CN104750838B - Towards the comprehensive situation quantitative Treatment method of big data analysis - Google Patents
Towards the comprehensive situation quantitative Treatment method of big data analysis Download PDFInfo
- Publication number
- CN104750838B CN104750838B CN201510158651.8A CN201510158651A CN104750838B CN 104750838 B CN104750838 B CN 104750838B CN 201510158651 A CN201510158651 A CN 201510158651A CN 104750838 B CN104750838 B CN 104750838B
- Authority
- CN
- China
- Prior art keywords
- entity
- attribute
- property value
- peripheral environment
- matrix
- 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
Abstract
The present invention relates to the comprehensive situation quantitative Treatment method towards big data analysis, including:The first step, system initialization;Second step, structure matrix;Third step, correction matrix;4th step, transition matrix;5th step, iterative processing;6th step, normalized.The present invention can overcome deficiency existing for existing comprehensive situation quantitative Treatment method, and accuracy is assessed conducive to being promoted.
Description
Technical field
The present invention relates to a kind of comprehensive situation quantitative Treatment method towards big data analysis, core is to big data depth
The improvement of analysis mechanisms is spent, belongs to technical field of the computer network.
Background technology
As far as the applicant is aware, at present in big data analysis, the core of target scene comprehensive situation quantitative Treatment method is thought
Think mainly to include:Using statistical function as the simple type quantitative Treatment method of core, using PageRank algorithms as the iterative type of core
Quantitative Treatment method and its mutation processing method.
In actual quantification processing procedure, using statistical function as the simple type quantitative Treatment method of core, with target scene
Each entity entity attribute value list in certain section of measurement period is object in (i.e. target large data sets), and statistical function is learned in access
(such as average value, mean square deviation), and then obtain entire target scene integrated situational awareness picture.Wherein, involved in entire calculating process
To all objects be referred to as entity, target scene can be considered the set of all entities, all ginsengs that can describe physical characteristics
Number is referred to as entity attribute.But this method does not consider the incidence relation between each entity, easily causes entity weight and adds up upwards
Phenomenon in turn results in the failure of quantitative Treatment method.
Using PageRank algorithms as the iterative type quantitative Treatment method of core and its mutation processing method, all it is in itself
To search for the searching algorithm for target, used morbid state entity repair mechanism is inherently a kind of " problem supervision ", and
Non- " crew's supervision ", therefore, this method easily causes peripheral environment entity and its direct precursor entity in entire quantitative Treatment system
Middle weight is excessive, and immediate successor entity weight is too small, in turn results in the failure of quantitative Treatment method.
Invention content
The technical problems to be solved by the invention are:In view of the problems of the existing technology, it provides a kind of towards big data
The comprehensive situation quantitative Treatment method of analysis can overcome deficiency existing for existing comprehensive situation quantitative Treatment method, conducive to promotion
Assess accuracy.
The technical solution that the present invention solves its technical problem is as follows:
A kind of comprehensive situation quantitative Treatment method towards big data analysis, it is characterized in that, include the following steps:
The first step, system initialization:Entity attribute collection e is formed with all entity attributes of entity in target scene, and with
All entity property value composition entity property value vector w (e)n×1=(w (e1),w(e2),...,w(en))T, n is entity attribute
Quantity;Iteration threshold ξ is set simultaneously;Go to second step;
Second step, structure matrix:Entity attribute association square can be transmitted according to the incidence relation setting between each entity attribute
Battle array R;Go to third step;
Third step, correction matrix:Setting entity property value is w (e)n+1Peripheral environment entity, in each entity and outer collarette
It is set respectively between the entity of border and two-way transmits relating attribute;Relating attribute is transmitted with each entity of peripheral environment entity direction
Row vector w (ie) under the weight of value composition peripheral environment entityn×1=(w (ie1),w(ie2),...,w(ien))T, and with each reality
Body is directed toward row vector w (oe) in the weight for transmitting relating attribute value composition peripheral environment entity of peripheral environment entityn×1=(w
(oe1),w(oe2),...,w(oen))T;Entity attribute incidence matrix R can be transmitted to be modified toGo to the 4th step;
4th step, transition matrix:According to correction matrix R(1)And each entity attribute can the transitive attribute downlink degree of association,
Operation, which obtains, can transmit entity attribute transition matrix Q;By entity property value vector w (e)n×1It is extended toGo to the 5th step;
5th step, iterative processing:By formulaTo entity property value vector w'
(e)(n+1)×1Processing is iterated, wherein,It is vectorial for entity property value after kth time iteration, and enableWhenWhen, iterative processing terminates, and after obtaining iterative processing
Entity property value vectorGo to the 6th step;
6th step, normalized:If W (e)n×1=(W (e1),W(e2),...,W(en))TTo normalize entity property value
Vector, whereinI ∈ { 1,2 ..., n }, j ∈ { 1,2 ..., n };The W (e)n×1I.e.
