CN101261717A - Subjective trust evaluation method based on cloud model - Google Patents

Abstract

A subjective trust quantification evaluation method based on the cloud model theory which is applied under a network environment in a network activity based on interpersonal relationship is mainly used for solving the trust decision problem of a trust body during the interpersonal interactive process in the network environment which aims at profit or is nonprofit. The technical points of the invention are that: 1. comprehensive quantification processing and evaluating are carried out according to the historical subjective credit rating data of a trust object and aiming at the nondeterminacy of subjective data; 2. the comprehensive quantification processing and evaluating of the credit rating change of the object are carried out according to the subjective credit rating changing data of the trust object; 3. the trust decision is accomplished based on the quantification data obtained in the steps of 1 and 2. The subjective trust quantification evaluation method of thte invention can effectively solve the nondeterminacy problem of the subjective trust quantification evaluation and upgrades the successful rate of online trading or cooperation and the satisfaction of the trust body.

Description

Subjective trust evaluation method based on cloud model

Technical field

The present invention relates under the internet environment, with profit and non-profit be purpose, based on trust evaluation in the network activity of interpersonal relation and trust decisions field, particularly, relate to a kind of subjective trust evaluation method based on cloud model.

Background technology

Along with the internet, applications field constantly enlarges and various service functions perfect, network application just by closed to open development, ecommerce and online transaction (B2B, C2C, B2C), Web Community, instant message etc. are more and more abundanter and popularize based on the application of internet.These application present the characteristics that are different from LAN: promptly number of users numerous and can not pre-determine, can be fully mutual between the user etc.This makes the Internet user present the trend of socialization, the user no longer is isolated mutually one by one individuality, but rely on certain network application, formed complicated network interpersonal relation, user can by network directly thousands of other users of contact and and they conclude the business and cooperate.In this case, traditional can not fully satisfy the requirement of numerous network applications based on the security mechanism of registration, authentication and authentication, need security mechanism more flexibly, Here it is based on the security mechanism of trusting.Trusting relationship is considered to than authorizing the relation security relationship of essence more, so the steadily operation of health of human society largely has benefited from the trusting relationship between individual, group and the tissue.

Because online transaction betides under the open internet environment, cause existence that the object of different services or similar service is provided in a large number.How correctly, reasonably a vital problem for main body is to initiate to differentiate final trading object and select object before the online transaction, i.e. trust decisions.For nonprofit Web Community, the cooperation between the user still set up with the basis of trusting each other on, when facing numerous cooperation candidate, trust main body and still face the trust decisions problem.

Current, about the research of trusting relationship, comprise two kinds of trusting relationships that are mutually related, promptly object is trusted and subjective trust.Wherein, the object during object is trusted is meant object or the entity of getting rid of people's subjective factor.Relation between object and object can fully be verified by appropriate evidence, is a kind of evidential method, accurately description, reasoning and checking.Reasoning based on hypothesis proves the basic research method that object is trusted, the trusting relationship in analyzing as security protocols such as BAN logics.Trust main body in the subjective trust refers to individuality or the colony that is made of people or procurator (as Agent), subjective trust (be called for short and trust) is a kind of mankind's a cognitive phenomenon, be to the special characteristic of trust object or other subjective judgement of a specific order of behavior, and this judgement is relatively independent of the feature and the behavior of main body.From certain non-formal angle, trust main body A the trust of trusting object B is meaned that A believes that B must carry out certain action under certain particular environment.The trust decisions problem that the present invention is primarily aimed under the subjective trust relation is studied, and provides subjective trust evaluation method.

Numerous researchers have carried out the research about problems such as subjective trust modeling, reasonings, and obtain certain achievement.There are multiple trust modeling, assessment and inference method, comprise probability theory method, its weak point is not consider the ambiguity of trust itself, and its reasoning is based upon on the pure probability model, show a kind of " excessively " formal trend, away from the reality of trust management; Based on fuzzy set theory subjective trust administrative model, linguistic variable and fuzzy logic are introduced subjective trust reasoning research.Fuzzy mathematics needs pre-defined membership function, and in case after determining membership function, certain numerical value just can not change for the subjection degree of specific concept, lacks certain dirigibility.Therefore, need a kind of can be rationally, effectively in conjunction with the quantization method of subjective assessment randomness and ambiguity characteristics, in directly perceived, succinct mode,, finally reach the purpose that promotes online transaction, cooperation quality and trust the main body satisfaction for the trust decisions of trusting main body provides support.

