Soft computing refers to a collection of computational techniques in computer science , artificial intelligence , machine learning and other engineering disciplines, which attempt to study, model, and analyze very complex phenomena, for which many conventional methods do not lead to low cost, and complete solutions. The key areas of soft computing are Neural networks , Fuzzy logic and Evolutionary computation to name a few. Fuzzy logic is provides a tool for managing knowledge discovery from uncertain data. Recognition of abnormal patterns in consumer management is very complicated and very much dependant on human behavior. This scenario indicates that experts are not capable of recognising and detecting these abnormal patterns. The best approach is to automatically generate rules that can be used to detect these abnormal patterns from large sets of data. Recently data mining techniques including Decision trees and Neural Networks have been deployed to recognise abnormal patterns in few applications like credit cards, Internet and communication networks. Electricity distribution companies are suffering from unfair use of electricity by some costing the companies huge amount of money. In this thesis, Electricity use of customers for one year period was evaluated for probable abnormal patterns based on data mining algorithms, for fuzzy c-means and fuzzy inference system. studying amount of actual deviation electrical energy consumption from modeled electrical energy consumption of electricity utility company customers, Based on normal pattern in an area, a fuzzy inference system generates ranked list of customers. By comparing of this list with actual examination of customers the performance of model with increases number of clusters improve. The implementing of this model on electricity customers showed that the purposed system is capable to detect suitable percentage of abnormal patterns.
