An empirical experimental evaluation on imbalanced data sets with varied imbalance ratio

摘要

Class imbalance presents a problem when traditional Classification algorithms are applied .In the previous years there are most important substitution and change has been carried out on data classification. Classification of data becomes difficult because of its unbalanced nature. The problem of imbalance class has developed into significant data mining issue. The class imbalance situation arises when one class is rare compared to the other, take place frequently in machine learning applications. Dataset of unbalanced learning is a new concept of machine learning which has applicability in real time, since all the datasets of real time are of unbalanced in nature. Researchers have rigorously studied several techniques to alleviate the problem of class imbalance, including resampling algorithms, ensemble learning and algorithmic modification for transforming vast amounts of skewed data efficiently into information and knowledge representation. In this paper, we conducted an empirical study on imbalance datasets. Experimental Results shows conclusion of some findings using Area Under Curve (AUC), precision, F-Measure, TN-rate TP-rate evaluation metrics.