ABSTRACT
The handling of the precise classification of medical imbalanced Big Data sets has set off a high-priority alarm in real-world applications. Standard algorithms and classifiers generally lead to the imprecise classification of imbalanced data sets. This chapter illustrates an enhanced data level oversampling technique, namely the Adjacent Extreme Mix Neighbours OverSampling Technique (AEMNOST). It proficiently handles the classification of medical imbalanced Big Data sets, incorporating its valuable data characteristics. The performance of the proposed technique (AEMNOST) is confirmed mostly across the UCI ML datasets. The validation of outcomes is planned with the standard assessment parameters like Geometric Mean and Area Under Curve values over three renowned classifiers. The experimentation is carried out on the Apache Hadoop framework underlying the MapReduce environment. The results obtained clearly demonstrate the superiority of AEMNOST over the other benchmarking techniques.
