CN106021377A - Information processing method and device implemented by computer - Google Patents
Information processing method and device implemented by computer Download PDFInfo
- Publication number
- CN106021377A CN106021377A CN201610310902.4A CN201610310902A CN106021377A CN 106021377 A CN106021377 A CN 106021377A CN 201610310902 A CN201610310902 A CN 201610310902A CN 106021377 A CN106021377 A CN 106021377A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- evaluation object
- new
- training
- knowledge graph
- 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
- 230000010365 information processing Effects 0.000 title claims abstract description 10
- 238000003672 processing method Methods 0.000 title claims abstract description 5
- 238000011156 evaluation Methods 0.000 claims abstract description 187
- 238000012549 training Methods 0.000 claims abstract description 100
- 239000013598 vector Substances 0.000 claims abstract description 60
- 238000000034 method Methods 0.000 claims abstract description 39
- 230000008569 process Effects 0.000 claims description 22
- 238000010586 diagram Methods 0.000 claims description 10
- 230000000875 corresponding effect Effects 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 6
- 238000003066 decision tree Methods 0.000 claims description 6
- 238000007637 random forest analysis Methods 0.000 claims description 6
- 230000004044 response Effects 0.000 claims description 6
- 238000012552 review Methods 0.000 claims description 6
- 241000406668 Loxodonta cyclotis Species 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 230000002596 correlated effect Effects 0.000 claims description 2
- 238000012550 audit Methods 0.000 abstract description 2
- 238000013179 statistical model Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 10
- 238000012360 testing method Methods 0.000 description 9
- 230000006399 behavior Effects 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 5
- 230000005855 radiation Effects 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 4
- 238000013178 mathematical model Methods 0.000 description 4
- 238000009412 basement excavation Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000007477 logistic regression Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013075 data extraction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000011158 quantitative evaluation 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
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/25—Integrating or interfacing systems involving database management systems
- G06F16/254—Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- General Business, Economics & Management (AREA)
- Technology Law (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides an information processing method and device implemented by a computer. The method comprises the following steps: acquiring data of a plurality of training evaluation objects, constructing a knowledge map, and creating a map database based on the knowledge map, wherein nodes of the knowledge map are evaluation objects or association objects having an association relation with the evaluation objects, and sides of the knowledge map describe the relation between the nodes and have corresponding weights; calculating feature vectors of the training evaluation objects based on the distances between the training evaluation objects, wherein the distance is at least based on a path weight between the training evaluation objects, and the feature vectors indicate the number of other training evaluation objects in the map databases with different dimensions within a certain distance for the training evaluation objects; using the feature vectors and the evaluation categories of the plurality of training evaluation objects as a sample set, and creating an evaluation module; responding to received data of a new evaluation object, calculating the feature vector and an evaluation index of the new evaluation object; and completing an audit operation of the new evaluation object by the computer according to discrimination conditions based on the evaluation index.
Description
Technical field
It relates to computer information processing field, particularly relate to computer implemented object credibility evaluation method and
Device.
Background technology
At present, personal information is audited and in the applied environment of reliability assessment, such as in social safety needing
Field, financial security field etc., generally by carrying out individual for personal information, Identity Association information, consumer sale information etc.
Portrait, sets up based on statistical machine learning mathematical model, and then carries out Analysis on confidence.
Along with appearance and the rise of social networks of the Internet, just createing for individual now than ever in our society
The information whenever can created in history will how information, there is between data extremely complex and countless ties contact.
But, when utilizing the data model of machine learning, people lack the deep excavation to personal data, the application to data simultaneously
The most limited, thus have impact on individual or the accuracy of the reliability assessment of individual behavior.
Summary of the invention
In order to solve problem suggested above, the disclosure provides a kind of deep exploitation individual's related data to pass through mathematics
Model carries out the computer implemented technical scheme of information processing.
First aspect according to the disclosure, it is provided that a kind of computer implemented information processing method, method includes: acquisition is many
The related data of individual Training valuation object, build knowledge graph, and create the chart database of knowledge based figure, wherein knowledge graph
Each node is evaluation object or has the affiliated partner of incidence relation with evaluation object, the limit Description of Knowledge figure of knowledge graph
Relation between node and there is respective weights；Based on the distance between Training valuation object, calculate each Training valuation pair
The characteristic vector of elephant, wherein distance is at least based on the path weight value between Training valuation object, and characteristic vector is for each training
Evaluation object instruction quantity of other Training valuation objects in the chart database of different dimensions in a certain distance；By feature to
The assessment categories of amount and multiple Training valuation objects is as sample set, to build assessment models；In response to receiving new assessment
The related data of object, calculates the characteristic vector of new evaluation object；Feature based on assessment models and new evaluation object to
Amount, calculates the evaluation index of new evaluation object；And evaluation index of based on new evaluation object, by computer according to differentiation
Condition completes the review operations about new evaluation object.
According to embodiment of the disclosure, wherein the distance between Training valuation object is calculated according to following formula:
Wherein xi、xjRepresent the i-th Training valuation object and jth Training valuation object, D (x respectivelyi,xj) represent xiAnd xjBetween away from
From,Represent xiAnd xjNode on the l of path, L represents xiAnd xjBetween the number in path,Represent xiAnd xjPath l
The weight of top；For the factor of influence of path l,
Represent the maximum path of factor of influence the interstitial content of process.
According to embodiment of the disclosure, wherein based on the distance between Training valuation object, calculate each Training valuation pair
The characteristic vector of elephant includes: for each property value of the particular community of Training valuation object, statistical distance D (xi,xjIt is less than in)
The number of paths of specific range threshold value；And based on the number of paths added up, obtain the characteristic vector of Training valuation object.
According to embodiment of the disclosure, wherein in response to receiving the related data of new evaluation object, calculate new assessment
The characteristic vector of object includes: projected in knowledge graph by the related data of new evaluation object；Based on new evaluation object with
The path weight value between other evaluation objects in knowledge graph, calculate between new evaluation object and other evaluation objects away from
From；And based on the distance between new evaluation object and other evaluation objects, calculate the characteristic vector of new evaluation object.
According to embodiment of the disclosure, wherein assessment models include following in one or more: decision tree, logistic regression
Model, Random Forest model.
According to embodiment of the disclosure, also include updating chart database and assessment models.
According to embodiment of the disclosure, also include: multiple new evaluation objects and corresponding assessment categories are updated and know
Know in figure；Based on the knowledge graph after updating and chart database, generate new assessment models；Determine the accurate of new assessment models
Degree；And accuracy of based on new assessment models, update assessment models.
