ABSTRACT
Numerous intelligent systems are based on different approaches to machine intelligence problems. Computational intelligence researchers focus on the modeling of natural systems, which can be considered as intelligent ones. Design with computer-based shell and human knowledge inside it is known as an expert system. There are various approaches to intelligent systems but fundamental difference is located in the following distinction: the methods under consideration can be described as symbolic versus holistic ones. All pattern recognition problems are examples of inductive reasoning tasks and need some machine learning approach that is both interesting and difficult. Moreover, the methods of automatic reasoning and inference rules must be constructed by the system designer and must be explicit to be built into the system’s structure. It is always difficult to do so and sometimes it is the source of limitations during the system’s development and exploitation.
