Abstract: We formulate a multi-matrices factorization model (MMF) for the missing sensor data estimation problem. The estimation problem is adequately transformed into a matrix completion one. With MMF, an n-by-t real matrix, R, is adopted to represent the data collected by mobile sensors from n areas at the time, T1, T2, ... , Tt, where the entry, Rij, is the aggregate value of the data collected in the ith area at Tj . We propose to approximate R by seeking a family of d-by-n probabilistic spatial feature matrices, U(1), U(2), ... , U(t), and a probabilistic temporal feature matrix, V E Rdxt, where Rj ≈ UT(j)Tj . We also present a solution algorithm to the proposed model. We evaluate MMF with synthetic data and a real-world sensor dataset extensively. Experimental results demonstrate that our approach outperforms the state-of-the-art comparison algorithms.
Keywords: matrix factorization; sensor data; probabilistic graphical model; missing estimation
Export to BibTeX
MDPI and ACS Style
Huang, X.-Y.; Li, W.; Chen, K.; Xiang, X.-H.; Pan, R.; Li, L.; Cai, W.-X. Multi-Matrices Factorization with Application to Missing Sensor Data Imputation. Sensors 2013, 13, 15172-15186.
Huang X-Y, Li W, Chen K, Xiang X-H, Pan R, Li L, Cai W-X. Multi-Matrices Factorization with Application to Missing Sensor Data Imputation. Sensors. 2013; 13(11):15172-15186.
Huang, Xiao-Yu; Li, Wubin; Chen, Kang; Xiang, Xian-Hong; Pan, Rong; Li, Lei; Cai, Wen-Xue. 2013. "Multi-Matrices Factorization with Application to Missing Sensor Data Imputation." Sensors 13, no. 11: 15172-15186.