For final gained entity property value vector;Processing terminates.
Further perfect technical solution is as follows by the present invention:
Preferably, the detailed process of the first step is:Entity attribute collection e=[e are set1,e2,...,en], entity property value to
Measure w (e)n×1=(w (e1),w(e2),...,w(en))T, iteration threshold ξ, wherein n be entity attribute quantity, i.e. n=| e |, w
(e1),w(e2),...,w(en) it is respectively entity attribute e1,e2,...,enEntity property value;Go to second step.
Preferably, the detailed process of second step is:Setting can transmit entity attribute incidence matrix R=(rij)n×n, i ∈ 1,
2 ..., n }, j ∈ { 1,2 ..., n };If there are entity attribute eiIt is directed toward entity attribute ejAssociation, then enable rij=| ei→ej|,
Otherwise r is enabledij=0;Go to third step.
Preferably, in third step, w (ie1),w(ie2),...,w(ien) it is respectively that peripheral environment entity is directed toward each entity
Relating attribute value, w (oe can be transmitted1),w(oe2),...,w(oen) it is respectively that each entity is directed toward transmitting for peripheral environment entity
Relating attribute value.
Preferably, in third step, correction matrixi∈
1,2 ..., n+1 }, j ∈ 1,2 ..., n+1 }.
Preferably, in the 4th step, entity attribute e is setiCan the transitive attribute downlink degree of associationSetting
Entity attribute transition matrix Q=(q can be transmittedij)(n+1)×(n+1), i ∈ 1,2 ..., n+1 }, j ∈ 1,2 ..., n+
1};If outi>0, then it enablesOtherwise q is enabledij=0.
Compared with prior art, beneficial effects of the present invention are as follows:
The present invention is by introducing peripheral environment entity, and build between peripheral environment entity and each entity two-way transmits pass
It is attribute, peripheral environment entity is formed to " equality and the comprehensively supervision " of each entity and each entity to peripheral environment entity
" anti-supervision ", and then the weight flowing being effectively ensured between each relating attribute, make each entity be obtained in comprehensive situation calculating process
To " fair, just treatment ", while in active undertaking comprehensive situation calculating process " anti-supervision obligation ".Based on this, this
Invention processing method is more in line with the actual motion flow of big data source scene, and can ensure that quantitative Treatment result it is more efficient and
Accurately, especially suitable for the information system integrated situation quantitative analysis taking human as behavioral agent.
Description of the drawings
Fig. 1 is the main process figure of present invention method.
Fig. 2, which is that each entity attribute is original in application case of the present invention, transmits incidence relation figure.
Fig. 3 is to transmit incidence relation figure based on peripheral environment entity in application case of the present invention.
Specific embodiment
The present invention is described in further detail with reference to the accompanying drawings and in conjunction with the embodiments.But the present invention is not limited to be given
The example gone out.
Embodiment
As shown in Figure 1, the present embodiment includes the following steps towards the comprehensive situation quantitative Treatment method of big data analysis:
The first step, system initialization:Entity attribute collection e is formed with all entity attributes of entity in target scene, and with
All entity property value composition entity property value vector w (e)n×1=(w (e1),w(e2),...,w(en))T, n is entity attribute
Quantity;Iteration threshold ξ is set simultaneously;Go to second step;
Detailed process is:Entity attribute collection e=[e are set1,e2,...,en], entity property value vector w (e)n×1=(w
(e1),w(e2),...,w(en))T, iteration threshold ξ, wherein n be entity attribute quantity, i.e. n=| e |, w (e1),w
(e2),...,w(en) it is respectively entity attribute e1,e2,...,enEntity property value;Go to second step.
Second step, structure matrix:Entity attribute association square can be transmitted according to the incidence relation setting between each entity attribute
Battle array R;Go to third step;
Detailed process is:Setting can transmit entity attribute incidence matrix R=(rij)n×n, i ∈ { 1,2 ..., n }, j ∈ 1,
2,...,n};If there are entity attribute eiIt is directed toward entity attribute ejAssociation, then enable rij=| ei→ej|, otherwise enable rij=0;Turn
It is walked to third.
