A New Fuzzy Adaptive Algorithm to Classify Imbalanced Data

Classification of imbalanced data is a well explored issue in the mining and machine learning community where one class representation overwhelmed by other classes. The Imbalanced distribution natural occurrence real world datasets, so needed to be dealt with carefully get important insights. In case imbalance sets, traditional classifiers have sacrifice their performances, therefore lead misclassifications. This paper suggests weighted nearest neighbor approach fuzzy manner deal this issue. We adapted ‘existing algorithm modification solution’ learn from datasets that classify without manipulating unlike popular balancing methods. K non-parametric classification method mostly used problems. Fuzzy clears belonging an instance classes optimal weights improved concept helping correctly data. proposed hybrid takes care nature reduces inaccuracies appear applications original classifiers. Results show it performs over existing strategies for learning.