Tikhonov Kullback Leibler Vuong Logistic Machine Learning Classifier for Early Disease Diagnosis Over Big Data

Abstract

With big data widening in healthcare groups, precise investigation of medical data conveniences early disease detection. However, the analysis accuracy is reduced when the parallel processing of medical data is not performed. Moreover, with curse of dimensionality as several regions discloses distinctive facets and if not properly filtered, relevant information’s are also discarded which may reduce the early prediction of disease outbreaks. To address these issues, in this work, a method using machine learning technique called, Polynomial Tikhonov Entropy and Kullback Vuong Logistic Classifier (PTE-KVLC) is presented. First, Inverse Polynomial Map Reduce Pre-processing is applied to the input data that both minimizes the signal to noise ratio and obtains computationally efficient features via parallel processing. This is turn provides a mean for early detection of epileptic seizures. Second, the feature extraction model is based on Entropy Tikhonov Regularization and is applied to the pre-processed features to identify the features pertinent to seizures. These features are then selected and fed into a Kullback–Leibler Vuong and Logistic Regressive Machine Learning Classifier for early epileptic seizure recognition. Experimental results demonstrate that the proposed method significantly classifies the epileptic seizure classes by means of specificity, sensitivity, and accuracy.