Summary of the invention

Concern subjectivity, uncertainty and the ambiguity that is had at subjective trust, and the object credit rating changes the influence to trust decisions, the present invention proposes a kind of subjective trust method for quantitatively evaluating based on subjective trust cloud and trust variation cloud, and the online transaction, the collaboration application that are used for solving based on the subjective trust degree are trusted main body to trusting the quantitatively evaluating problem of object credit rating.This method can solve the trust decisions problem of trusting main body preferably in conjunction with the analysis of object credit rating historical information and change information thereof.

The inventive method is specific as follows:

Step 1:

The design of subjective trust cloud comprises the type selecting of trusting cloud, and general the selection has the one dimension normal state cloud of universality well, but does not get rid of the selection of using the other types cloud; The setting of STD credit rating upper and lower bound; STD credit rating series of discrete and successional setting are as then being provided with initial credit degree value and credit rating spacing value for discrete series; Can set the qualitativing concept rank of STE in case of necessity.

Step 2:

According to the design result of subjective trust cloud, from the credit rating database, obtain the historical credit rating evaluation information of many objects.The ground that is without loss of generality, whether the present invention unifies not do the pressure requirement to the scoring of different objects, and different objects can be marked by different evaluation mechanism, but may influence the accuracy of evaluation.Therefore, in identical evaluation mechanism and scoring system, the effect of the method for the invention can be better.

Step 3:

Credit rating data age weights are set.Subjective credit rating evaluating data has certain ageing, and apart from current trust decisions data constantly far away more, its validity that correctly reflects the object credit rating descends thereupon constantly in scoring.

Step 4:

According to historical credit rating of object and ageing weights thereof, adopt the reverse cloud generating algorithm of weighting, generate the numerical characteristic value of subjective trust cloud.The reverse cloud generating algorithm of weighting is as follows:

Input: N water dust x1, x2 ..., xN}, and the set of the weights of water dust w1, w2 ..., wN}, and have $1=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{w}_i;$

Output: expectation value Ex, entropy En and the super entropy He of the qualitativing concept that N water dust is represented;

Concrete steps are as follows:

1) calculates sample average according to xi and weights thereof $\stackrel{\‾}{X}=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{w}_i{x}_i,$ Single order sample Absolute Central Moment Sample variance $S^{}$ ${}^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{w}_i{(x_i-\stackrel{\‾}{X})}^2.$

2) estimated value of Ex is $\hat{E}x=\stackrel{\‾}{X}.$

3) estimated value of He is $\hat{H}e=\sqrt{\frac{\mathrm{\π}}{2}}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{w}_i|x_i-\hat{E}x|.$

4) estimated value of En is $E\hat{n}=\sqrt{S^2-\frac{1}{3}\×H{\hat{e}}^2}.$

Trust decisions process involved in the present invention need comprise a plurality of trust objects (be at least two and trust object), therefore, needs to preserve the subjective trust cloud numerical characteristic value of (needing interim preservation at least) each trust object.

Step 5:

Make up the subjective trust proper vector, and carry out quantitatively evaluating trusting object.The subjective trust proper vector is made of the numerical characteristic Ex and the He value of subjective trust cloud, and shape is as＜Ex, He 〉.According to the subjective trust proper vector, when trusting the object quantitatively evaluating, the physical significance of Ex and He is respectively: Ex is as the representative value that the subjective credit rating of object is estimated, i.e. the average credit degree level of object credit rating; He represents the departure degree of object credit rating and average credit degree, can think that He has embodied the stability of object credit rating.When Ex is big, can think that the object credit rating is higher, on the contrary lower; When He hour, can think that the object credit rating is more stable, on the contrary its less stable.Main body can be according to a plurality of object＜Ex, He〉subjective judgement of value, carry out trust decisions.As select Ex higher, the less object of while He.

Step 6:

When having a plurality of object＜Ex, He〉when close, and main body can't carry out subjectivity when selecting, and then changes step 7 over to.

