Improved Vine Copula-Based Dependence Description for Multivariate Process Monitoring Based on Ensemble Learning

Abstract

This paper proposes a boosting vine copula-based dependence description (BVCDD) method for multivariate and multimode process monitoring. The BVCDD aims to improve the standard vine copula-based dependence description (VCDD) method by establishing an ensemble of submodels from sample directions based on a boosting strategy. The generalized Bayesian inference-based probability (GBIP) index is introduced to assess the degrees of a VCDD model (submodel) to depict different samples, which means how likely an observation is under the probabilistic model for the system. Every sample is weighted individually according to the depiction degree. The weights are then used to choose a certain number of samples for each succeeding submodel. In this way, the samples with large error in the preceding model can be selected for training the next submodel. Moreover, the number of submodels as well as the number of training samples chosen for every submodel are determined adaptively in the ensemble learning process. The proposed BVCDD method can not only solve weak sample problems but also remove redundant information in samples. To examine the performance, empirical evaluations have been conducted to compare the BVCDD method with some other state-of-the-art methods in a numerical example, the Tennessee Eastman (TE) process, and an acetic acid dehydration process. The results show that the developed BVCDD models are superior to those obtained by the counterparts on weak samples in both accuracy and stability.

Distributed Statistical Process Monitoring Based on Multiblock Canonical Correlation Analysis

“…To conceptually validate the superiority and effectiveness of the proposed MBCCA-based distributed process monitoring approach over its counterparts, comparisons with MBPCA-, DPCA-, and DICA-based methods are implemented with application to the Tennessee Eastman (TE) process. − …”

“…The TE process has become a well-known benchmark platform for validating different data-driven process monitoring methods − because of its relatively complicated configuration. The flowchart of the TE benchmark process is depicted in Supporting Information Figure S1, which consists of multiple units: a reactor, a condenser, a vapor–liquid separator, a recycle compressor, a stripper, and a plant-wide control system .…”

Nonlinear Non-Gaussian and Multimode Process Monitoring-Based Multi-Subspace Vine Copula and Deep Neural Network

“…To detect process faults, a generalized Bayesian inference probability (GBIP) index had been used to monitor the process status. To improve the quality of VCDD, Zhou et al proposed boosting VCDD by establishing an ensemble of sub-models from sample directions based on a boosting strategy and combined the VCDD method and active learning strategy to solve the problem of missing samples . To enhance the robustness of the VCDD method, a copula double-subspace (CDS) including margin distribution subspace (MDS) and dependence structure subspace (DSS) was proposed .…”

Artificial Intelligence Approach To Investigate the Longevity Drug

“…It is reported that a new drug optimization method based on the nanopore force spectroscopy has been proposed. By all-atom molecular dynamics simulations, the binding strength between a drug molecule and a target protein could be electrically detected. , Ensemble learning is a method that integrates multiple machine learning algorithms by using some rules, so as to achieve better effects than a single machine learning algorithm . There are two kinds of ensemble learning; one is boosting, and another is bagging.…”

“…30,31 Ensemble learning is a method that integrates multiple machine learning algorithms by using some rules, so as to achieve better effects than a single machine learning algorithm. 32 There are two kinds of ensemble learning; one is boosting, and another is bagging. The former's characteristic is the dependencies between weak learners, while for the latter, there are no dependencies between weak learners; all weak learners could be parallel fitted.…”