According to embodiment of the disclosure, wherein the node of figure has the affiliated partner bag of incidence relation with evaluation object
Include in following item is one or more: the personal data object that evaluation object is associated；The collage-credit data that evaluation object is associated
Object；The social relations data object that evaluation object is associated；The social network data object that evaluation object is associated；It is right to assess
As the communication data object being associated.
According to embodiment of the disclosure, wherein the weight on the limit in figure at least based in following item one or more come really
Fixed: the relationship type of the node associated by limit；The correlation degree between node associated by limit.
Second aspect according to the disclosure, it is provided that a kind of computer implemented information processor, device includes: diagram data
Storehouse creates device, is configured to obtain the related data of multiple Training valuation object, builds knowledge graph, and creates knowledge based
The chart database of figure, wherein each node of knowledge graph is evaluation object or have associating of incidence relation with evaluation object right
As, relation between the node of the limit Description of Knowledge figure of knowledge graph and have respective weights；First calculates device, is configured to
Based on the distance between Training valuation object, calculating the characteristic vector of each Training valuation object, wherein distance is at least based on instruction
Practicing the path weight value between evaluation object, characteristic vector is for each Training valuation object instruction different dimensional in a certain distance
The quantity of other Training valuation objects in the chart database of degree；Assessment models construction device, is configured to characteristic vector with many
The assessment categories of individual Training valuation object is as sample set, to build assessment models；Second calculates device, is configured to respond to
In receiving the related data of new evaluation object, calculate the characteristic vector of new evaluation object；And based on assessment models with new
The characteristic vector of evaluation object, calculate the evaluation index of new evaluation object；And examination & verification processing means, be configured to based on
The evaluation index of new evaluation object, is completed the review operations about new evaluation object by computer according to criterion.
According to embodiment of the disclosure, wherein the distance between Training valuation object is calculated according to following formula:
Wherein xi、xjRepresent the i-th Training valuation object and jth Training valuation object, D (x respectivelyi,xj) represent xiAnd xjBetween away from
From,Represent xiAnd xjNode on the l of path, L represents xiAnd xjBetween the number in path,Represent xiAnd xjPath l
The weight of top；For the factor of influence of path l,
Represent the maximum path of factor of influence the interstitial content of process.
According to embodiment of the disclosure, first calculates device is also configured to the particular community for Training valuation object
Each property value, statistical distance D (xi,xjLess than the number of paths of specific range threshold value in)；And based on the number of path added up
Amount, obtains the characteristic vector of Training valuation object.
According to embodiment of the disclosure, second calculates device is also configured to throw the related data of new evaluation object
It is mapped in knowledge graph；Based on the path weight value between other evaluation objects in new evaluation object and knowledge graph, calculate new
Distance between evaluation object and other evaluation objects；And based between new evaluation object and other evaluation objects away from
From, calculate the characteristic vector of new evaluation object.
According to embodiment of the disclosure, wherein assessment models include following in one or more: decision tree, logistic regression
Model, Random Forest model.
According to embodiment of the disclosure, also include updating device, be configured to update chart database and assessment models.
According to embodiment of the disclosure, updating device is also configured to multiple new evaluation objects and corresponding assessment
Classification updates in knowledge graph；Based on the knowledge graph after updating and chart database, generate new assessment models；Determine new assessment
The accuracy of model；And accuracy of based on new assessment models, update assessment models.
According to embodiment of the disclosure, wherein the node of figure has the affiliated partner bag of incidence relation with evaluation object
Include in following item is one or more: the personal data object that evaluation object is associated；The collage-credit data that evaluation object is associated
Object；The social relations data object that evaluation object is associated；The social network data object that evaluation object is associated；It is right to assess
As the communication data object being associated.
According to embodiment of the disclosure, wherein the weight on the limit in figure at least based in following item one or more come really
Fixed: the relationship type of the node associated by limit；The correlation degree between node associated by limit.
The disclosure sets up assessment models by having the knowledge graph of radiation feature, by data by one-dimensional characteristic vector exhibition
Open as model based on figure relation, and data are converted into machine are appreciated that data mode, utilize computer to realize data message
Process, take full advantage of the advantage of data mining and machine learning.
Accompanying drawing explanation
In conjunction with accompanying drawing and with reference to described further below, the feature of the presently disclosed embodiments, advantage and other aspects will become
Must become apparent from, existing by only by example, nonrestrictive mode shows some embodiments of the disclosure, in the accompanying drawings:
Fig. 1 shows and processes block diagram according to the schematic information that embodiment of the disclosure；
Fig. 2 shows according to the method flow diagram that embodiment of the disclosure；
Fig. 3 shows according to the example knowledge graph that embodiment of the disclosure；And
Fig. 4 shows according to the device schematic diagram that embodiment of the disclosure.
Detailed description of the invention
Each exemplary embodiment of the disclosure is described in detail below with reference to accompanying drawing.Flow chart and block diagram in accompanying drawing illustrate
Architectural framework in the cards, function and the operation of the method and system of the various embodiments according to the disclosure.It should be noted that,
Each square frame in flow chart or block diagram can represent a module, program segment or a part for code, described module, program
Section or the part of code can include one or more for realizing holding of the logic function of defined in each embodiment
Row instruction.It should also be noted that in some realization alternately, the function marked in square frame can also be attached according to being different from
The order marked in figure occurs.Such as, two square frames succeedingly represented can essentially perform substantially in parallel, or it
Sometimes can also perform in a reverse order, this depends on involved function.It should also be noted that flow chart
And/or the combination of the square frame in each square frame in block diagram and flow chart and/or block diagram, it is possible to use perform the merit of regulation
The special hardware based system that can or operate realizes, or specialized hardware can be used to come with the combination of computer instruction
Realize.
Term as used herein " includes ", " comprising " and similar terms are understood to the term of opening, i.e.
" include/including but not limited to ", expression can also include other guide.Term "based" is " being based at least partially on ".Term
" embodiment " expression " at least one embodiment "；Term " another embodiment " expression " at least one further embodiment ".
As it was previously stated, the disclosure is intended to deep excavation information data based on incidence relation, utilize the radiation relation of information
Feature, and then it is based on statistical model to individual or the reliability assessment of individual behavior to use computer to realize.Computer realizes
Statistical model algorithm (i.e. machine learning algorithm) be that a class automatically analyzes from data and obtains rule, and assimilated equations is to the unknown
The algorithm that data are predicted.In this article, term " assessment models " is mathematical model based on statistics, and it can be used for commenting
Estimate object to be estimated.Term " evaluation object " for the needs determined according to various different application scenes be estimated right
As, such as individual or individual behavior, in an embodiment of the disclosure, evaluation object can be the loaning bill needing to carry out auditing
Or loan, in another embodiment, evaluation object can be the personal information needing examination & verification.Term " characteristic vector " represents basis
The characteristic information extracted from sample (such as, evaluation object) and the multi-C vector built, the dimension of vector is equal to characteristic information
The number of type.
