Choice of experiments and analysis of approximation accuracy in a Tagaki-Sugeno fuzzy system
Farida Benmakrouha, Pierre Yves Glorennec · 2003
We are concerned with two important points in identification of fuzzy systems, namely choosing training data for optimal coverage of input space and establishing approximation error bounds. We present a method to determine a good distribution of experimental data and we establish in this case approximation bound errors.