A Novel Technique on Class Imbalance Big Data using Analogous over Sampling Approach

摘要

Big data consists of huge volumes of data which are used to analyses and discover the hidden knowledge. Class imbalance nature is a conventional issue which is present in all real world datasets. The class imbalance nature in the big data reduces the performance of the existing classification algorithms. In this paper, we propose a novel algorithm known as Over Sampling on Imbalance Big Data (OSIBD) which uses analogous oversampling strategy to improve the knowledge discovery from the class imbalance big datasets. The experimental simulations are conducted on eight moderately large class imbalance datasets which are obtained from UCI machine learning repository. The experimental results suggest that the proposed OSIBD algorithm had performed better than the existing C4.5 algorithm on class imbalance big datasets.