It is appreciated that these exemplary embodiments described below merely to enable those skilled in the art the most geographical
Solve and realize embodiment of the disclosure, and limiting the scope of the present disclosure the most by any way.
Fig. 1 shows schematic information processing block Figure 100 of an embodiment according to the disclosure.It it is noted that these public affairs
The information processing solution that dispersing and elevation goes out does not limits concrete application scenarios, it will be understood by those skilled in the art that at needs
Client or user profile or client or user behavior information are analyzed in the control environment processed, embodiment of the disclosure
All can be applied.
Figure 1 illustrates the data carried out by making full use of the relatedness of data to process and assessment models foundation
Process.Fig. 1 shows the aspect that data source or data relate to: user data 101, the Internet public data 102, historical data
103, these data can be stored in one or more data base.Generally, the number that user data 101 provides for user self
According to, such as age, education degree, working condition, family background, communication data etc..The Internet public data 102 is in social activity
Disclosed, any data can checked per capita, such as social network data, web data etc. in network.Historical data 103 is
Apply the mechanism of this programme or the historical relevance data that system is known, such as, in the scene of financial institution, can be consumption
Data, bank data, collage-credit data etc., and in the scene of search application, can be the historical search data etc. of user.
Fig. 1 has been diagrammatically only by the aspect that data may relate to, it is possible to understand that, it is possible to make full use of the potential data aspect of the relatedness of data
The most within the scope of the present disclosure.
Data extraction process 104 can such as gather various data source, and data source can include such as user data 101, mutually
Networking public data 102, historical data 103 etc., extraction wherein can data, to form chart database based on figure relation
105, and and then generate training data 107.Therefore, in data extraction process 104, to training object and the pass that is mutually related thereof
Connection object and related data carry out deep discovery and excavation.These training objects and the affiliated partner formation figure that is mutually related close
The node of system, each node is identified by a globally unique ID and is indexed, and node has several attributes simultaneously, to describe
The characteristic of node.Such as, when node is individual, its attribute can be name, identification card number, home address etc.；At node
During for object for sale, its attribute can be to sell type, sell amount of money etc.；When node is phone, its attribute can be electricity
Call type, telephone number, institute possession etc..And the relation between these nodes is described by the limit of figure, e.g. is-a closes
System, represents that a node is the one of another node, or has-a relation, represents that a node has another node etc.
Deng, such relation is all used to describe the incidence relation between entity.Meanwhile, this relation needs to be converted into machine and can manage
The expression solved, uses the weight on limit to characterize correlation degree in an embodiment of the disclosure.
Chart database 105 stores various data based on figure relation, such as, node as above, node ID, genus
Property, limit, the relation on limit, weight etc..Information processing system can utilize computer to carry out various to the data in chart database
Process, such as based on figure relation traversal, add up, search, renewal etc..
Data in chart database 105 are analyzed, extract and process by relation analysis and extraction process 106, to be formed
Training data 107.According to the training data of required acquisition, based on chart database is trained the various incidence relations of object, carry
Take out the characteristic vector of training object, and training object is carried out label, be input to assessment models as sample data and trained
Journey 108, finally gives assessment models 109.
In assessment models training process 108, the sample of the label chosen is divided into training set and test set, such as
Can be using the 75% of sample as training set, the 25% of sample is as test set.The sample data of training set is inputted statistics mould
In type, building statistical model, such as statistical model is that GBDT (Gradient Boost Decision Tree) model, logic are returned
Return model etc., and continuous iteration sample in the process, until training terminates.The statistics afterwards test sample input built
Model, carries out test checking and can adjust the parameter of statistical model, is finally completed assessment models training process 108, is commented
Estimate model 109.When there being new evaluation object to need to be estimated, pass through the data analysis to this new evaluation object and extraction,
Its characteristic vector is inputted assessment models 109, to be calculated its evaluation index.
It is described in detail below in conjunction with specific embodiment technical scheme of this disclosure.Fig. 2 shows according to these public affairs
The flow chart of the method 200 of the embodiment opened.In step 202, obtain the related data of multiple Training valuation object, build
Knowledge graph, and create the chart database of knowledge based figure, wherein each node of knowledge graph be evaluation object or with assessment
Object has the affiliated partner of incidence relation, relation between the node of the limit Description of Knowledge figure of knowledge graph and have corresponding power
Weight.
For the ease of the purpose understood, in a specific embodiment, such as, for personal credit's scene, node definition shows
Anticipate as shown in the table.
Table 1
As shown in table 1, in such scene, node can be evaluation object (such as loan) or evaluation object is correlated with
The affiliated partner (such as bank card, loan distribution person, Email etc.) of connection, it addition, as seen from Table 1, each node has
Several attributes corresponding.Based on several evaluation objects and the affiliated partner that is associated thereof, build knowledge graph, in such field
Jing Zhong, the limit of figure such as can be as shown in table 2.
Table 2
As shown in table 2, based on the relation between the node described by limit and the type of associated nodes, can be that each limit is divided
Joining corresponding weighted value, this weighted value characterizes the correlation degree between node, or in different application scene, characterize for
The influence degree of the object of needs assessment, it is quantificational expression.For example, between the object if desired assessed by man and wife,
The relation such as colleague connects, then this contact generally can have the higher power that interacts, can be allocated one higher
Weighted value；If being connected by such as examination & approval behavior, i.e. one approver has examined two objects, and this contact is generally of relatively low
The power that interacts, a relatively low weighted value can be allocated.
According to the node extracted from data source and the related data on limit, knowledge graph model can be built.A concrete reality
Execute in example, construct knowledge graph signal 300 as shown in Figure 3.As it is shown on figure 3, have the incidence relation of complexity between each node,
Such as, for node 301 (loan 2), it has direct correlation with node 302 (borrower 2), node 303 (loan distribution person 1)
Relation, the node 305 (borrower 1) simultaneously associated with node 302 (borrower 2), node 306 (father), node 307 (hands
Machine), node 308 (Email), node 309 (company) etc. there is nearer indirect association relation, with node 304 (loan 3),
Phone that node 306 (father) has, company work other people etc. there is indirect association relation farther out.It is true that
In different application scene, the relation aspect of concern is different, can build applicable application scenarios, the most illustrated in Figure 3
Knowledge graph.
Knowledge graph model is stored in chart database 105, can the most such as matrix, adjacency list, chained list etc.
Store, be convenient to the operation such as graph traversal and search.Building of knowledge graph (or the chart database) being appreciated that in the disclosure
Vertical, it is provided that a kind of to utilize the solution of relation radiation characteristic between data, to be preferably applied for assessment or the field of prediction
Scape.