Step 7:

The object credit rating is carried out the sequential processing, generate object credit rating delta data.Object credit rating delta data is to generate to trust the data basis that changes cloud, and subjective trust cloud sequential method is as follows:

Input: the subjective trust cloud expectation value SetEx={Ext0 of N+1 band time mark, Ext1 ..., ExtN}; Output: N (∞, ∞) the water dust Drops={d1 of the variation of the expression degree of belief the open interval in, d2 ... dN}; Concrete steps are as follows:

1) according to time mark to Ext0, Ext2 ..., ExtN} sorts again, generate new S set etEx '=Ex0, Ex1 ..., ExN}, and Time (Exi)＜Time (Exj), 0＜=i＜j＜=N, Time () they are the function of Exi → ti, ti is the rise time of Exi;

2)num＝1，low＝0，sup＝1；

3) by formula, $d_{\mathrm{num}}=\frac{(\sup -\mathrm{low})}{(\mathrm{Time}\left({\mathrm{Ex}}_{\sup }\right)-\mathrm{Time}\left({\mathrm{Ex}}_{\mathrm{low}}\right))}$ Calculate the rate of change of two adjacent subjective trust cloud expectation values;

4)num＝num+1，low＝low+1，sup＝sup+1；

5) repeating step (3)-(4) are till num＞N.

The time interval among the S set etEx between element does not need strict equating, but must guarantee that the time interval has identical time measure unit, as week, the moon, season etc.

Step 8:

According to Drops={d1, d2 ... the data of dN}, and the ageing weights of step 3 kind adopt the reverse cloud generating algorithm of weighting, generate to trust the numerical characteristic value that changes cloud.For the selection principle of the type that trust to change cloud, identical with subjective trust varieties of clouds type selection principle in ' step 1 '.

Step 9:

Under the similar prerequisite of subjective trust proper vector,, the credit rating of trusting object is carried out quantitatively evaluating further by the subjective trust variation characteristic vector of more a plurality of objects.Subjective trust variation characteristic vector is by the numerical characteristic Ex that trusts cloud _EAnd He _EValue constitutes, and shape is as＜Ex _E, He _E.Trusting the thought that changes cloud is, it is the essential characteristic of trusting relationship that the object credit rating changes in time, also is the key factor that the subjective trust evaluation need be considered simultaneously.The foundation of trusting relationship and development are long-term processes between subject and object, and the time is the key factor that main body is investigated the object confidence level.This mode of thinking meets the objective law of daily trusting relationship focusing on people.

The physical significance of subjective trust variation characteristic vector is:

1) ExE represents the variation pattern of object credit rating, represents the credit rating no change when ExE equals 0; When ExE represents that credit rating is decline less than 0 the time; When representing credit rating greater than 0 the time, ExE promotes.

2) HeE represents the stability that the object credit rating changes, and the more little expression credit rating of HeE changes stable more, otherwise then unstable more.

Main body can be according to a plurality of object＜Ex _E, He _EThe value subjective judgement, carry out trust decisions.As select Ex higher, the less object of while He.

Step 10:

When there being a plurality of object＜Ex _E, He _EWhen close, and main body can't carry out subjectivity when selecting, and then changes step 11, otherwise finishes the trust decisions process.

Step 11:

Use forward cloud algorithm, generate water dust.Choose all bigger trust object of water dust value and water dust degree of certainty.Normal state forward cloud generating algorithm is as follows:

Input: 3 numerical characteristic value Ex of expression qualitativing concept C, En, He (En 〉=He 〉=0); Water dust is counted N;

Output: the quantitative values of N water dust and each water dust belong to the degree of certainty of notion C.

Step: generate a uniform random number En ' on interval [En-He, En+He];

1) a uniform random number En ' in the generation interval [En-He, En+He];

2) generating with Ex is expectation value, (En ') ²A normal random number x for variance;

3) make that x is the once concrete quantized value of qualitativing concept C, be called water dust;

4) calculate $y=e^{-{(x-\mathrm{Ex})}^2/2{\left({\mathrm{En}}^{\′}\right)}^2},$ Make that y is the degree of certainty that x belongs to qualitativing concept C, { x, y} have intactly reflected the full content of qualitative, quantitative conversion this time;

(5) repeating step (1)-step (4) is up to producing N water dust

Step 12: finish the trust decisions process.

Description of drawings

Fig. 1 prototype system structural representation

Fig. 2 trust decisions schematic flow sheet

Embodiment

Understand and enforcement the present invention the existing embodiments of the invention of describing in conjunction with the accompanying drawings for ease of persons skilled in the art.