Third step, correction matrix:Setting entity property value is w (e)n+1Peripheral environment entity, in each entity and outer collarette
It is set respectively between the entity of border and two-way transmits relating attribute;Relating attribute is transmitted with each entity of peripheral environment entity direction
Row vector w (ie) under the weight of value composition peripheral environment entityn×1=(w (ie1),w(ie2),...,w(ien))T, and with each reality
Body is directed toward row vector w (oe) in the weight for transmitting relating attribute value composition peripheral environment entity of peripheral environment entityn×1=(w
(oe1),w(oe2),...,w(oen))T;Entity attribute incidence matrix R can be transmitted to be modified toI ∈ 1,2 ..., n+1 }, j ∈ 1,2 ..., n
+1};Go to the 4th step;
Wherein, w (ie1),w(ie2),...,w(ien) be respectively peripheral environment entity be directed toward each entity transmit association
Property value, w (oe1),w(oe2),...,w(oen) be respectively each entity be directed toward peripheral environment entity transmit relating attribute value.
4th step, transition matrix:According to correction matrix R(1)And each entity attribute can the transitive attribute downlink degree of association,
Operation, which obtains, can transmit entity attribute transition matrix Q;By entity property value vector w (e)n×1It is extended toGo to the 5th step;
Wherein, setting entity attribute eiCan the transitive attribute downlink degree of associationSetting can transmit entity category
Property transition matrix Q=(qij)(n+1)×(n+1), i ∈ 1,2 ..., n+1 }, j ∈ 1,2 ..., n+1 };If outi>0, then
It enablesOtherwise q is enabledij=0.
5th step, iterative processing:By formulaTo entity property value vector w'
(e)(n+1)×1Processing is iterated, wherein,It is vectorial for entity property value after kth time iteration, and enableWhenWhen, iterative processing terminates, and after obtaining iterative processing
Entity property value vectorGo to the 6th step.
6th step, normalized:If W (e)n×1=(W (e1),W(e2),...,W(en))TTo normalize entity property value
Vector, whereinI ∈ { 1,2 ..., n }, j ∈ { 1,2 ..., n };The W (e)n×1As
Final gained entity property value vector;Processing terminates.
Application case:
Target scene is related to 6 entity attributes, and the entity property value of each entity attribute is 1, and between each other original can
Incidence relation is transmitted as shown in Fig. 2, the circle of label 1 to 6 represents each entity attribute in figure.At the present embodiment method
Reason.
(1) first step, system initialization:
Enable e=[e1,e2,...,e6], w (e)6×1=(w (e1),w(e2),...,w(e6))T=(1,1 ..., 1)T, n=
6, ξ=0.0001.Go to second step.
(2) second step, structure matrix:Entity attribute association can be transmitted according to the incidence relation setting between each entity attribute
Matrix R:
Go to third step.
(3) third step, correction matrix:As shown in figure 3, setting entity property value is w (e)7=1 peripheral environment entity 7,
Set respectively between each entity and peripheral environment entity it is two-way transmit relating attribute, each entity is directed toward with peripheral environment entity
Transmit relating attribute value composition peripheral environment entity weight under row vector w (ie)6×1=(w (ie1),w(ie2),...,w
(ie6))T=(1,1 ..., 1)T, and peripheral environment is formed with the relating attribute value of transmitting of each entity direction peripheral environment entity
Row vector w (oe) in the weight of entity6×1=(w (oe1),w(oe2),...,w(oe6))T=(1,1 ..., 1)T。
Entity attribute incidence matrix R can be transmitted to be modified to:
Go to the 4th step.
(4) the 4th steps, transition matrix:According to correction matrix R(1)And each entity attribute can transitive attribute downlink association
Degree, operation, which obtains, can transmit entity attribute transition matrix Q:
By entity property value vector w (e)6×1It is extended toIt goes to
5th step.
(5) the 5th steps, iterative processing:
It enablesFor iteration start vector, and by formulaTo entity
Property value vector w'(e)7×1It is iterated processing;
1st iteration, becauseThen after iteration: And have at this time Therefore continue iterative processing;
2nd iteration, because Then after iteration: And have at this timeTherefore continue at iteration
Reason;
…
11st iteration, because Then after iteration: And have at this timeTherefore after
It is continuous to be iterated processing;
…
17th iteration, because Then after iteration: And have at this timeTherefore
Iterative processing terminates;
Entity property value is vectorial after obtaining iterative processing at this time Go to the 6th step.
