Dimensionality Reduction Using PCA and SVD in Big Data: A Comparative Case Study

Abstract

With the advancement in technology, data produced from different sources such as Internet, health care, financial companies, social media, etc. are increases continuously at a rapid rate. Potential growth of this data in terms of volume, variety and velocity coined a new emerging area of research, Big Data (BD). Continuous storage, processing, monitoring (if required), real time analysis are few current challenges of BD. However, these challenges becomes more critical when data can be uncertain, inconsistent and redundant. Hence, to reduce the overall processing time dimensionality reduction (DR) is one of the efficient techniques. Therefore, keeping in view of the above, in this paper, we have used principle component analysis (PCA) and singular value decomposition (SVD) techniques to perform DR over BD. We have compared the performance of both techniques in terms of accuracy and mean square error (MSR). Comparative results shows that for numerical reasons SVD is preferred PCA. Whereas, using PCA to train the data in dimension reduction for an image gives good classification output.