TS-MC: Two Stage Matrix Completion Algorithm for Wireless Sensor Networks

Abstract

Wireless sensor network (WSN) data is prone to huge losses and corruption. Hence, the existing matrix completion algorithms experience high estimation errors in such scenarios. Therefore, a robust matrix completion algorithm is required for WSN data to meet the above challenges. This paper proposes a robust "two stage matrix completion (TS-MC)" algorithm to recover data from missing and corrupted values. The proposed TS-MC algorithm consists of two stages. For the first stage, two different methods have been proposed for recovering the incomplete data that exploit the double DCT sparsity as WSN data varies smoothly in both time and spatial domain. In the second stage, the recovered data of the first stage is de-noised in the matrix factorization framework, wherein the rank of the data is estimated from the data recovered from the first stage. Simulations are performed on two real datasets of Intel Lab and Data Sensing Lab. Results demonstrate that the proposed TS-MC algorithm achieves high accuracy even when 90% of the data is missing.