Soft computing-based feature extraction method of abnormal communications in Blockchain-based healthcare systems

Abstract

With the advent of pervasive, dependable, and near-instant wireless connectivity for humans and machines, Blockchain technology is serving as the backbone of digital revolution of this era. For ensuring successful and secured communication in Blockchain databases, an accurate extraction of abnormal communication features is imperative and innovative mechanisms are required to achieve the seamless communication for Blockchain-enabled transactions. In order to achieve this objective, a novel feature extraction method for anomalous communication is proposed in this article for secured Blockchain-enabled transactions. In the proposed technique, wavelet transformation is employed to deconstruct the aberrant network communication signals in high and low frequency bands to detect the anomaly. The corresponding parameters for phase space reconstruction and nonlinear dimensionality reduction methods are devised based on the distribution features of the data in the frequency range. Using the KPCA approach (which combines principal component analysis and kernel learning), the wavelet packet decomposition coefficients are recreated after noise reduction to achieve nonlinear noise of abnormal signals as well as abnormal data. The high-dimensional feature space is mapped to the de-noised abnormal communication to protect the data in Blockchain-based transactions. The principal component is analyzed according to the nonlinear function in the mapped feature space, and the nonlinear function is resolved by using the self-organizing neural network. The results reveal that the proposed strategy works with 92 percent accuracy and have practical implication in real-time Blockchain-based transaction to find the anomalies in the data for safeguarding the Blockchain-based transactions.