Return to the method 200 shown in Fig. 2, after establishing chart database 105, in step 204, based on Training valuation
Distance between object, calculates the characteristic vector of each Training valuation object, and wherein this distance is at least based on Training valuation object
Between path weight value, characteristic vector for each Training valuation object instruction in a certain distance the diagram data of different dimensions
The quantity of other Training valuation objects in storehouse.
In order to make it easy to understand, still as a example by Fig. 3, node 302 (borrower 2) and node 305 (borrower 1) SEPARATE APPLICATION
The loan represented with node 301 (loan 2) and node 310 (provide a loan 1), from this exemplary plot it can be seen that node 310 is in
Bad credit state, and node 302 (borrower 2) and node 305 (borrower 1) have linked character close together, that weight is higher,
In this case, node 301 (loan 2) have greatly may probability have high can not reliability.In order to reach distinguish and
The target of quantitative evaluation, in an embodiment of the disclosure, by calculating " distance " between evaluation object, by evaluation object
Characteristic vector pickup be the quantity of other evaluation objects in the chart database of different dimensions in a certain distance, and then can
Distinguish the assessment categories of evaluation object.Therefore, this distance can characterize with associate that impact is big, there is insincere (or credible) because of
These objects of the distance of " closely " between the object of element, it should assign to significant assessment categories.
In a specific embodiment of the disclosure, two evaluation object xi、xjBetween distance D (xi,xj) can basis
Following formula is calculated:
WhereinRepresent xiAnd xjNode on the l of path, L represents xiAnd xjBetween the number in path,Represent xiAnd xj
The weight of l top, path；For the factor of influence of path l,
Represent the maximum path of factor of influence the interstitial content of process.
That is, obtain evaluation object distance between any two according to formula (1), this distance is all roads between evaluation object
The path that in footpath, factor of influence is maximum the interstitial content of process.Based on this distance, for each evaluation object, can be according to need
Attribute to be assessed, selects suitable radiation level, obtains the characteristic vector for mathematical model.
In a specific embodiment of the disclosure, such as in personal credit's application scenarios, in order to evaluation object (example
Such as loan) extract characteristic vector, according to node as shown in table 1 and nodal community example, evaluation object (such as can be borrowed
Money) the property value (the most various state) of attribute (such as loan status) do such as the definition of table 3, as shown in table 3, evaluation object
Particular community have 6 property values.
Table 3
| Loan status | Identifier |
| Apply for | applying |
| Refund | current |
| Close | complete |
| Exceed the time limit less than 30 days | M1 |
| Exceed the time limit more than 30 days, less than 90 days | M2M3 |
| Bad credit | default |
Therefore, it can for evaluation object xi, statistical distance less than certain distance threshold value Dis quantity as feature to
Amount, i.e.
cntJ=1,2 ..., N(D(xi,xj) < Dis) (formula 2)
Wherein N represents N number of property value of particular community of evaluation object.Evaluation object x is obtained according to formula (2)iN-dimensional special
Levy vector
Characteristic vectorFor each evaluation object instruction in a certain distance in the chart database of different dimensions other
The quantity of evaluation object.
When during the training of statistical model, the distance between Training valuation object can be calculated in the manner described above
And obtain the characteristic vector of Training valuation object.After assessment models creates, for new evaluation object, it is also possible to according to
Aforesaid way calculates the distance between new evaluation object and other evaluation objects, and obtain the feature of new evaluation object to
Amount.
It is appreciated that the calculating of distance described above and the acquisition of characteristic vector, in concrete computer realizes, can
To use depth-first traversal based on figure and/or breadth first traversal, calculate the interstitial content meeting radiation condition.
Turn again to Fig. 2.It follows that method 200 proceeds to step 206, by characteristic vector and multiple Training valuation object
Assessment categories as sample set, to build assessment models.Such as the description above in relation to Fig. 1, assessment models can be according to answering
Select suitable mathematical model by scene, in a specific embodiment of the disclosure, GBDT model can be used.GBDT model
(or GBDT algorithm) calculating each time is the residual error in order to reduce last calculating, and in order to eliminate residual error, can be in residual error
A new model is set up on the gradient direction reduced.It is to say, in GBDT algorithm, the foundation of each new model be for
Before making, the residual error of model reduces toward gradient direction.
Such as, for GBDT model, for evaluation object xi, its N kind probability distribution over states be probability distribution over states be F1
(x),F2(x),…,FnX (), result is that to belong to the probability of classification k be pk(x).Logistic conversion is as follows:
Result after Logistic is converted, loss function is:
Wherein, ykThe estimated value of the sample data for inputting, as a sample xiWhen belonging to classification k, yk=1, otherwise yk=
0。
Formula (4) is brought into loss function formula (5), and to its derivation, the gradient of loss function can be obtained,
For a sample, optimal gradient is the gradient closer to 0.It is trained drawing assessment mould by above formula
Type.The concrete training process such as description above in association with Fig. 1, is divided into training set and test by the sample of the label chosen
Collection, such as, as test set, can input the 75% of sample the sample data of training set as training set, the 25% of sample
In statistical model, build statistical model, and continuous iteration sample in the process, until training terminates.Afterwards by test sample
The statistical model that input builds, carries out test checking and can adjust the parameter of statistical model, being finally completed assessment models training
Process 108, obtains assessment models 109.
The statistical model being appreciated that in embodiment of the disclosure is not limited to GBDT model, it would however also be possible to employ such as its
His decision-tree model, Logic Regression Models, Random Forest model etc..
After constructing assessment models, method 200 proceeds to step 208, in response to receiving the relevant of new evaluation object
Data, calculate the characteristic vector of new evaluation object.In this step, according to a specific embodiment of the disclosure, needs are worked as
To new evaluation object xnewWhen being predicted or assess, in response to receiving new evaluation object xnewRelated data, by this
A little data projection are in the knowledge graph built, to form the graph structure associated by new evaluation object.Afterwards, can be according to formula
(1), formula (2) and formula (3) calculate new evaluation object xnewCharacteristic vector
In step 210, based on assessment models and the characteristic vector of new evaluation object, calculate the assessment of new evaluation object
Index.In an embodiment of the disclosure, can be by the GBDT model prediction built or evaluate evaluation index p (xnew), p
(xnew) value the highest then represent credibility the lowest.
In step 212, evaluation index based on new evaluation object, computer complete about new according to criterion
The review operations of evaluation object.In this step, computer is according to the criterion preset, such as threshold value (such as Pthreshold、
PhighAnd/or PlowDeng) condition, as evaluation index p (xnew)>PhighShowing that credibility is the highest, computer is refused automatically, i.e. audits
Do not pass through；p(xnew)<PlowShowing with a high credibility, computer accepts automatically, i.e. examination & verification is passed through；Plow<p(xnew)<PhighTime, can
Add in the way of other auditing standardses or get involved manual examination and verification etc. using.
