Genetics-based self-adjusting expert systems
Ramakrishnan Pakath · 2002
The conventional rule-based expert system (ES) paradigm has been criticized because of its inability to display characteristics of true intelligence and its relatively static knowledge storehouse and dependence on rote-learned rule fragments as a basis for reasoning. An approach to constructing self-adjusting ESs is described. Self-adjusting ESs are characterized by their dynamic knowledge bases that improve with experience and their ability to pursue unfocused reasoning. An implementation of the paradigm is examined and its traits in light of the criticisms leveled against its predecessor are evaluated.>