Computational Intelligence: An Introduction - Computational Intelligence in Time Series Forecasting - page 8

clear distinction between them. Tettamanzi and Tomassini (2001) rather view the
scope of computational intelligence as the broader of the two methodologies,
because computational intelligence encompasses most various techniques for
describing and modelling of complex systems, which is not the case with the scope
of soft computing. This is in accordance with the view of Zadeh (1993, 1996,
1999), which defines computational intelligence as the combination of soft
computing and numerical processing. But still, Engelbrecht (2002) suggests
conceiving soft computing as an extension of computational intelligence in the
sense that the probabilistic methods are added to the paradigms of computational
intelligence.
In fact, the boundary of the disciplines associated with computational
intelligence are still not finally defined. They are still growing up to include new
emerging disciplines. For example, the agenda of the 2002 IEEE World Congress
on Computational Intelligence includes neuroinformatics and neurobiology as
new constituents. In the meantime, computational intelligence is viewed as a new-
generation artificial intelligence for human-like data and knowledge processing,
professionally known as High Machine Intelligence Quotient ( HMIQ )
technology. Most recently, the convergence of the core computational technologies
- neural networks, fuzzy systems, and evolutionary computation - to a common
frontier has drawn strong attention from the computational intelligence society. A
related term was coined: autonomous mental development (Wenig, 2003).
1.6 Hybrid Computational Technology
In the 1990s we witnessed a new trend in computational intelligence. A growing
number of publications on its applications have been published reporting on
successful combination of intelligent computational technologies - neural, fuzzy,
and evolutionary computation - in solving advanced artificial intelligence
problems. The hybrid computational technology created in this way is rooted
mainly in integrating various computational algorithms in order to implement more
advanced algorithms required for solving more complex problems. For instance,
neural networks have been combined with fuzzy logic to result in neuro-fuzzy or
fuzzy-neuro systems in which:
x Neural networks tune the parameters of the fuzzy logic systems, which are
used in building of adaptive fuzzy controllers, as implemented in the
Adaptive Network-Based Fuzzy Inference System (ANFIS) proposed by
Jang (1993).
x Fuzzy logic systems monitor the performance of the neural network and
adapt its parameters optimally, for instance in order to achieve the
nonlinear mapping and/or the function approximation to any desired
accuracy (Wang, 1992).
x Fuzzy logic is used to control the learning rate of neural networks to avoid
the creeping phenomenon in the network when approaching the solution
minimum (Arabshahi et al ., 1992).