The major focus of this study is to describe the structure of a solution designed for robustly detecting and delineating the arterial blood pressure (ABP) signal events. To meet this end, first, the original ABP signal is pre-processed by application of à trous discrete wavelet transform (DWT) for extracting several dyadic scales. Then, a fixed sample size sliding window is moved on the appropriately selected scale and in each slid, six features namely as summation of the nonlinearly amplified Hilbert transform, summation of absolute first-order differentiation, summation of absolute second-order differentiation, curve length, area and variance of the excerpted segment are calculated. Then, all feature trends are normalized and utilized to construct a newly proposed principal components analyzed geometric index (PCAGI) (to be used as the segmentation decision statistic (DS)) by application of a linear orthonormal projection. After application of an adaptive-nonlinear transformation for making the DS baseline stationary, the histogram parameters of the enhanced DS are used to regulate the α-level Neyman–Pearson classifier for false alarm probability (FAP)-bounded delineation of the ABP events. In order to illustrate the capabilities of the presented algorithm, it was applied to all 18 subjects of the MIT-BIH Polysomnographic Database (359,000 beats) and the end-systolic and end-diastolic locations of the ABP signal as well as dicrotic notch pressure were extracted and values of sensitivity and positive predictivity Se = 99.86% and P+ = 99.95% were obtained for the detection of all ABP events. High robustness against measurement noises, acceptable detection-delineation accuracy of the ABP events in the presence of severe heart valvular and arrhythmic dysfunctions within a tolerable computational burden (processing time) and having no parameters dependency to the acquisition sampling frequency can be mentioned as the important merits and capabilities of the proposed PCAGI-based ABP events detection-segmentation algorithm.
Unsupervised multilayer long short-term memory autoencoder (LSTM-AE) models are proposed for monitoring nonlinear batch processes. The methodology is demonstrated for a simulation-based study of an industrialscale penicillin process and for an industrial vaccine manufacturing process, using production data. The LSTM-AE model was trained with two different loss functions: minimizing mean square error (MSE) between the input and reconstructed data and maximizing the average fault detection rate (FDR) in the training data set. Two algorithms are also proposed for obtaining contribution plots for the diagnosis of faults. For the industrial case study, where the faults are not known a priori, the contribution plots are found to be a valuable tool for identifying possible sources of faults. Furthermore, a semisupervised procedure has been proposed to select the normal process region for training the model. Two metrics are also presented to evaluate the performance of the proposed methodology: one for the simulator case study in which fault knowledge is available and one for the industrial case study in which fault knowledge is not available a priori. The proposed unsupervised algorithms exhibit a clear improvement in accuracy over linear methods or nonlinear techniques that do not explicitly account for dynamic behavior.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.