A Principal Component Analysis, Sampling and Classifier strategies for dealing with concerns of class imbalance in datasets with a ratio greater than five

This research looks at how the imbalance ratio and classifier affect how well various resampling techniques work when dealing with imbalanced data sets. The research examines the effects on learning when previously imbalanced data is changed into fictitiously balanced class distributions using a variety of resampling strategies. When data sets are significantly uneven, analyses utilising attribute selection, four different classifiers, and six performance evaluation criteria demonstrate that over-sampling (95 percent accuracy) the minority class routinely outperforms under-sampling (77 percent accuracy) the majority class.