As shown in Figure 1, for implementing the prototype system structural representation of method described in the present invention, by at least one user terminal apparatus 1, at least one trust decisions center 2, and at least one credit rating database 3 is formed.User terminal apparatus 1 can be any wired or wireless communication ability that has, and the device that can carry out message exchange with trust decisions center 2, as PC, mobile phone etc.User terminal 1 is by any wire communication line 401, or wireless communication line 402, carries out alternately with trust decisions center 2.Credit rating database 3 is used to store the information such as subjective historical credit rating of trusting object, and carries out data interaction by any wire communication line 401 or order control linkage 403 with trust decisions center 2.

Trust decisions center 2, by at least one master control set 201, at least one subjective trust evaluating apparatus 202, at least one trusts variation evaluating apparatus 203, and at least one forward cloud generating apparatus 204 is formed.Master control set 201 is connected with device 202,203,204 by order control linkage 403, and the execution of responsible whole evaluation procedure and finish information interaction with user terminal apparatus 1.Device 202 trust evaluation of being responsible for finishing based on the subjective trust cloud, device 203 are responsible for finishing the trust evaluation that changes cloud based on trusting, and the forward that device 204 is finished water dust generates task.

In the specific implementation process of the present invention, can dispose a plurality of distributed trust decisions center 2 as required, and a plurality of trust decisions center 2 can be carried out information and data interaction mutually.In addition, the physical centralization configuration can be carried out in trust decisions center 2 and credit rating database 3, also can carry out decentralized configuration.

As shown in Figure 2, existing in conjunction with prototype system structural representation shown in Figure 1 for implementing the trust decisions process flow diagram of method described in the present invention, concrete implementation step is described in detail as follows:

Step 1:

User terminal apparatus 1 by 401 or 402 and the trust decisions center connect.By and the message exchange of 201 of master control sets, trust object for selected one group, as the evaluation object of trust evaluation.Credit rating according to selected trust object is estimated mechanism and scoring system, master control set 201 is finished the design of subjective trust cloud, comprising: set credit rating and estimate type, set STD credit rating upper and lower bound, set the discrete or continuity of STD, set the qualitativing concept grade of STE in case of necessity.

Step 2:

Master control set 201 obtains the historical credit rating data of trusting object from credit rating database 3.

Step 3:

According to ageing setting principle, 201 pairs of credit rating data of master control set are carried out ageing weights setting.The cardinal rule that ageing model of credit rating and weights are determined is as follows:

If the ageing model M of credit rating=＜X, tc, tb, T 〉:

1) X={x ₁, x ₂..., x _nBe the complete or collected works of the historical credit rating evaluating data of object, for arbitrary evaluating data x _i, Time (x _i) represent credit rating evaluation constantly.

2) tc represents current trust decisions constantly, and as timeorigin, tb is certain moment on the forward time shaft, as the time threshold of judging credit rating evaluating data validity.

3) T={t ₁, t ₂..., t _M-iBy m-1 ordered set that moment value is formed between tc and tb, for arbitrary t _i, di=|t _i-tc| is called t _iTo the time gap of tc, and satisfy:

● $\∀d_i(1\≤i\≤m-1)\→d_i\≤|t_c-t_b|$

● $\&ForAll;d_i,d_j(1\&le;i<j\&le;m-1)\&RightArrow;d_i<d_j$

According to Time (x _i), tb can be decomposed into X two subclass X1 ' and X2 ', and meets the following conditions:

1) X=X1 ' ∪ X2 ', and X1 ' ∩ X2 '=Φ;

2) $\&ForAll;x_i\&Element;{X^{\&prime;}}_1(1\&le;i\&le;n)\&RightArrow;\left(\right|\mathrm{Time}\left(x_i\right)-t_c|\&le;|t_c-t_b\left|\right);$

3) $\&ForAll;x_i\&Element;{X^{\&prime;}}_2(1\&le;i\&le;n)\&RightArrow;\left(\right|\mathrm{Time}\left(x_i\right)-t_c|>|t_c-t_b\left|\right);$

The decomposition of set X, with | Time (x _i)-tc and | the difference of tc-tb| is a foundation, and tc is a timeorigin, | tc-tb| is the time threshold of credit rating evaluating data validity.According to | tc-tb| is decomposed into subset X 1 ' and X2 ' with X.The time gap of arbitrary element and tc all is less than or equal to threshold value among the X1 ', and the time gap of middle arbitrary element of X2 ' and tc is all greater than threshold value.Therefore, can think the evaluation time of credit rating data among the X2 ', far away apart from the current decision-making moment, can not correctly reflect the object credit rating of current time; Comprise all among the X1 ' and participate in the data that the object credit rating is estimated.

