Computational Intelligence: An Introduction - Computational Intelligence in Time Series Forecasting - page 10

Data mining is a strategy for rapid collection, storage and processing of huge
amounts of data (Mitra and Mitra, 2002) in some particular application areas, such
as in production and financial engineering (Heider, 1996; Major and Riedinger,
1992), surgery (Blum, 1982), telecommunication networks (Pedrycz, Vasilakos,
and Karnouskos, 2003/2004), Internet (Etzioni, 1996), etc .
Multisensor data fusion , again, is an advanced area of signal processing that
deals with the simultaneous collection of multiple sensor values related to a
physical system or to any observable phenomenon. It is the most useful technique
for solving the problems of pattern recognition and pattern interpretation (Bloch,
1996). For instance, in analysis of remotely sensed satellite images the multisensor
image interpretation plays a crucial role. Here, the reflected radiation values from
different sensors build a feature vector, which subsequently undergoes the feature
extraction and classification process (Bloch, 1996). In engineering, multisensor
data fusion has been applied to solve the problems of systems performance
monitoring and the problems of fault diagnosis of rotating machinery based on
vibration measurements (Emmanouilidis et al ., 1998). In addition, the multisensor
data fusion approach has been particularly applied in monitoring of operability of
individual sensors (Taniguchi and Dote, 2001). In recent years, on-line fault
detection and diagnosis of dynamic systems based on a reliable model of the
overall system behaviour under normal operating conditions have been the subjects
of research by the soft computing experts. Remarkable results have been reported
in this field of research by Akhimetiv and Dote, (1999).
In systems engineering , the application of soft computing encompasses the
activities that are essential for system study, optimal system design, and design of
adaptive system control concepts: identification and model building of dynamic
systems (Tzafestas, 1999; Zurada et al ., 1994). Here, model building and parameter
estimation of dynamic systems are the initial steps in the generation of a
mathematical description of dynamic systems behaviour, based on experimental
data. The methodology of computational intelligence helps generally in
implementation of advanced neuro and fuzzy controllers and supports the evolving
of adaptive controllers.
Optimal path planning is a soft computing application area widely needed in
manufacturing, primarily in job-shop scheduling and rescheduling, in optimal
routing in very large-scale integration layouts, and in robotics for optimal path
planning of robots and manipulators.
As a systems designer's tool, computational intelligence helps in styling the
circuit layout in microelectronics (Bosacci, 1997), optimal product shaping, etc .
1.8 Applications in Industry
In the industrial reality, there is a growing need for employing completed machine
and process automation, which includes not only the motion or process control, but
also their performance monitoring, diagnosis, and similar tasks. Owing to the
increasing complexity of the tasks, advanced intelligent computational tools, such
as soft computing and computational intelligence, are called upon to help in
handling the execution of the tasks efficiently. The application capabilities of both