I have a project on function approximation by fuzzy decision trees and I want to compare my results with some other methods improved by fuzzy logic.
I think that your fuzzy decision trees can be compared with a fuzzy bayesian networks, taking into account the same states and knowledge used for creating the fuzzy decision trees. It is important to consider the use of Fuzzy Entropy Measures to reduce the dimensionality of fuzzy decision trees. The above has also been applied in fuzzy bayesian networks. Other alternative is to add a decision-making based on Multiatribute or Multicriteria in order to observe the performance the fuzzy decision trees respect to the fuzzy bayesian networks.
you must to know what do you want?
fuzzy decission trees can be compared with some fuzzy method but one of the best method is bayesian network. I think you must to implement your method and bayesian method in a system and compare it with eachother.
I have other idea: use two fuzzy approximation called Disjunctive normal form (DNF) or Conjunctive normal form (CNF) to build-up a fuzzy decision and compare it with respect to your fuzzy decision tree. I've read some papers about fuzzy approximation as Computational Rule of Inference (CRI) and a special transformation called Fuzzy (F)-Transform, in order to get several options to compare.
You can read my papers about infinite fuzzy logic controllers (Fuzzy Sets and Systems, 1995, InterStat 2003) if you want work with the fuzzy approximations of Lebesque functions or fuzzy probability)
It will be good use the notice that the maximum of entropy product is equivalent to maximum entropy principle because of the properties of ln functions.
- Badges
- Science topic