Set T is separated into m sub-range with the time interval of tc and tb formation, is called time window, is designated as Wt.Time window is divided into m credit rating evaluating data subclass, Xt with X1 ' ₁, Xt ₂..., Xt _m, and meet the following conditions:

1) arbitrary time window $W_{\mathrm{ti}}=<t_{\mathrm{low}}^i,t_{\sup }^i>,$ t _Low ⁱBe Wt _iMoment lower limit, t _Sup ⁱBe Wt _iThe moment upper limit, and satisfy $|t_{\mathrm{low}}^i-t_c|<|t_{\sup }^i-t_c|;$ $\left|W_{\mathrm{ti}}\right|=|t_{\sup }^i-t_{\mathrm{low}}^i|$ Be called Wt _iLength of window;

2) X1 '=Xt ₁∪ Xt ₂∪ ..., ∪ Xt _m, and $\&ForAll;{\mathrm{Xt}}_i,{\mathrm{Xt}}_j(1\&le;i\&le;m,1\&le;j\&le;m)\&RightArrow;({\mathrm{Xt}}_i\&cap;{\mathrm{Xt}}_j)=\mathrm{\&phi;};$

3) $\&ForAll;y\&Element;{\mathrm{Xt}}_i,z\&Element;{\mathrm{Xt}}_j(1\&le;i<j\&le;m)\&RightArrow;|\mathrm{Time}\left(y\right)-t_c|<|\mathrm{Time}\left(z\right)-t_c|;$

During design set T, should take all factors into consideration | the time span of tb-bc| and | the sum of credit rating evaluating data in the tb-bc|.Set T further is decomposed into m subclass with X1 ', and according to time window under each subclass of X1 ', Xt ₁, Xt ₂..., Xt _mBetween have strict time sequencing.The credit rating data of time value in identical time window have identical ageing weights, can be arbitrary subset X t of X1 ' _i(1＜=i＜=m) gives an ageing weight w t _i, wt _iExpression Xt _iIn the credit rating evaluating data to the influence degree of the whole credit rating evaluation result of object, and need satisfy the constraint condition of formula (1) and (2).According to the constraint condition of formula (1) and (2), the invention provides a kind of simple weights that satisfy formula (3) and determine method.These weights are determined method hypothesis, and along with the increase of the time gap between credit rating evaluating data time value and tc, the ageing decay of credit rating evaluating data is very fast, and the mode that ratio such as employing descends is expressed this attenuation trend.

$\&ForAll;x_i\&Element;{\mathrm{Xt}}_K,x_j\&Element;{\mathrm{Xt}}_l(1\&le;k<l\&le;m)\&RightArrow;({\mathrm{wt}}_k<{\mathrm{wt}}_l)---\left(1\right)$

$\left(\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{\mathrm{wt}}_i\right)=1---\left(2\right),$ $\frac{w{t}_{i-1}}{{\mathrm{wt}}_i}=\frac{{\mathrm{wt}}_i}{{\mathrm{wt}}_{i+1}}(2\&le;i\&le;m-1)---\left(3\right)$

Step 4:

Realize that according to the reverse cloud generating algorithm of weighting device 202 calculates the numerical characteristic value of object subjective trust cloud, and it is back to master control set 201.Enter step 5.

Step 5:

Master control set 201 is trusted object according to the result of the device 202 subjective trust cloud numerical characteristic values that return for each, makes up subjective trust proper vector＜Ex, He 〉.And the result is back to user terminal apparatus 1.

Step 6:

Trust main body the subjective trust proper vector of a plurality of trust objects is judged,, then finish the trust decisions process if main body is selected certain object; Otherwise need to select the further object of evaluation, and, send the request and the object information of further trust decisions to master control set 201 by user terminal apparatus 1.

