|
|
|
|
A membership function is a slice we cut off a pie: 0.2 of one pie may be a different size piece of 0.2 of another pie! |
|
|
|
| It has been proposed by Sasikala and Petrou that the membership functions are allowed to take values higher than 1, up to a value that reflects the relative importance between the different factors that are to be combined. | |
|
|
| In their paper in Fuzzy Sets and Systems, Sasikala and
Petrou proposed a training based scheme, where the appropriate operators
for combining information at the disjunctive and the conjunctive levels of
reasoning as well as the relative importance of the different factors
are learned using training data classified by experts. During the operational
stage of such a system the operators used and the relative weights of
the various factors are fixed. To read more on this,
click here for the acrobat version of this
paper.
|
|
|
|
 |
Now you are an expert on the Fuzzy approach, click here to learn on
the Probabilistic
approach |
| |
