Minimum Covariance Determinant Algorithm Research Articles

Structural health monitoring under environmental and operational variations using MCD prediction error

This paper proposes a novel technique that aims at detecting the effect of damage on structural frequency signals as “bad” outliers. To this end, a procedure is developed based on the Variational Mode Decomposition (VMD), Minimum Covariance Determinant (MCD), and Recurrent Neural Network (RNN) with Bi-directional Long-Short Term Memory (BiLSTM) cells. The VMD is first used in a pre-processing stage to denoise the signals and remove the seasonal patterns in them. Then, the proposed method seeks to learn the rules behind calculation of the Mahalanobis distances of the points from their distribution, using the parameters obtained from the MCD algorithm, through training an RNN on signals obtained from the inferior state of the structure (healthy state). It will be shown that, since the rule behind the effect of damage on the Mahalanobis distances has not been learnt by the trained RNN, the prediction errors of these values will increase significantly as soon as damage occurs using the data obtained from the posterior state of the structure (including damage). The performance of the proposed method is first tested on a numerical example and further validated through solving an experimental example of the Z24 bridge. Moreover, the proposed method is compared against a PCA-based method. The results demonstrate the superiority of the proposed method in long-term condition monitoring of civil infrastructures. The proposed method is an output-only condition monitoring method that requires only a couple of lowest structural natural frequency signals measured over a long-term monitoring of the structure. Therefore, it is recommended for cases when the measurements from the EOV are not available. Also the proposed method can be used along with other output-only or input-out methods to either improve or confirm the validity of their results.

Identifying Outliers in Response Quality Assessment by Using Multivariate Control Charts Based on Kernel Density Estimation

Abstract When monitoring industrial processes, a Statistical Process Control tool, such as a multivariate Hotelling T 2 chart is frequently used to evaluate multiple quality characteristics. However, research into the use of T 2 charts for survey fieldwork–essentially a production process in which data sets collected by means of interviews are produced–has been scant to date. In this study, using data from the eighth round of the European Social Survey in Belgium, we present a procedure for simultaneously monitoring six response quality indicators and identifying outliers: interviews with anomalous results. The procedure integrates Kernel Density Estimation (KDE) with a T 2 chart, so that historical “in-control” data or reference to the assumption of a parametric distribution of the indicators is not required. In total, 75 outliers (4.25%) are iteratively removed, resulting in an in-control data set containing 1,691 interviews. The outliers are mainly characterized by having longer sequences of identical answers, a greater number of extreme answers, and against expectation, a lower item nonresponse rate. The procedure is validated by means of ten-fold cross-validation and comparison with the minimum covariance determinant algorithm as the criterion. By providing a method of obtaining in-control data, the present findings go some way toward a way to monitor response quality, identify problems, and provide rapid feedbacks during survey fieldwork.

Residual misregistration noise estimation in hyperspectral anomalous change detection

We propose a novel method to estimate the first- and second-order statistics of the residual misregistration noise (RMR), which severely affects the performance of anomalous change detection techniques. Depending on the specific distribution of the RMR, the estimates allow for precisely defining the size of the uncertainty window, which is crucial when dealing with misregistration noise, as in the local coregistration adjustment approach. The technique is based on a sequential strategy that exploits the well-known scale-invariant feature transform (SIFT) algorithm cascaded with the minimum covariance determinant algorithm. The SIFT procedure was originally developed to work on gray-level images. The proposed method adapts the SIFT procedure to hyperspectral images so as to exploit the complementary information content of the numerous spectral channels, further improving the robustness of the outliers filtering by means of a highly robust estimator of multivariate location. The approach has been tested on different real hyperspectral datasets with very high spatial resolution. The analysis highlighted the effectiveness of the proposed strategy, providing reliable and very accurate estimation of the RMR statistics.

Fault detection using robust multivariate control chart

We introduce a new multivariate statistical process control chart for fault detection using robust statistics and principal component analysis. The proposed approach consists of two main steps. In the first step, a robust covariance matrix is determined using the minimum covariance determinant algorithm. In the second step, an eigen-analysis of the robust correlation matrix is performed to derive the robust control limits of the proposed multivariate chart. Our experimental results illustrate the much better fault detection performance of the proposed method in comparison with existing statistical monitoring and process controlling charts.