Scribd
Upload
2 views3 pages

Fuzzy Set Approximation Using Polar2

The document presents a method for fuzzy set approximation called FEAT-p, which improves the approximation of relationships between antecedent and consequent universes by utilizing all available rules rather than just two surrounding rules. The method involves shifting fuzzy sets to align with an observation's reference point and calculating new set shapes using polar cuts, with a focus on the influence of proximity through weighted factors. Numerical examples are provided to illustrate the method's effectiveness and the impact of different weighting factors.

Uploaded by

sukaryaditrisakt11
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
2 views3 pages

Fuzzy Set Approximation Using Polar2

The document presents a method for fuzzy set approximation called FEAT-p, which improves the approximation of relationships between antecedent and consequent universes by utilizing all available rules rather than just two surrounding rules. The method involves shifting fuzzy sets to align with an observation's reference point and calculating new set shapes using polar cuts, with a focus on the influence of proximity through weighted factors. Numerical examples are provided to illustrate the method's effectiveness and the impact of different weighting factors.

Uploaded by

sukaryaditrisakt11
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

3 Fuzzy Set Approximation by Shifting of Linguistic

Terms (FEAT-p)
The proposed method is based on the assumption that a better approximation of
the real relation between the antecedent and consequent universes can be attained
by taking into consideration all the available rules in the rule base. This supposal
appeared already right at the beginning of the history of the fuzzy rule
interpolation e.g. in [7] or later e.g. in [13]. In spite of the possible advantages
most of the methods use only two rules that surround the observation. The
technique being presented serves the determination of the antecedent and
consequent sets of the new rule in the first and third stage in the first step of the
generalized methodology. The method is the same regardless of being applied for
an antecedent or a consequent dimension.

Figure 1
The original partition and the result of the shifting

The starting point is a fuzzy partition with the reference points of the sets
determined in advance and the reference point of the observation in the actual
dimension/partition. All the sets in the partition belong to the antecedent part of
one or more rules.
First all the sets are shifted horizontally in order to reach the coincidence of their
reference points with the reference point of the observation. This idea is similar to
the concept in [2], but that method uses and translates only the two flanking sets
into the location of the observation.

Figure 2
Next the shape of the new set is determined from the collection of the overlapped
sets. There are several solutions for this task. In [5] the authors present a solution
with low computational complexity based on α-cuts (FEAT-α) for the most
popular case of the convex and normal fuzzy (CNF) sets.
Further on a polar cut based technique (FEAT-p) is presented. It can also be
applied in cases when the normality condition is not satisfied for all the sets
participating in the approximation process, i.e. the height of one or more sets is
smaller than 1. Similar to the choice of the reference point the selection of the
calculation mode of the shape is also a tuning point.
The suggested shape calculation technique is based on the assumption that an
extension and a resolution principle of the fuzzy sets can be defined for polar cuts.
The method uses a polar co-ordinate system whose origin coincides with the
abscissa of the reference point of the observation. A polar cut is defined by a value
pair {rho, θ} (Fig. 2) that determines a point on the shape of the linguistic term.
For each polar cut of the approximated set the value rho is calculated as weighted
average of the rho values of the shifted sets for the same θ angle using the formula
(2).
nj

∑w jk { }
⋅rho A jkθ
{ }=
rho A a
jθ
k =1
nj
(2)

∑w k =1
jk

where rho denotes the length of a polar cut, j is the actual antecedent dimension, θ
is the angle of the actual cut, nj is the number of the sets in the partition, Ajkθ is the
a
polar cut of the kth set, wjk is the weighting factor of the kth set and A jθ is the
approximated polar cut.
It seems to be natural that the sets whose original position were in the
neighbourhood of the reference point of the observation to exercise higher
influence as those ones situated in farther regions of the universe of discourse.
Therefore the weighting factor should be dependent on distance. The simplest
weighting factor is the reciprocal value of the distance, which can be expressed by
the formula (3) with p=1, but there are several recommendations in the literature
for more or less analogue cases. For example in [7] the square of the reciprocal
value of the distance is suggested (p=2). The authors of [11] propose the use of the
reciprocal value of the distance on the mth power (p=m), where m is number of the
antecedent dimensions.
1
w jk = (3)
(
d A*j , A jk ) p
*
where A j is the fuzzy set corresponding to the jth dimension of the observation. In
[13] three variants of the weighting factor called extensibility functions are
introduced. These can be described by (4) with p=1 respective p=2 and (5).
1
w jk = (4)
λ ⋅ d (A*j , A jk )
p

(
− λ ⋅d A*j , A jk )
w jk = e (5)

where λ is a positive constant determining the effective extensibility distance. The


choice of the weighting factor can add a free parameter to the formula (2) to adjust
the sensitivity.
In case of the weighting factors (3) and (4) the distance between the observation
and the kth set of the partition is in the denominator. Therefore these weighting
factors require the supplement of the formula (2), namely if the distance is zero
the polar cuts of the approximated set should coincide with the respective polar
cuts of the set Ajk. Thus the endpoints of the polar cuts are calculated using the
formula (6). Hereby the method FEAT-p becomes a fuzzy set interpolation
technique. It ensures the fulfilment of the condition 4 from [6], namely the
compatibility with the rule base, for the rule interpolation method based on the
above mentioned method. It means that if the observation coincides with the
antecedent part of a rule, the estimated conclusion should coincide with the
consequent part of that rule, too.

⎧ nj
⎪ ∑ w jk ⋅rho A jkθ { }
⎪ k =1 ( )
d A*j , A jk > 0
{ } ⎪ nj
=⎨
∑
a
rho A jθ w jk (6)
⎪ k =1
⎪
⎪ rho A jkθ
⎩ { } ( )
d A*j , A jk = 0

4 Numerical Examples
This section intends to present some relevant features of the suggested method
outlining the effect of the weighting factor.

You might also like