Time-Series Forecasting Using Fuzzy-Neural System with Evolutionary Rule Base
Arit Thammano, Sirinda Palahan · Journal of Robotics and Mechatronics · 2006
This paper proposes a new hybrid time-series forecasting system which is the fusion of fuzzy systems and artificial neural networks. The proposed fuzzy-neural system consists of 5 layers: an input layer, fuzzification layer, rule layer, hidden layer, and output layer. The artificial neural network is used as the fuzzy inference engine and the genetic algorithm is used to optimize the fuzzy rule base. This proposed system was tested with 20 time-series datasets. The results obtained were very encouraging.