決策層特征融合decisionlevelidentityfusion(編輯修改稿)
2025-01-30 21:49 本頁(yè)面
【文章內(nèi)容簡(jiǎn)介】 discernment is the miniature “world” we are trying to observe and understand. ={A1, A2, …, An} ?DempsterShafer’s method ? 2n general propositions may be developed by Boolean binations. ? One important general proposition ? If evidence is assigned to it is equivalent to a general level of uncertainty. 1 1 2 1 1 22 { , { } , ..., { } , { } , ..., { } , ..., { ... } }n n n nA A A A A A A A A? ?? ? ? ? ? ? ?12 ... nA A A? ? ? ? ?DempsterShafer’s method ? The DS method assigns evidence to both single and general propositions instead of assigning probability to hypotheses (Bayesian). ? Probability mass, m( ), to represent assigned evidence. ? m( ), denotes a probability mass assigned either to an elementary proposition or to a general proposition. The sum of all mass function assigned to elementary and general propositions is ( ) 1m ????1( ) 1nim????DempsterShafer’s method ? The probability of a proposition Ai is given by summing the probability masses for the pertinent elements in Θ and 2Θ. iA , 2P r oba bil it y { A } = ( , 2 )im ??????We sum m(Θ) for the element of Θ that contains Ai exactly and in addition, sum the m(Θ) for those general proposition in 2Θ that contain Ai as an element. DempsterShafer’s method E1 E2 . . . . . . . Ek H1 H2 . . . . . . . HN Bayes Assignment of Evidence Evidence Hypotheses E1 E2 . . . . . . . Ek H1 H2 . . . . . . . HN DS Assignment of Evidence Evidence Hypotheses DempsterShafer’s method ? Evidential interval, [spt(Bi), Pls(Bi)] ? The support for a proposition Bi is ? If Bi is a simple proposition (Bi = Ai), then the spt(Bi) is simply the probability of Ai。 ? If Bi is a general proposition, (., Bi = A1∨ A2 ∨ A3)then the support for Bi is the sum of probability masses contributing to all elements of Bi. iB , 2( ) ( , 2 )iiSpt B m ???????眾信度函數(shù) 1 2 3 1 2 1 2 2 3 1 3 1 2 3( ) ( ) ( ) ( ) ( ) ( ) ( )Sp t A A A m A m A m A A m A A m A A m A A A? ? ? ? ? ? ? ? ? ? ? ? ?DempsterShafer’s method ? Evidential interval, [spt(Bi), Pls(Bi)] ? The plausibility of a proposition Ai is ? Which means lack of evidence that refutes the proposition 似真度函數(shù) ( ) 1 ( )iiP ls A Sp t A??A useful feature of DS approach is the ability to establish a general level of uncertainty. The DS method provides a means to explicitly account for unknown possible causes of observational data. DempsterShafer’s method Sensor 1 Observables Classifier Declaration Sensor 2 ETC Sensor n ETC Compute or Enumerate Mass Distribution for Given declaration ETC ETC Combine/Fuse Distributions Via Dempster’s Rules of Combination M(Oj)=F(mi(Oj)) Decision Logic Mi(Oj) Fused Indentity Declaration DempsterShafer’s method ? Dempster’s Rule of Combination Proposition 1=u0=hypothesis A is true Proposition 2=u1=hypothesis B is true Proposition 3=u2=hypothesis A or B is true S1 S2 m2(u0) m2(u1) m2(u2) m1(u0) m(u0)=m1(u0)m2(u0) k10=m1(u0)m2(u1) m(u0)=m1(u0)m2(u2) m1(u1) k01=m1(u1)m2(u0) m(u0)=m1(u1)m2(u1) m(u0)=m1(u1)m2(u2) m1(u2) m(u0)=m1(u2)m2(u0) m(u0)=m1(u2)m2(u1) m(u0)=m1(u2)m2(u2) DempsterShafer’s method ? Dempster’s rule of bination for two independent sources is 112,1( ) ( )()1ijijA B um A m Bmuc????12
法律信息相關(guān)推薦
鄂ICP備17016276號(hào)-1