It may be noted that after new evaluation object is made examination & verification, new evaluation object can be removed from knowledge graph,
To ensure the stability of assessment models.In order to preferably provide prediction or assessment result, can comment with on-line study and renewal
Estimate model.After assessment models is disposed, such as, can periodically new data and result be updated in knowledge graph model, by
In the increase of data dimension, model can learn to new feature and be used, to improve its predictablity rate.Data are more
Newly, train and test and terminate after, contrast the accuracy of new assessment models and existing assessment models, and if only if, and new model is accurate
Really rate just disposes new assessment models when improving, and otherwise waits for lower whorl learning training process.
According to embodiment of the disclosure, also provide for a kind of computer implemented information processor 400.As shown in Figure 4, dress
Put 400 to include: chart database creates device 401, be configured to obtain the related data of multiple Training valuation object, build knowledge
Figure, and create the chart database of knowledge based figure, wherein each node of knowledge graph is evaluation object or and evaluation object
There is the affiliated partner of incidence relation, relation between the node of the limit Description of Knowledge figure of knowledge graph and there is respective weights；
First calculates device 402, the distance between being configured to based on Training valuation object, calculates the feature of each Training valuation object
Vector, wherein distance is at least based on the path weight value between Training valuation object, and characteristic vector is for each Training valuation object
Instruction quantity of other Training valuation objects in the chart database of different dimensions in a certain distance；Assessment models construction device
403, it is configured to the assessment categories of characteristic vector and multiple Training valuation object as sample set, to build assessment mould
Type；Second calculates device 404, is configured to respond to receive the related data of new evaluation object, calculates new evaluation object
Characteristic vector；And based on assessment models and the characteristic vector of new evaluation object, the assessment calculating new evaluation object refers to
Mark；And examination & verification processing means 405, it is configured to evaluation index based on new evaluation object, by computer according to differentiating bar
Part completes the review operations about new evaluation object.
According to embodiment of the disclosure, wherein the distance between Training valuation object can be calculated according to following formula:
Wherein xi、xjRepresent the i-th Training valuation object and jth Training valuation object, D (x respectivelyi,xj) represent xiAnd xjBetween away from
From,Represent xiAnd xjNode on the l of path, L represents xiAnd xjBetween the number in path,Represent xiAnd xjPath l
The weight of top；For the factor of influence of path l,
Represent the maximum path of factor of influence the interstitial content of process.
According to embodiment of the disclosure, first calculates device 402 is also configured to the specified genus for Training valuation object
Each property value of property, statistical distance D (xi,xjLess than the number of paths of specific range threshold value in)；And based on the path added up
Quantity, obtains the characteristic vector of Training valuation object.
According to embodiment of the disclosure, second calculates device 404 is also configured to the related data of new evaluation object
Project in knowledge graph；Based on the path weight value between other evaluation objects in new evaluation object and knowledge graph, calculate new
Evaluation object and other evaluation objects between distance；And based between new evaluation object and other evaluation objects away from
From, calculate the characteristic vector of new evaluation object.
According to embodiment of the disclosure, wherein assessment models include following in one or more: decision tree, logistic regression
Model, Random Forest model.
According to embodiment of the disclosure, also include updating device 406, be configured to update chart database and assessment models.
According to embodiment of the disclosure, updating device 406 is also configured to multiple new evaluation objects and comments accordingly
Estimate classification to update in knowledge graph；Based on the knowledge graph after updating and chart database, generate new assessment models；Determine new commenting
Estimate the accuracy of model；And accuracy of based on new assessment models, update assessment models.
According to embodiment of the disclosure, wherein the node of figure has the affiliated partner bag of incidence relation with evaluation object
Include in following item is one or more: the personal data object that evaluation object is associated；The collage-credit data that evaluation object is associated
Object；The social relations data object that evaluation object is associated；The social network data object that evaluation object is associated；It is right to assess
As the communication data object being associated.
According to embodiment of the disclosure, wherein the weight on the limit in figure at least based in following item one or more come really
Fixed: the relationship type of the node associated by limit；The correlation degree between node associated by limit.
By the teaching gone out given in above description and relevant drawings, many modification of the disclosure described herein
Will be appreciated by disclosure those skilled in the relevant art with other embodiment.Therefore, the it being understood that disclosure
Embodiment is not limited to disclosed detailed description of the invention, and modification and other embodiment are intended to be included in this
Within scope of disclosure.Although additionally, above description and relevant drawings are in some example combination form of parts and/or function
Under background, example embodiment is described, it will be appreciated that, parts can be provided by alternate embodiment
And/or the different combinations of function are without departing from the scope of the present disclosure.On this point, such as, with explicitly described above
Different parts and/or other combining form of function be also expected within being in the scope of the present disclosure.Although here
Have employed concrete term, but they only are not intended to limit so that general and illustrative implication uses.
Claims (18)
1. a computer implemented information processing method, described method includes:
Obtain the related data of multiple Training valuation object, build knowledge graph, and create diagram data based on described knowledge graph
Storehouse, each node of wherein said knowledge graph is evaluation object or to have associating of incidence relation with described evaluation object right
As, the limit of described knowledge graph describes the relation between the described node of described knowledge graph and has respective weights；
Based on the distance between described Training valuation object, calculate the characteristic vector of each Training valuation object, wherein said away from
From at least based on the path weight value between described Training valuation object, described characteristic vector is for each described Training valuation object
Instruction quantity of other Training valuation objects in the described chart database of different dimensions in a certain distance；
Using the assessment categories of described characteristic vector and the plurality of Training valuation object as sample set, to build assessment mould
Type；
In response to receiving the related data of new evaluation object, calculate the characteristic vector of described new evaluation object；
Based on described assessment models and the described characteristic vector of described new evaluation object, calculate commenting of described new evaluation object
Estimate index；And
Described evaluation index based on described new evaluation object, by described computer according to criterion complete about described newly
The review operations of evaluation object.
Method the most according to claim 1, the distance between wherein said Training valuation object is calculated according to following formula:
Wherein xi、xjRepresent the i-th Training valuation object and jth Training valuation object, D (x respectivelyi,xj) represent xiAnd xjBetween distance,Represent xiAnd xjNode on the l of path, L represents xiAnd xjBetween the number in path,Represent xiAnd xjPath l on institute
State the weight on limit；For the factor of influence of path l,
Represent the maximum path of factor of influence the interstitial content of process.
