Teed when utilizing this method. GANs automatically understand the properties from the target bio-signal by utilizing competitive networks (i.e., generator and discriminator) . Luo et al.  recommended a QL-IX-55 supplier conditional Wasserstein GAN for EEG data augmentation to enhance the accuracy of emotion recognition. Zhang and Liu  utilized a conditional deep convolutional GAN process to Esflurbiprofen Biological Activity generate artificial EEG information. Nevertheless, GANs need a long coaching time as well as a significant number of information samples . Hence, when only a smaller number of bio-signal samples are available, a GAN can’t create high-quality artificial information.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.Copyright: 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed below the terms and situations from the Inventive Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).Appl. Sci. 2021, 11, 9388. https://doi.org/10.3390/apphttps://www.mdpi.com/journal/applsciAppl. Sci. 2021, 11,two ofCMD procedures happen to be extensively utilised to make stochastic signals due to the fact they contemplate the correlation in between functions . CMD will not require complicated coaching; hence, its calculation time is extremely brief. Additionally, CMD offers high-quality data devoid of an incredibly huge database. Owing to these positive aspects, CMD is an adequate information augmentation strategy for bio-signals. CMD-generated artificial datasets improve the classification accuracy for brainwave, electromyography, and electrocardiography signals [21,22]. For that reason, in this study, CMD was used to produce artificial brainwave signals. Accordingly, this study aimed to develop a more versatile CMD model than previous CMD models. CMD needs random noise to synthesize the artificial data. To preserve the correlations observed within the original information, the mean on the random noise must be zero, and its variance have to be uniform. On the other hand, previous models impose one more restriction on this random noise; they use only a standard standard distribution, despite the fact that this restriction is not associated to correlation preservation. Hence, this study focused on releasing this restriction to provide higher flexibility to the CMD. The proposed model modifies the skewness and kurtosis of random noise by utilizing a generalized regular distribution (GND). Then, the effects of skewness and kurtosis on accuracy have been investigated for brainwave signals. The remainder of this paper is organized as follows. Section 2 describes the motor imagery brainwave dataset utilised in this study and provides a detailed description from the proposed CMD method. Section three describes the artificial brainwave signals generated by the proposed process. The classification accuracies more than distinctive values of GND skewness and kurtosis are also compared. Ultimately, Section 4 summarizes the study and concludes this paper. two. Supplies and Techniques 2.1. Data Description A dataset applied in BCI competitors III (Dataset I) was made use of to investigate the effects of information augmentation on classification . The subject imagined the movement of a left modest finger (Class 1) and tongue (Class 2) for 3 s. The brainwave information (i.e., electrocorticography) have been measured at a 1000 Hz sampling frequency. From the original dataset, 160 samples (80 samples for finger and tongue each and every) have been used in the instruction dataset, and one hundred samples (50 samples per class) had been employed to receive the test ac.