Step 7:

Master control set 201 receives the request of user terminal apparatus 1, and the historical credit information of object is passed to trust variation evaluating apparatus 203.Device 203 carries out the sequential processing to each object credit information, generates the water dust Drops={d1 that degree of belief changes, d2 ... dN}.

Step 8:

Trust variation evaluating apparatus 203 and change water dust, trust object for each and generate the numerical characteristic that trust changes cloud, and it is returned master control set 201 according to trusting.

Step 9:

The trust that master control set 201 returns according to device 203 changes the result of cloud numerical characteristic value, trusts object for each, makes up and trusts variation characteristic vector＜Ex _E, He _E.And the result is back to user terminal apparatus 1.

Step 10:

Trust main body the trust variation characteristic vector of a plurality of trust objects is judged,, then finish the trust decisions process if main body is selected certain object; Otherwise need to select the further object of evaluation, and, send the request and the object information of further trust decisions to master control set 201 by user terminal apparatus 1.

Step 11:

Master control set 201 receives the request of user terminal apparatus 1, and the historical credit information of object is passed to trust variation evaluating apparatus 204.The device 204 subjective trust cloud numerical characteristic values according to each object use forward cloud generating algorithm, generate a water dust.Select water dust value and all bigger object of degree of certainty, be back to device 201.With final object selection result, return to the user by device 201 by installing 1.

Step 12:

Finish the trust decisions process.

The inventive method is rationally finished the selection of trusting object for main body guidance is provided with directly perceived, simple, effective and efficient manner.Solved randomness and ambiguity problem beyond expression of words in the subjective trust evaluation.Particularly will trust object credit rating changing factor and introduce the trust decisions process, and can more reasonably promote the accuracy and the main body satisfaction of decision-making.The trust decisions proper vector of cloud has directly perceived, succinct quantification characteristics, has avoided the trust main surface to multiple evaluation mechanism, reaches the decision error that is produced when multiple evaluation result represents form.

In the specific embodiment of the invention, be example, introduced the embodiment that implements the inventive method with the normal state cloud.Cloud model type of the present invention, and forward, the weighting reverse cloud generating algorithm relevant with the cloud model type can be chosen according to actual needs; And, in the invention process process, do not require that the subjective trust cloud is in full accord with the type of trusting the variation cloud in principle.Therefore, at cloud model type, relevant cloud operative algorithm, and the variation of aspect such as two kinds of cloud model type consistance selection, all belong to the protection domain of claim of the present invention.

Claims (13)

1. in the interpersonal reciprocal process in the internet based on cloud model, method for quantitatively evaluating to subjective trust scoring with randomness, ambiguity characteristics, it is characterized in that: the numerical characteristic value that (1) uses the subjective trust cloud, carry out quantitatively evaluating to the historical credit rating of many trusts object; (2) use the numerical characteristic value of trusting the variation cloud, quantitatively evaluating is carried out in the credit rating variation of many trusts object; (3) according to the quantification numerical characteristic value of subjective trust cloud and trust variation cloud, finish the trust decisions in the interpersonal reciprocal process in internet.

2. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1, it is characterized in that the subjective trust cloud adopts subjective trust degree space S TD to represent quantitative credit rating evaluation of estimate in (1), quantitative domain interval [0, n] an orderly numerical value set, be designated as " 0; n ", this set can be made of continuous or discrete dull numerical value; N is any positive integer, and dividing another name 0 and n is the degree of belief lower limit and the upper limit of STD; And the subjective trust degree then represents the subjective credit rating of object low more more near 0, more near n, then represents the subjective credit rating of object high more.

3. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1, it is characterized in that subjective trust cloud described in (1) adopts subjective trust space S TS to gather as the credit rating qualitativing concept, can also can ignore the setting of reliability rating for STS default 1 or a plurality of reliability rating.

4. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1, it is characterized in that trust described in (2) changes cloud and adopts degree of belief variation space E TD to represent that quantitative credit rating changes evaluation, (∞ ∞) is the open interval of real number axis-∞ to ∞ to ETD=; And have, $\&ForAll;x\&Element;\mathrm{ETD},$ If x=0 represents the degree of belief no change; X＜0 expression degree of belief descends, and x＞0 expression degree of belief promotes.

5. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1, it is characterized in that trusting described in (2) and change the cloud courier with trusting the qualitativing concept set that variation space E TS changes as credit rating, can also can ignore the setting of reliability rating for ETS default 1 or a plurality of reliability rating.

6. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1 is characterized in that (1) further comprises: the design of subjective trust cloud, subjective trust cloud numerical characteristic generates, and makes up subjective trust proper vector＜Ex, He 〉.

7. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1 is characterized in that (2) further comprise: subjective trust cloud sequentialization, subjective trust cloud numerical characteristic generates, and makes up subjective trust variation characteristic vector＜Ex _E, He _E.

8. subjective trust cloud method for designing according to claim 6, it is characterized in that further comprising: the type of subjective trust cloud is chosen, and the degree of belief upper and lower bound in subjective trust degree space is set, and selects the discreteness or the continuity Characteristics of degree of belief sequence.

9. generate according to the described subjective trust cloud of claim 6 numerical characteristic, it is characterized in that: adopt the reverse cloud generating algorithm of weighting, the subjective trust cloud numerical characteristic of completeness credit rating evaluating data generates.

10. according to the reverse cloud generating algorithm of the described weighting of claim 9, it is characterized in that, its be input as N water dust x1, x2 ..., xN}, and the set of the weights of water dust w1, w2 ..., wN}, and have $1=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{w}_i;$ It is output as expectation value Ex, entropy En and the super entropy He of the represented qualitativing concept of N water dust; And, comprise following steps:

1) calculates sample average according to xi and weights thereof $\stackrel{\&OverBar;}{X}=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{w}_i{x}_i,$ Single order sample Absolute Central Moment

Sample variance $S^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{w}_i{(x_i-\stackrel{\&OverBar;}{X})}^2;$

2) estimated value of Ex is $\hat{E}x=\stackrel{\&OverBar;}{X};$

3) estimated value of He is $\hat{H}e=\sqrt{\frac{\mathrm{\&pi;}}{2}}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{w}_i|x_i-\hat{E}x|;$

4) estimated value of En is $E\stackrel{\mathrm{\&Lambda;}}{n}=\sqrt{S^2-\frac{1}{3}\&times;H{\hat{e}}^2}.$

11. according to the described subjective trust cloud of claim 7 sequential method, it is characterized in that: it is input as the subjective trust cloud expectation value SetEx={Ext0 of N+1 band time mark, Ext1 ..., ExtN}; It is output as N (∞, ∞) the water dust Drops={d1 of the variation of the expression degree of belief the open interval in, d2 ... dN}; And comprise following steps:

1) according to time mark to Ext0, Ext2 ..., ExtN} sorts again, generate new S set etEx '=Ex0, Ex1 ..., ExN}, and Time (Exi)＜Time (Exj), 0＜=i＜j＜=N, Time () they are the function of Exi → ti, ti is the rise time of Exi;

2)num＝1，low＝0，sup＝1；

3) by formula, $d_{\mathrm{num}}=\frac{(\sup -\mathrm{low})}{(\mathrm{Time}\left({\mathrm{Ex}}_{\sup }\right)-\mathrm{Time}\left({\mathrm{Ex}}_{\mathrm{low}}\right))}$ Calculate the rate of change of two adjacent subjective trust cloud expectation values;

4)num＝num+1，low＝low+1，sup＝sup+1；

5) repeating step (3)-(4) are till num＞N.

12. generate according to the described subjective trust cloud of claim 7 numerical characteristic, it is characterized in that: according to Drops={d1, d2 ... the data of dN} adopt the reverse cloud generating algorithm of weighting, generate to trust the numerical characteristic value that changes cloud.

13. the subjective trust method for quantitatively evaluating based on cloud model according to claim 1 is characterized in that (3) comprise following steps:

1) the subjective trust proper vector＜Ex of more a plurality of trust objects, He 〉, select Ex bigger, and the less trust object of He, if having a plurality of＜Ex, He〉close object, then can select arbitrary object at random, perhaps change step 2);

2) the subjective trust variation characteristic vector＜Ex of more a plurality of trust objects _E, He _E, select Ex _EBigger, and He _ELess trust object is if exist a plurality of＜Ex _E, He _EClose object, then can select arbitrary object at random, perhaps change step 3);

3) use the forward cloud generator, according to subjective trust cloud numerical characteristic, to a plurality of＜Ex, He〉and＜Ex _E, He _EEach object in the close object generates a water dust, selects water dust value and all bigger object of degree of certainty.

Images