Fuzzy Verdict Reveal with Pruning and Rule Extraction

Fuzzy min-max is supervised neural network classifier that creates hyperboxes for classification. In FMM trying to enhance classification numerous hyperboxes are created in the network. For network modification, rule of extraction algorithm is newly added to the FMM confidence and threshold is used for calculation of each FMM hyperbox. User defined limit is used to prune the hyperboxes with very less confidence factor. The multiple hyperboxes are created in the network by adding some changes to enhance the execution of min-max. The smaller set of rules extract from FMM to define prediction on it. detecting fault and problem with classification a set of sensor is created from power plant is assessed utilizing FMM. Multi level fuzzy (MLF) min-max neural network classifier is known as system for supervised learning. MLF uses basic concept of the fuzzy min max strategy in multilevel design pattern this strategy uses separate classifier for smaller hyperboxes examples are found in area of overlapping. In this paper we proposed pruning and rule for extraction. Pruning reduces the size of decision tree and it is also used for classification. Rules for extraction are produce by using MLF algorithm.