(6) the 6th steps, normalized:
If W (e)6×1=(W (e1),W(e2),...,W(e6))TIt is vectorial for normalization entity property value, whereinI ∈ { 1,2 ..., 6 }, j ∈ { 1,2 ..., 6 };With reference to having obtainedFinally:
W(T)(ei)6×1=[0.1026,0.1539,0.1026,0.2051,0.1539,0.2820], this i.e. final gained of this application case
Entity property value vector;Processing terminates.
So far the comprehensive situation quantitative Treatment towards big data analysis is completed, entire processing method is more in line with big data
The actual motion flow of source scene can ensure that quantitative Treatment result is more efficient and accurate.
In addition to the implementation, the present invention can also have other embodiment.It is all to use equivalent substitution or equivalent transformation shape
Into technical solution, all fall within the present invention claims protection domain.
Claims (5)
1. a kind of comprehensive situation quantitative Treatment method towards big data analysis, it is characterized in that, include the following steps:
The first step, system initialization:Entity attribute collection e is formed, and with all entity attributes of entity in target scene with all
Entity property value composition entity property value vector w (e)n×1=(w (e1),w(e2),...,w(en))T, n is the number of entity attribute
Amount;Iteration threshold ξ is set simultaneously;Go to second step;
Second step, structure matrix:Entity attribute incidence matrix R can be transmitted according to the incidence relation setting between each entity attribute;
Go to third step;
Third step, correction matrix:Setting entity property value is w (e)n+1Peripheral environment entity, in each entity and peripheral environment entity
Between set two-way transmit relating attribute respectively;The relating attribute value of transmitting that each entity is directed toward with peripheral environment entity forms
Row vector w (ie) under the weight of peripheral environment entityn×1=(w (ie1),w(ie2),...,w(ien))T, and be directed toward with each entity
Row vector w (oe) in the weight for transmitting relating attribute value composition peripheral environment entity of peripheral environment entityn×1=(w (oe1),
w(oe2),...,w(oen))T;Entity attribute incidence matrix R can be transmitted to be modified toTurn
To the 4th step;
4th step, transition matrix:According to correction matrix R(1)And each entity attribute can the transitive attribute downlink degree of association, operation
Entity attribute transition matrix Q can be transmitted by obtaining;By entity property value vector w (e)n×1It is extended to
Go to the 5th step;
5th step, iterative processing:By formulaTo entity property value vector w'(e)(n+1)×1
Processing is iterated, wherein,It is vectorial for entity property value after kth time iteration, and enable
WhenWhen, iterative processing terminates, and obtains entity property value vector after iterative processingGo to the 6th step;
6th step, normalized:If W (e)n×1=(W (e1),W(e2),...,W(en))TFor normalization entity property value to
Amount, whereinI ∈ { 1,2 ..., n }, j ∈ { 1,2 ..., n };The W (e)n×1As
Final gained entity property value vector;Processing terminates;
In third step, w (ie1),w(ie2),...,w(ien) be respectively peripheral environment entity be directed toward each entity transmit association
Property value, w (oe1),w(oe2),...,w(oen) be respectively each entity be directed toward peripheral environment entity transmit relating attribute value.
2. according to claim 1 towards the comprehensive situation quantitative Treatment method of big data analysis, it is characterized in that, the first step
Detailed process is:Entity attribute collection e=[e are set1,e2,...,en], entity property value vector w (e)n×1=(w (e1),w
(e2),...,w(en))T, iteration threshold ξ, wherein n be entity attribute quantity, i.e. n=| e |, w (e1),w(e2),...,w(en)
Respectively entity attribute e1,e2,...,enEntity property value;Go to second step.
3. according to claim 1 towards the comprehensive situation quantitative Treatment method of big data analysis, it is characterized in that, second step
Detailed process is:Setting can transmit entity attribute incidence matrix R=(rij)n×n, i ∈ { 1,2 ..., n }, j ∈ 1,2 ...,
n};If there are entity attribute eiIt is directed toward entity attribute ejAssociation, then enable rij=| ei→ej|, otherwise enable rij=0;Go to third
Step.
