MATLAB toolbax for fuzzy type 2 is ready for order

You can train fuzzy type 2 based on a novel training algorithm.

The noise reduction property of T2FLSs that use a novel type-2 fuzzy membership function (ellipsoidal type-2 membership function) is studied in this dissertation. The proposed type-2 membership function has certain values on both ends of the support and the kernel, and some uncertain values for the other values of the support. In this part of the dissertation, the parameter tuning rules of a T2FLS that uses such a membership function are derived using the gradient descent (GD) learning algorithm. There exists a number of papers in literature which claim that the modeling and control performance of T2FLSs is better than T1FLSs under noisy conditions. This is
attempted to be shown via simulation (or real-time) studies only for some specific systems. However in this dissertation, a simpler T2FLS is considered with the novel membership function proposed in which the effect of input noise in the rule base is shown numerically in a general way. The proposed type-2 fuzzy neuro structure is tested on different input-output data sets, and it is shown that the T2FLS in combination with the proposed novel membership function has a better noise reduction property when compared to its type-1 counterpart.