Method the most according to claim 2, wherein based on the distance between described Training valuation object, calculates each training
The characteristic vector of evaluation object includes:
For each property value of the particular community of described Training valuation object, add up described distance D (xi,xjLess than specific in)
The number of paths of distance threshold；And
Described number of paths based on statistics, obtains the characteristic vector of described Training valuation object.
Method the most according to claim 1, wherein in response to receiving the related data of new evaluation object, calculate described newly
The characteristic vector of evaluation object include:
The related data of described new evaluation object is projected in described knowledge graph；
Based on the path weight value between other evaluation objects in described new evaluation object and described knowledge graph, calculate described newly
Evaluation object and other evaluation objects described between distance；And
Based on the distance between described new evaluation object and other evaluation objects described, calculate the spy of described new evaluation object
Levy vector.
Method the most according to claim 1, wherein said assessment models include following in one or more: decision tree,
Logic Regression Models, Random Forest model.
Method the most according to claim 1, also includes: update described chart database and described assessment models.
Method the most according to claim 6, also includes:
Multiple described new evaluation objects and corresponding assessment categories are updated in described knowledge graph；
Based on the described knowledge graph after updating and chart database, generate new assessment models；
Determine the accuracy of described new assessment models；And
Described accuracy based on described new assessment models, updates described assessment models.
Method the most according to claim 1, has incidence relation with described evaluation object in the node of wherein said figure
Affiliated partner include in following item one or more: the personal data object that evaluation object is associated；Evaluation object is correlated with
The collage-credit data object of connection；The social relations data object that evaluation object is associated；The social networks number that evaluation object is associated
According to object；The communication data object that evaluation object is associated.
Method the most according to claim 1, the weight on the described limit in wherein said figure is at least based in following item
Or multinomial determine: the relationship type of the node associated by described limit；The correlation degree between node associated by described limit.
10. a computer implemented information processor, described device includes:
Chart database creates device, is configured to obtain the related data of multiple Training valuation object, builds knowledge graph, and creates
Building chart database based on described knowledge graph, each node of wherein said knowledge graph is evaluation object or right with described assessment
As having the affiliated partner of incidence relation, the limit of described knowledge graph describe the relation between the described node of described knowledge graph and
There is respective weights；
First calculates device, the distance between being configured to based on described Training valuation object, calculates each Training valuation object
Characteristic vector, wherein said distance is at least based on the path weight value between described Training valuation object, described characteristic vector pin
To each described Training valuation object instruction other Training valuation in the described chart database of different dimensions in a certain distance
The quantity of object；
Assessment models construction device, is configured to make the assessment categories of described characteristic vector and the plurality of Training valuation object
For sample set, to build assessment models；
Second calculates device, is configured to respond to receive the related data of new evaluation object, calculates described new assessment right
The characteristic vector of elephant；And based on described assessment models and the described characteristic vector of described new evaluation object, calculate described newly
The evaluation index of evaluation object；And
Examination & verification processing means, be configured to described evaluation index based on described new evaluation object, by described computer according to
Criterion completes the review operations about described new evaluation object.
11. devices according to claim 10, the distance between wherein said Training valuation object calculates according to following formula
Arrive:
Wherein xi、xjRepresent the i-th Training valuation object and jth Training valuation object, D (x respectivelyi,xj) represent xiAnd xjBetween distance,Represent xiAnd xjNode on the l of path, L represents xiAnd xjBetween the number in path,Represent xiAnd xjPath l on institute
State the weight on limit；For the factor of influence of path l,
Represent the maximum path of factor of influence the interstitial content of process.
12. devices according to claim 11, described first calculates device is also configured to
For each property value of the particular community of described Training valuation object, add up described distance D (xi,xjLess than specific in)
The number of paths of distance threshold；And
Described number of paths based on statistics, obtains the characteristic vector of described Training valuation object.
13. devices according to claim 10, described second calculates device is also configured to
The related data of described new evaluation object is projected in described knowledge graph；
Based on the path weight value between other evaluation objects in described new evaluation object and described knowledge graph, calculate described newly
Evaluation object and other evaluation objects described between distance；And
Based on the distance between described new evaluation object and other evaluation objects described, calculate the spy of described new evaluation object
Levy vector.
14. devices according to claim 10, wherein said assessment models include following in one or more: decision-making
Tree, Logic Regression Models, Random Forest model.
15. devices according to claim 10, also include updating device, are configured to update described chart database and described
Assessment models.
16. devices according to claim 15, described updating device is also configured to
Multiple described new evaluation objects and corresponding assessment categories are updated in described knowledge graph；
Based on the described knowledge graph after updating and chart database, generate new assessment models；
Determine the accuracy of described new assessment models；And
Described accuracy based on described new assessment models, updates described assessment models.
17. devices according to claim 10, having with described evaluation object in the node of wherein said figure associates
It is one or more that the affiliated partner of system includes in following item: the personal data object that evaluation object is associated；Evaluation object phase
The collage-credit data object of association；The social relations data object that evaluation object is associated；The social networks that evaluation object is associated
Data object；The communication data object that evaluation object is associated.
