Nonsmooth training of fuzzy neural networks
โ Scribed by C. Eitzinger
- Book ID
- 106168697
- Publisher
- Springer
- Year
- 2004
- Tongue
- English
- Weight
- 424 KB
- Volume
- 8
- Category
- Article
- ISSN
- 1432-7643
No coin nor oath required. For personal study only.
๐ SIMILAR VOLUMES
This paper presents a practical algorithm for training neural networks with fuzzy number weights, inputs, and outputs. Typically, fuzzy number neural networks are di cult to train because of the many -cut constraints implied by the fuzzy weights. A transformation is used to eliminate these constrain
In this paper, we describe a method for nonlinear fuzzy regression using neural network models. In earlier work, strong assumptions were made on the form of the fuzzy number parameters: symmetric triangular, asymmetric triangular, quadratic, trapezoidal, and so on. Our goal here is to substantially