Balancing of the Imbalance data classification using Data Intrinsic characteristics

Abstract

Data mining is the current buzz of the data processing world. There are large number of applications which are producing the large amount of data. This data need to be processed for the extraction of the analyzed data. So that this analysis can be used in the decision making purpose in relevant organizations. But the data collected from the row sources are of imbalance in nature. This data needs to be processed to generate the balancing dataset. Because processing imbalance data will be highly in efficient process. There are various types of discrepancies lies into the dataset. In current research paper these different types of discrepancies are to be neutralized. So that processing and analysis of the dataset will become easy and more efficient. These discrepancies lies into the dataset are like small disjuncts, lack of density in data, overlapping between the classes, impact of noisy data, the borderline etc. using different techniques these problems into the dataset are being neutralized. So that balanced data can be inputted for the data mining purpose.