211.

[Cover] [Contents] [Index]

Page 288

 

The belief for other subsets is zero. The plausibilities for subsets {B}, {B, F}, and {B, P} are all equivalent to one, because

 

The second example is an extension of the first. Suppose that there is a third observation that confirms {P} to the degree of 0.7, denoted by mc, and let ma mb=md. The orthogonal sum mc md is illustrated in Table 7.2. There are two null intersections, i.e. Ø. One contains the value 0.294, and the other 0.196. Therefore, the normalisation constant K and the final evidence combination results are computed by:

 

and for all other subsets, mc md=0.

Both belief and plausibility can be derived in the similar way, as explained in example one, and so will not be duplicated here. Readers may also wish to combine bpa using the above examples, but in a different sequence. For example, one may compute the orthogonal sum first using evidence 1 and 3, and then combine to evidence 2. This exercise should show that the results of evidence combination are the same, irrespective of the ordering.

Table 7.2 Extended example of Dempster’s rule combination

md

mc

md({B})=(0.42)

md({B, F})=(0.28)

md({B,P})=(0.18)

md(θ)=(0.12)

mc({P})(0.7)

Ø=0.294

Ø=0.196

{P}=0.126

{P}=0.084

mc(θ)(0.3)

{B}=0.126

{B, F}=0.084

{B, P}=0.054

θ=0.036

[Cover] [Contents] [Index]


Classification Methods for Remotely Sensed Data
Classification Methods for Remotely Sensed Data, Second Edition
ISBN: 1420090720
EAN: 2147483647
Year: 2001
Pages: 354

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net