The Centroid Decomposition (CD) is a matrix
decomposition technique that decomposes an input matrix X (consisting of multiple time-series
as columns) into the product of two matrices L (loading matrix) and R
(relevance matrix), such that X = L * RT. CD allows to efficiently
perform recovery of missing values in large time series, both in batch mode and streaming mode.
In addition to the recovery, ReVival can be used as an online calculator to compute the Centroid
Cecomposition and to visualize different sign vector maximization strategies that can be used by CD.
- Mourad Khayati, Ines Arous, Zakhar Tymchenko, and Philippe Cudré-Mauroux. “ORBITS: Online Recovery of Missing Blocks in Multiple Time Series Streams.” In Proceedings of the VLDB Endowment, Vol. 14, 2021
- Mourad Khayati, Alberto Lerner, Zakhar Tymchenko, and Philippe Cudré-Mauroux. “Mind the Gap: An Experimental Evaluation of Imputation of Missing Values Techniques in Time Series.” In Proceedings of the VLDB Endowment, Vol. 13, 2020.
- Ines Arous, Mourad Khayati, Philippe Cudré-Mauroux, Ying Zhang, Martin Kersten, and Svetlin Stalinlov. “RecovDB: Accurate and Efficient Missing Blocks Recovery for Large Time Series.” In 35th IEEE International Conference on Data Engineering (ICDE 2019). Macau, China, 2019.
- Mourad Khayati, Michael H. Böhlen, and Johann Gamper. “Memory-Efficient Centroid Decomposition for Long Time Series.” In IEEE 30th International Conference on Data Engineering (ICDE 2014), Chicago, ICDE 2014, IL, USA, March 31 - April 4, 2014, 100–111, 2014.