Reconstruction of Missing Data Completely at Random for Trains Based on Improved GAN

Reconstruction of missing data for heavy-haul trains is critical to ensuring safe train operation. However, existing generative model training methods require a complete dataset, making it difficult for them to address the issue of missing data completely at random. To address this issue, this study proposes a new attention-generative adversarial network to reconstruct missing data. First, a mask matrix is designed to locate the missing data, and the gradient descent algorithm is applied in combination with the output probability matrix of the discriminator so that the mask matrix can still fill up the data well in the case of an incomplete data set. Subsequently, the prompt matrix is derived based on the mask matrix to solve the problem of model overfitting and accelerate the convergence. Finally, an attention mechanism is introduced into the entire generative adversarial network to improve the expression of data features using the feature extraction network. The experimental results show that the mean square error and mean absolute error percentage indexes of reconstruction accuracy can be maintained below 1.5 for measurement data at different missing rates, and the reconstructed data can also well conform to the distribution law of measurement data.