18. devices according to claim 10, the weight on the described limit in wherein said figure is at least based in following item
One or more determine: the relationship type of the node associated by described limit；Association journey between node associated by described limit
Degree.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610310902.4A CN106021377A (en) | 2016-05-11 | 2016-05-11 | Information processing method and device implemented by computer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610310902.4A CN106021377A (en) | 2016-05-11 | 2016-05-11 | Information processing method and device implemented by computer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106021377A true CN106021377A (en) | 2016-10-12 |
Family
ID=57099167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610310902.4A Pending CN106021377A (en) | 2016-05-11 | 2016-05-11 | Information processing method and device implemented by computer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106021377A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106933949A (en) * | 2017-01-20 | 2017-07-07 | 浙江大学 | The planing method of influence power outburst in a kind of control social networks |
| CN106997488A (en) * | 2017-03-22 | 2017-08-01 | 扬州大学 | A kind of action knowledge extraction method of combination markov decision process |
| CN107153847A (en) * | 2017-05-31 | 2017-09-12 | 北京知道创宇信息技术有限公司 | Predict method and computing device of the user with the presence or absence of malicious act |
| CN107239533A (en) * | 2017-05-31 | 2017-10-10 | 北京知道创宇信息技术有限公司 | Generate abnormal patterns, determine method and computing device of the user with the presence or absence of malicious act |
| CN107729424A (en) * | 2017-09-28 | 2018-02-23 | 华为技术有限公司 | A kind of data visualization method and equipment |
| CN107957998A (en) * | 2016-10-14 | 2018-04-24 | 富士通株式会社 | The method of relation between the method and estimation entity pair of production Methods computation model |
| CN108038700A (en) * | 2017-12-22 | 2018-05-15 | 上海前隆信息科技有限公司 | A kind of anti-fraud data analysing method and system |
| CN108710999A (en) * | 2018-05-03 | 2018-10-26 | 上海电机学院 | The confidence level automatic evaluation method of shared resource under a kind of environment based on big data |
| CN109360106A (en) * | 2018-09-27 | 2019-02-19 | 北京京东金融科技控股有限公司 | Portrait construction method, system, medium and computer system |
| CN109582808A (en) * | 2018-11-22 | 2019-04-05 | 北京锐安科技有限公司 | A kind of user information querying method, device, terminal device and storage medium |
| CN109634995A (en) * | 2018-09-10 | 2019-04-16 | 阿里巴巴集团控股有限公司 | Main body is assessed to the method, apparatus and server of relationship |
| CN109657918A (en) * | 2018-11-19 | 2019-04-19 | 平安科技（深圳）有限公司 | Method for prewarning risk, device and the computer equipment of association assessment object |
| CN109933698A (en) * | 2019-02-27 | 2019-06-25 | 腾讯科技（深圳）有限公司 | A kind of the source method of calibration and device of user's portrait |
| CN109934302A (en) * | 2019-03-23 | 2019-06-25 | 大国创新智能科技(东莞)有限公司 | New category recognition methods and robot system based on fuzzy theory and deep learning |
| CN110348215A (en) * | 2019-07-16 | 2019-10-18 | 深圳众赢维融科技有限公司 | Exception object recognition methods, device, electronic equipment and medium |
| CN111537831A (en) * | 2020-04-01 | 2020-08-14 | 华中科技大学鄂州工业技术研究院 | Power distribution network line fault positioning method and device |
| CN111581446A (en) * | 2020-05-09 | 2020-08-25 | 北京金山云网络技术有限公司 | Graph relation generation method, device, system, equipment and storage medium |
| CN113282689A (en) * | 2021-07-22 | 2021-08-20 | 药渡经纬信息科技(北京)有限公司 | Retrieval method and device based on domain knowledge graph and search engine |
| CN113456033A (en) * | 2021-06-24 | 2021-10-01 | 江西科莱富健康科技有限公司 | Physiological index characteristic value data processing method and system and computer equipment |
| US20220285030A1 (en) * | 2019-08-21 | 2022-09-08 | Nec Corporation | Information processing method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101986299A (en) * | 2010-10-28 | 2011-03-16 | 浙江大学 | Multi-task personalized web service method based on hypergraph |
| CN102298606A (en) * | 2011-06-01 | 2011-12-28 | 清华大学 | Random walking image automatic annotation method and device based on label graph model |
| US20120072402A1 (en) * | 2008-12-19 | 2012-03-22 | Thales | Method and system for merging data or information |
| CN103593792A (en) * | 2013-11-13 | 2014-02-19 | 复旦大学 | Individual recommendation method and system based on Chinese knowledge mapping |
| CN103778145A (en) * | 2012-10-22 | 2014-05-07 | 中国移动通信集团广东有限公司 | Information recommendation method and system |
| CN104408149A (en) * | 2014-12-04 | 2015-03-11 | 威海北洋电气集团股份有限公司 | Criminal suspect mining association method and system based on social network analysis |
| CN104992266A (en) * | 2015-06-15 | 2015-10-21 | 广东电网有限责任公司电力调度控制中心 | Method of determining power grid node importance degree and system thereof |
| WO2015162458A1 (en) * | 2014-04-24 | 2015-10-29 | Singapore Telecommunications Limited | Knowledge model for personalization and location services |
| CN105205180A (en) * | 2015-10-27 | 2015-12-30 | 无锡天脉聚源传媒科技有限公司 | Knowledge map evaluation method and device |
-
2016
- 2016-05-11 CN CN201610310902.4A patent/CN106021377A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120072402A1 (en) * | 2008-12-19 | 2012-03-22 | Thales | Method and system for merging data or information |
| CN101986299A (en) * | 2010-10-28 | 2011-03-16 | 浙江大学 | Multi-task personalized web service method based on hypergraph |
| CN102298606A (en) * | 2011-06-01 | 2011-12-28 | 清华大学 | Random walking image automatic annotation method and device based on label graph model |
| CN103778145A (en) * | 2012-10-22 | 2014-05-07 | 中国移动通信集团广东有限公司 | Information recommendation method and system |
| CN103593792A (en) * | 2013-11-13 | 2014-02-19 | 复旦大学 | Individual recommendation method and system based on Chinese knowledge mapping |
| WO2015162458A1 (en) * | 2014-04-24 | 2015-10-29 | Singapore Telecommunications Limited | Knowledge model for personalization and location services |
| CN104408149A (en) * | 2014-12-04 | 2015-03-11 | 威海北洋电气集团股份有限公司 | Criminal suspect mining association method and system based on social network analysis |
| CN104992266A (en) * | 2015-06-15 | 2015-10-21 | 广东电网有限责任公司电力调度控制中心 | Method of determining power grid node importance degree and system thereof |
| CN105205180A (en) * | 2015-10-27 | 2015-12-30 | 无锡天脉聚源传媒科技有限公司 | Knowledge map evaluation method and device |
Cited By (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107957998A (en) * | 2016-10-14 | 2018-04-24 | 富士通株式会社 | The method of relation between the method and estimation entity pair of production Methods computation model |
| CN106933949A (en) * | 2017-01-20 | 2017-07-07 | 浙江大学 | The planing method of influence power outburst in a kind of control social networks |
| CN106997488A (en) * | 2017-03-22 | 2017-08-01 | 扬州大学 | A kind of action knowledge extraction method of combination markov decision process |
| CN107239533B (en) * | 2017-05-31 | 2021-12-07 | 北京知道创宇信息技术股份有限公司 | Method and computing device for generating abnormal pattern and determining whether malicious behaviors exist in user |
| CN107153847A (en) * | 2017-05-31 | 2017-09-12 | 北京知道创宇信息技术有限公司 | Predict method and computing device of the user with the presence or absence of malicious act |
