how to create hyperbox for fuzzy set? and how to calculate hyperbox membership degree?

Question

I want to know that while creating hyperbox (Fuzzy min max Neural network) all the parameters of input dataset (WBC or PID ) should be used or work can be done with single parameter also And want to know weather the size of hyperbox depend on input pattern ?And how calculate membership degree of hyperbox?

or

Is there any easier way to generate hyperbox for fuzzy set that you know.

Answer 1

While creating Hyperbox you required dataset and dataset contains number of attributes and there will be one attribute or field called class .Hyperbox should be created using that attribute (class). For example PID data set contain 9 attribute but the last attribute called Class as field name is specified in PID.name file. Hyperbox will be created for that class.

Answer 2

Check out this link Videos regarding HyperBox in detail

Consider "iris" dataset from UCI Repository which contains n-features. Link to dataset

For example: 5.1,3.5,1.4,0.2,Iris-setosa

This sample is single point in n-dimensional plane. First 4 values are the n-dimensions of that point. Here n is 4. and Last term notifies the Class Name to which it belongs.

Initially, there is no Hyperbox present. So to create a Hyperbox, consider the 4 values as Min and Max values of Hyperbox. Because there is only one point at the moment, Min and Max points will be same. This Hyperbox will be called as Point Hyperbox. If second point that is next feature sample belongs to same class, it is necessary to check whether existing Hyperbox can be expanded or not using Expand function. If it can be expanded Update the Min-Max Point if necessary else there is no need of Hyperbox expansion. If Hyperbox can not be expanded or second point does not belongs to same class as of existing Hyperbox, create a new Hyperbox.