4. according to claim 1 towards the comprehensive situation quantitative Treatment method of big data analysis, it is characterized in that, third step
In, correction matrixI ∈ 1,2 ..., and n+1 }, j ∈
{1,2,......,n+1}。
5. according to claim 4 towards the comprehensive situation quantitative Treatment method of big data analysis, it is characterized in that, the 4th step
In, entity attribute e is setiCan the transitive attribute downlink degree of associationSetting can transmit entity attribute transition matrix
Q=(qij)(n+1)×(n+1), i ∈ 1,2 ..., n+1 }, j ∈ 1,2 ..., n+1 };If outi>0, then it enablesOtherwise q is enabledij=0.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510158651.8A CN104750838B (en) | 2015-04-03 | 2015-04-03 | Towards the comprehensive situation quantitative Treatment method of big data analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510158651.8A CN104750838B (en) | 2015-04-03 | 2015-04-03 | Towards the comprehensive situation quantitative Treatment method of big data analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104750838A CN104750838A (en) | 2015-07-01 |
CN104750838B true CN104750838B (en) | 2018-06-12 |
Family
ID=53590522
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510158651.8A Active CN104750838B (en) | 2015-04-03 | 2015-04-03 | Towards the comprehensive situation quantitative Treatment method of big data analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104750838B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105447186B (en) * | 2015-12-16 | 2019-03-08 | 汉鼎宇佑互联网股份有限公司 | A kind of user behavior analysis system based on big data platform |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002169826A (en) * | 2000-12-04 | 2002-06-14 | Just Syst Corp | Data converting device and storage medium stored with data conversion processing program |
CN101201926A (en) * | 2006-12-15 | 2008-06-18 | 深圳职业技术学院 | Synthesis assessment analytical system for formations health service computer network |
CN104050197A (en) * | 2013-03-15 | 2014-09-17 | 腾讯科技(深圳)有限公司 | Evaluation method and device for information retrieval system |
-
2015
- 2015-04-03 CN CN201510158651.8A patent/CN104750838B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002169826A (en) * | 2000-12-04 | 2002-06-14 | Just Syst Corp | Data converting device and storage medium stored with data conversion processing program |
CN101201926A (en) * | 2006-12-15 | 2008-06-18 | 深圳职业技术学院 | Synthesis assessment analytical system for formations health service computer network |
CN104050197A (en) * | 2013-03-15 | 2014-09-17 | 腾讯科技(深圳)有限公司 | Evaluation method and device for information retrieval system |
Non-Patent Citations (1)
Title |
---|
基于模糊Petri网的综合态势分析;李香亭等;《电子测试》;20110930;第43-45页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104750838A (en) | 2015-07-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112949837B (en) | Target recognition federal deep learning method based on trusted network | |
CN108446711B (en) | Software defect prediction method based on transfer learning | |
Ye et al. | Predicting positive and negative links in signed social networks by transfer learning | |
CN110210538B (en) | Household image multi-target identification method and device | |
CN107679183A (en) | Grader training data acquisition methods and device, server and storage medium | |
CN109492076B (en) | Community question-answer website answer credible evaluation method based on network | |
CN103617259A (en) | Matrix decomposition recommendation method based on Bayesian probability with social relations and project content | |
CN110348665A (en) | A kind of low-voltage platform area electric power system data quality evaluating method and device | |
He et al. | Precipitation deficits increase high diurnal temperature range extremes | |
Ramos et al. | The Long Term Fr\'echet distribution: Estimation, Properties and its Application | |
Zhao et al. | Socialtransfer: Transferring social knowledge for cold-start cowdsourcing | |
CN106997379A (en) | A kind of merging method of the close text based on picture text click volume | |
de Azevedo et al. | Ancient remains and the first peopling of the A mericas: Reassessing the Hoyo Negro skull | |
CN104750838B (en) | Towards the comprehensive situation quantitative Treatment method of big data analysis | |
CN116259078B (en) | Pesticide recommendation method, device, equipment and storage medium | |
Li et al. | Asymptotic analysis of endemic equilibrium to a brucellosis model | |
CN108491477A (en) | Neural network recommendation method based on multidimensional cloud and user's dynamic interest | |
Zhang et al. | A method of deep belief network image classification based on probability measure rough set theory | |
WO2023125456A1 (en) | Multi-level variational autoencoder-based hyperspectral image feature extraction method | |
CN105447315A (en) | D-S evidence law fusion improved method for evidence conflict | |
Zhou et al. | Peculiar statistical properties of Chinese stock indices in bull and bear market phases | |
DE102022120731A1 (en) | MULTIMODAL SENSOR FUSION FOR CONTENT IDENTIFICATION IN HUMAN-MACHINE INTERFACE APPLICATIONS | |
Tapsoba et al. | Joint modeling of survival time and longitudinal data with subject‐specific changepoints in the covariates | |
WO2021134944A1 (en) | Mobile news client-based evaluation method and system therefor | |
Shi et al. | Application on stock price prediction of Elman neural networks based on principal component analysis method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |