weights, and calculating the similarity of the users respectively sampled from the second characteristic set and the fourth characteristic set to obtain a similar characteristic index Q1;
s2.2, according to the Q1 and the user classification model, carrying out similarity analysis on the users correspondingly sampled from each feature value in the first feature set to obtain a similar feature index Q2;
s2.3, establishing a behavior data feature set according to Q1 and Q2.
6. A method as claimed in claim 1, wherein the step of first modifying the user classification model comprises: analyzing the dynamic behavior data of each type of user according to the user classification model to obtain a behavior characteristic evolution model; the behavior feature evolution model comprises the change of the feature values and the corresponding weights, and the user classification model and the user data feature set are corrected according to the behavior data feature set and the behavior feature evolution model.
7. A user representation retrieval method as claimed in claim 6 wherein each of the behavioral characteristic values corresponds to a corresponding number of users, the corresponding weights being derived from the different numbers of users corresponding to the behavioral characteristic values; and hierarchically sampling the behavior data feature set according to the weight corresponding to the behavior feature value, analyzing the sampled sample to obtain a second behavior feature correction factor, and correcting the weight of the first feature set again by using the second behavior feature correction 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 CN108647729A (en) | 2018-10-12 |
| CN108647729B true 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) |
Families Citing this family (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 |
| CN109872242B (en) * | 2019-01-30 | 2020-10-13 | 北京字节跳动网络技术有限公司 | Information pushing method and device |
| 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 |
| CN111159763B (en) * | 2019-12-26 | 2022-05-31 | 银江技术股份有限公司 | System and method for analyzing portrait of law-related personnel group |
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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108647729A (en) | 2018-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108647729B (en) | User portrait acquisition method | |
| Amar et al. | Low-level components of analytic activity in information visualization | |
| US20040177069A1 (en) | Method for fuzzy logic rule based multimedia information retrival with text and perceptual features | |
| WO2019129520A1 (en) | Systems and methods for combining data analyses | |
| CN110737805B (en) | Method and device for processing graph model data and terminal equipment | |
| CN111897528B (en) | Low-code platform for enterprise online education | |
| US20170323316A1 (en) | Method for Documenting a Customer's Journey Using an Online Survey Platform | |
| CN112508743B (en) | Technology transfer office general information interaction method, terminal and medium | |
| US20220327398A1 (en) | Technology maturity judgment method and system based on science and technology data | |
| CN111143689A (en) | Method for constructing recommendation engine according to user requirements and user portrait | |
| Groen et al. | Is there Really a Need for Using NLP to Elicit Requirements? A Benchmarking Study to Assess Scalability of Manual Analysis. | |
| CN107239964A (en) | User is worth methods of marking and system | |
| CN111159559A (en) | Method for constructing recommendation engine according to user requirements and user behaviors | |
| CN109165351B (en) | Service component search recommendation method based on semantics | |
| CN110442736A (en) | A kind of semantically enhancement subspace cross-media retrieval method based on quadratic discriminatory analysis | |
| KR100557874B1 (en) | Method of scientific information analysis and media that can record computer program thereof | |
| CN113220565B (en) | Method and device for processing crowdsourcing test report | |
| CN114969375A (en) | Method and system for giving artificial intelligence learning to machine based on psychological knowledge | |
| TW202006617A (en) | Cloud self-service analysis platform and analysis method thereof | |
| Jain et al. | Building query optimizers for information extraction: the sqout project | |
| CN113972009A (en) | Medical examination consultation system based on clinical examination medical big data | |
| CN111274331A (en) | Relational data management maintenance system and method | |
| CN111209375A (en) | Universal clause and document matching method | |
| CN112948544B (en) | Book retrieval method based on deep learning and quality influence | |
| CN116089624B (en) | Knowledge graph-based data recommendation method, device and system |
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 |