The Application Research On The DM Based Line Loss Automatically Generated System

Data Mining is a very popular subject in recent years, it is a process that extract hidden, unknown, potential and value information or models from a large amount of data that is stored database or data warehouse. It is considered to be one of important leading in the database system. And clustering analysis is one of the very important technologies in the data mining, it is based on a simple idea that"similar things always like to gather together". It is this kind of process which divides those physical or abstract objects into some categories or clusters that are composed of some similar objects. It makes the same cluster of objects have a high similarity and the different clusters of objects have larger difference, it can solve the numerical incomplete problems in data pre-processing stage. But the existence of isolate pots is not always meaningless. And it may be very useful, the outlier data mining is very popular in its'research.In this text, based on analyzing some kinds of outlier data mining ways , the author improves distance based outlier algorithm, and what's more applies it on the line loss automatically generated system of the Dalian power supply company.The main tasks in this text:(1) Elaborating the related theories of Data Mining.(2) Make a description about the related ideas of the Clustering Algorithm.(3) Make a description about outlier data mining and some outlier DM algorithm.(4) Making related researches for line loss automatically generated system of the Dalian power supply company(included : background, identify issues)(5) Based on the actual situation of this system, improved distance based outlier algorithm and use it on the line loss automatically generated system of the Dalian power supply company, and realizing the goals of reducing losses and fight back the act of stealing electricity.