|Author(s)||Phan Thi-Thu-Hong1, 2, Poisson Caillault Emilie1, Bigand Andre1, Lefebvre Alain3|
|Affiliation(s)||1 : Univ Littoral Cote dOpale, EA LISIC 4491, F-62228 Calais, France.
2 : Vietnam Natl Univ Agr, Dept Comp Sci Hanoi, Hanoi, Vietnam.
3 : IFREMER, LER BL, F-62321 Boulogne Sur Mer, France.
|Meeting||27th IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Int House Japan, Tokyo, JAPAN, SEP 25-28, 2017|
|Source||Proceedings of the 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP). 2017. 6 p.|
|Keyword(s)||Imputation, Uncorrelated multivariate time series, Missing data, Dynamic Time Warping, Similarity measures|
Missing data are inevitable in almost domains of applied sciences. Data analysis with missing values can lead to a loss of efficiency and unreliable results, especially for large missing sub-sequence(s). Some well-known methods for multivariate time series imputation require high correlations between series or their features. In this paper, we propose an approach based on the shape-behaviour relation in low/un-correlated multivariate time series under an assumption of recurrent data. This method involves two main steps. Firstly, we find the most similar sub-sequence to the sub-sequence before (resp.after) a gap based on the shape-features extraction and Dynamic Time Warping algorithms. Secondly, we fill in the gap by the next (resp.previous) sub-sequence of the most similar one on the signal containing missing values. Experimental results show that our approach performs better than several related methods in case of multivariate time series having low/non-correlations and effective information on each signal.
Phan Thi-Thu-Hong, Poisson Caillault Emilie, Bigand Andre, Lefebvre Alain (2017). DTW-Approach for uncorrelated multivariate time series imputation. Proceedings of the 2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP). 2017. 6 p. https://archimer.ifremer.fr/doc/00429/54082/