| CN107239533A (en) * | 2017-05-31 | 2017-10-10 | 北京知道创宇信息技术有限公司 | Generate abnormal patterns, determine method and computing device of the user with the presence or absence of malicious act |
| CN107729424A (en) * | 2017-09-28 | 2018-02-23 | 华为技术有限公司 | A kind of data visualization method and equipment |
| CN107729424B (en) * | 2017-09-28 | 2021-07-09 | 华为技术有限公司 | Data visualization method and equipment |
| CN108038700A (en) * | 2017-12-22 | 2018-05-15 | 上海前隆信息科技有限公司 | A kind of anti-fraud data analysing method and system |
| CN108710999A (en) * | 2018-05-03 | 2018-10-26 | 上海电机学院 | The confidence level automatic evaluation method of shared resource under a kind of environment based on big data |
| CN109634995A (en) * | 2018-09-10 | 2019-04-16 | 阿里巴巴集团控股有限公司 | Main body is assessed to the method, apparatus and server of relationship |
| CN109634995B (en) * | 2018-09-10 | 2023-06-30 | 创新先进技术有限公司 | Method, device and server for evaluating relationship of main body pairs |
| CN109360106B (en) * | 2018-09-27 | 2021-03-30 | 京东数字科技控股有限公司 | Sketch construction method, system, medium and computer system |
| CN109360106A (en) * | 2018-09-27 | 2019-02-19 | 北京京东金融科技控股有限公司 | Portrait construction method, system, medium and computer system |
| CN109657918A (en) * | 2018-11-19 | 2019-04-19 | 平安科技（深圳）有限公司 | Method for prewarning risk, device and the computer equipment of association assessment object |
| CN109657918B (en) * | 2018-11-19 | 2023-07-18 | 平安科技（深圳）有限公司 | Risk early warning method and device for associated evaluation object and computer equipment |
| CN109582808A (en) * | 2018-11-22 | 2019-04-05 | 北京锐安科技有限公司 | A kind of user information querying method, device, terminal device and storage medium |
| CN109933698A (en) * | 2019-02-27 | 2019-06-25 | 腾讯科技（深圳）有限公司 | A kind of the source method of calibration and device of user's portrait |
| CN109933698B (en) * | 2019-02-27 | 2021-06-08 | 腾讯科技（深圳）有限公司 | User portrait source verification method and device |
| CN109934302A (en) * | 2019-03-23 | 2019-06-25 | 大国创新智能科技(东莞)有限公司 | New category recognition methods and robot system based on fuzzy theory and deep learning |
| CN110348215A (en) * | 2019-07-16 | 2019-10-18 | 深圳众赢维融科技有限公司 | Exception object recognition methods, device, electronic equipment and medium |
| CN110348215B (en) * | 2019-07-16 | 2021-08-06 | 深圳众赢维融科技有限公司 | Abnormal object identification method, abnormal object identification device, electronic equipment and medium |
| US20220285030A1 (en) * | 2019-08-21 | 2022-09-08 | Nec Corporation | Information processing method |
| CN111537831B (en) * | 2020-04-01 | 2022-06-24 | 华中科技大学鄂州工业技术研究院 | Power distribution network line fault positioning method and device |
| CN111537831A (en) * | 2020-04-01 | 2020-08-14 | 华中科技大学鄂州工业技术研究院 | Power distribution network line fault positioning method and device |
| CN111581446B (en) * | 2020-05-09 | 2023-02-03 | 北京金山云网络技术有限公司 | Graph relation generation method, device, system, equipment and storage medium |
| CN111581446A (en) * | 2020-05-09 | 2020-08-25 | 北京金山云网络技术有限公司 | Graph relation generation method, device, system, equipment and storage medium |
| CN113456033A (en) * | 2021-06-24 | 2021-10-01 | 江西科莱富健康科技有限公司 | Physiological index characteristic value data processing method and system and computer equipment |
| CN113456033B (en) * | 2021-06-24 | 2023-06-23 | 江西科莱富健康科技有限公司 | Physiological index characteristic value data processing method, system and computer equipment |
| CN113282689A (en) * | 2021-07-22 | 2021-08-20 | 药渡经纬信息科技(北京)有限公司 | Retrieval method and device based on domain knowledge graph and search engine |
| CN113282689B (en) * | 2021-07-22 | 2023-02-03 | 药渡经纬信息科技(北京)有限公司 | Retrieval method and device based on domain knowledge graph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106021377A (en) | Information processing method and device implemented by computer | |
| CN110413707A (en) | The excavation of clique's relationship is cheated in internet and checks method and its system | |
| CN105335157B (en) | A kind of demand classes sort method for integrating subjective and objective evaluation and system | |
| CN106407999A (en) | Rule combined machine learning method and system | |
| CN105528387A (en) | Segmentation discovery, evaluation and implementation platform | |
| CN105931116A (en) | Automated credit scoring system and method based on depth learning mechanism | |
| CN107230108A (en) | The processing method and processing device of business datum | |
| CN112417176B (en) | Method, equipment and medium for mining implicit association relation between enterprises based on graph characteristics | |
| CN113011895B (en) | Associated account sample screening method, device and equipment and computer storage medium | |
| CN108734567A (en) | A kind of asset management system and its appraisal procedure based on big data artificial intelligence air control | |
| Nakhkob et al. | Predicted increase enrollment in higher education using neural networks and data mining techniques | |
| Aher et al. | Mining association rule in classified data for course recommender system in e-learning | |
| CN105354644A (en) | Financial time series prediction method based on integrated empirical mode decomposition and 1-norm support vector machine quantile regression | |
| Ekong et al. | A Machine Learning Approach for Prediction of Students’ Admissibility for Post-Secondary Education using Artificial Neural Network | |
| Sugumar et al. | A technique to stock market prediction using fuzzy clustering and artificial neural networks | |
| CN109410035B (en) | Method and tool for assisting anti-fraud analysis of group structure | |
| Prasanth et al. | Intelligent Loan Eligibility and Approval System based on Random Forest Algorithm using Machine Learning | |
| Li | Construction of intelligent service system of university library based on internet of things in artificial intelligence environment | |
| Mullainathan et al. | From predictive algorithms to automatic generation of anomalies | |
| Shaik et al. | Customer loan eligibility prediction using machine learning | |
| Hou et al. | Prediction of learners' academic performance using factorization machine and decision tree | |
| CN117171428B (en) | Method for improving accuracy of search and recommendation results | |
| Kabra et al. | Student’s Performance Prediction Using Genetic Algorithm | |
| Wahyono et al. | Matching User in Online Learning using Artificial Intelligence for Recommendation of Competition | |
| Kumar et al. | Data Mining in Credit Scoring and Future Application |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20170323 Address after: 200122 Shanghai City Bureau of Huangpu District Road, No. eight, No. 457 bridge, building four, floor 3 Applicant after: Shanghai point information technology Co., Ltd. Address before: 200002 floor, building 5-7, No. 207, Mengzi Road, Shanghai, Huangpu District Applicant before: Shanghai Rong Rong Financial Information Service Co., Ltd. |
|
| TA01 | Transfer of patent application right | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1228536 Country of ref document: HK |
|
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20220208 |
|
| AD01 | Patent right deemed abandoned | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: WD Ref document number: 1228536 Country of ref document: HK |