Hybrid Hierarchical Clustering — Piecewise Aggregate Approximation, with Applications

Abstract

PublishedPiecewise Aggregate Approximation (PAA) provides a powerful yet computationally e±cient tool for dimensionality reduction and Feature Extraction (FE) on large datasets compared to previously reported and well-used FE techniques, such as Principal Component Analysis (PCA). Nevertheless, performance can degrade as a result of either regional information insu±ciency or over-segmentation, and because of this, additional relatively complex modi¯cations have subsequently been reported, for instance, Adaptive Piecewise Constant Approximation (APCA). To recover some of the simplicity of the original PAA, whilst addressing the known problems, a distance-based Hierarchical Clustering (HC) technique is now proposed to adjust PAA segment frame sizes to focus segment density on information rich data regions. The e±cacy of the resulting hybrid HC-PAA methodology is demonstrated using two application case studies on non-timeseries data viz. fault detection on industrial gas turbines and ultrasonic biometric face identi¯-cation. Pattern recognition results show that the extracted features from the hybrid HC-PAA provide additional bene¯ts with regard to both cluster separation and classi¯cation performance, compared to traditional PAA and the APCA alternative. The method is therefore demonstrated to provide a robust readily implemented algorithm for rapid FE and identi¯cation for datasets.

Detection of Emerging Faults on Industrial Gas Turbines Using Extended Gaussian Mixture Models

“…Datasets from the identified transient operation can also be used for fault detection through start-up analysis and shutdown analysis and during load changes [6,24,25], which is not included in the current paper. The most relevant features are then extracted from the steady-state data and a statistical "fingerprint" for the extracted features is obtained through the application of GMMOC.…”

“…Feature Extraction. Feature extraction provides an essential tool for reducing the dimensionality of raw data whilst keeping informative features [25]. Many feature extraction techniques have been reported and successfully applied, including the use of the Fast Fourier Transform (FFT) and Discrete Wavelet Transform (DWT), all involving elaborate time-frequency transforms [31,32].…”

“…VBGMMs are able to classify steady-state operation that can occur under full-or part-load conditions (e.g., 50% load). Additionally therefore, as well as identifying steadystate operation, the remaining data, including that associated with start-ups, shutdowns, and load changing conditions, is also naturally separated and it can therefore be used as a preanalysis tool for alternative dynamic scenarios, as reported in [24,25], for instance.…”

A Deep Learning Based Explainable Control System for Reconfigurable Networks of Edge Devices

“…Multi-attribute time series data has been used to identify the similarity between the multiple time series data. In the realm of motif discovery in multivariate sequences, many of the applications are designed to identify motifs that matches all dimensions [10], [16], [17] or a subset of dimensions [9], [18], [20]. Many of the multi-variate approaches apply distancebased methods such as lower bounding of Euclidean distance [14] and Dynamic Time Wrapping (DTW) [15].…”

“…To describe DNN behaviour, the clusters of the DNN response needs to be compared against corresponding environmental conditions. The existing clustering algorithms, do not provide such functionality [9], [10].…”