A data imputation method for multivariate time series based on generative adversarial network.

•We present a new GAN based imputation method for multivariate time series (MTS). To the best of our knowledge, our work is the first one that employs GAN to the imputation of MTS.•We propose a new GAN variant which deals with the demerits of existing GANs in modeling MTS distribution and is capable of generate realistic MTS.•Compared with other imputation approaches, the presented MTS imputation method not only achieves a higher imputation accuracy under different missing rates, but also performs more robustly as the missing rate increases.