Cumulative Growth of Production Rules with Fuzzy Hierarchy (PRFH)

R. Kandwal and K.K. Bharadwaj (India)


Machine learning, Cumulative Growth, Hierarchical Production Rule, Production Rules with Fuzzy Hierarchy


This paper suggests a cumulative growth approach for Production Rules with Fuzzy Hierarchy (PRFH) system by merging the knowledge discovered during current episode into the previous episode knowledge. A PRFH system takes a generalized view of inheritance by associating each class Dk in the fuzzy hierarchy with three categories of properties–public, special and private properties. The PRFH for the class Dk (1 ≤ k ≤ n) is defined as follows: Pk→ Dk Generality [general class] Specificity [Dk1 (d1),…..,Dki(di),….Dkj(dij)], Where Pk is the set of preconditions= (Ppub) k∪(Pspl) k ∪(Ppvt) k and specificity element Dki(di) means that Dki is a specific class of Dk with degree of subsumption di. A set of related PRFH is called a cluster and is represented as a PRFH-tree. The aim of this approach is to merge the newly discovered knowledge into one of the clusters of knowledge discovered in the past. The suggested growth algorithm incrementally incorporates new discovered PRFH appropriately into one of the already existing PRFH-tree by maintaining consistency as well as minimizing redundancy. The proposed methodology would be useful in dynamic restructuring of knowledge bases especially in data streams.

Important Links:

Go Back