Multiscale Principal Component Analysis Research Articles

Reservoir Inflow Prediction under GCM Scenario Downscaled by Wavelet Transform and Support Vector Machine Hybrid Models

Climate change has significant impacts on changing precipitation patterns causing the variation of the reservoir inflow. Nowadays, Indonesian hydrologist performs reservoir inflow prediction according to the technical guideline of Pd-T-25-2004-A. This technical guideline does not consider the climate variables directly, resulting in significant deviation to the observation results. This research intends to predict the reservoir inflow using the statistical downscaling (SD) of General Circulation Model (GCM) outputs. The GCM outputs are obtained from the National Center for Environmental Prediction/National Center for Atmospheric Research Reanalysis (NCEP/NCAR Reanalysis). A new proposed hybrid SD model named Wavelet Support Vector Machine (WSVM) was utilized. It is a combination of the Multiscale Principal Components Analysis (MSPCA) and nonlinear Support Vector Machine regression. The model was validated at Sutami Reservoir, Indonesia. Training and testing were carried out using data of 1991–2008 and 2008–2012, respectively. The results showed that MSPCA produced better extracting data than PCA. The WSVM generated better reservoir inflow prediction than the one of technical guideline. Moreover, this research also applied WSVM for future reservoir inflow prediction based on GCM ECHAM5 and scenario SRES A1B.

Fault Detection and Monitoring Using Multiscale Principal Component Analysis at a Sewage Treatment Plant

Safety, environmental regulations, the cost of maintenance and the operation of sewage treatment plants are some of the many reasons researchers have carried out countless research studies into fault detection and monitoring over the years. Conventional principal component analysis (PCA) in particular has been used in the field of fault detection, where the technique is able to separate useful information from multivariate data. However, conventional PCA can only be used on data that has a constant mean, which is rare in sewage treatment plants. Consequently, the success of combining wavelet and conventional PCA has attracted many researchers to apply it to fault detection where the wavelet is capable of separating data into several time scales. The separated data will be approximated to a constant mean. In addition, the conventional PCA only captures the correlation across the data, unlike multiscale PCA (MSPCA) which captures the correlation within the data and across the data. Therefore, in this work, MSPCA is introduced to improve the performance of PCA in fault detection. The objective of this paper is to reduce false alarms that exist in PCA fault detection and monitoring. Data from the Bunus sewage treatment plant (Bunus STP) is used and analysed using conventional PCA with Hotelling’s T2 and the squared prediction error (SPE). MSPCA with Hotelling’s T2 and SPE is used to improve the efficiency of fault detection and monitoring performance in conventional PCA. Therefore, MSPCA is successful in improving conventional PCA in fault detection and monitoring by reducing false alarms.

The effect of multiscale PCA de-noising in epileptic seizure detection.

In this paper we describe the effect of Multiscale Principal Component Analysis (MSPCA) de-noising method in terms of epileptic seizure detection. In addition, we developed a patient-independent seizure detection algorithm using Freiburg EEG database. Each patient contains datasets called "ictal" and "interictal". Window length of 16 s was applied to extract EEG segments from datasets of each patient. Furthermore, Power Spectral Density (PSD) of each EEG segment was estimated using different spectral analysis methods. Afterwards, these values were fed as input to different machine learning methods that were responsible for seizure detection. We also applied MSPCA de-noising method to EEG segments prior to PSD estimation to determine if MSPCA can further enhance the classifiers' performance. The MSPCA drastically improved both the sensitivity and the specificity, increasing the overall accuracy of all three classifiers up to 20%. The best overall detection accuracy (99.59%) was achieved when Eigenvector analysis was used for frequency estimation, and C4.5 as a classifier. The experiment results show that MSPCA is an effective de-noising method for improving the classification performance in epileptic seizure detection.

Effects on stratospheric moistening by rates of change of aerosol optical depth and ozone due to solar activity in extra-tropics

The solar-induced changes in ozone and aerosol optical depth have relative effects on stratospheric moistening at upper troposphere/lower stratosphere region. Wavelet-based multi-scale principal component analysis technique has been applied to de-noise component of quasi-biennial oscillation and El Nino-Southern Oscillation from ozone and aerosol optical depth variations. Rate of change of aerosol optical depth sharply increases indicating a positive gradient whereas rate of change of ozone sharply decreases indicating a negative gradient with solar activity during the years 2004–2010. It is also observed that with increase of rate of change of aerosol optical depth, there is a sharp increase of stratospheric moistening caused by enhanced deep convection. On the contrary, with the increase of stratospheric moistening, there is a sharp decrease of rate of change of ozone resulting in a cross-over between the two parameters. An increase in aerosol optical depth may cause a significant increase in the gradient of vertical temperature profile, as well as formation of cloud condensation nuclei, clouds and hence rainfall. This may lead to formation of strong convective system in the atmosphere that is essential for vertical transfer of water vapour in the tropics percolating tropical tropopause layer and depleting stratospheric ozone in the extra-tropics.

Effect of Multiscale PCA De-noising in ECG Beat Classification for Diagnosis of Cardiovascular Diseases

Current trends in clinical applications demand automation in electrocardiogram (ECG) signal processing and heart beat classification. This paper examines the design of an effective recognition meth...

Cross-correlations and structures of stock markets based on multiscale MF-DXA and PCA

Based on the daily closing prices of US and Chinese stock markets, we conduct empirical analysis of the cross-correlations and multiscale distances among the US stock markets and China stock markets using the proposed multiscale multifractal detrended cross-correlation analysis method (MSMF-DXA). It is demonstrated that stock markets appear to be far more complex than hitherto reported in the studies using a fixed time scale. In order to identify and compare the interactions and structures of stock markets during financial crisis, as well as in the pre-crisis and post-crisis periods, all the entire samples are divided into four sub-periods. We are thus able to deeply examine the dynamics of linkages among stock markets during different considered periods by MSMF-DXA and principal component analysis technology. The empirical results indicate that both two financial crises cause the increase of international transmission of stock markets, but the global financial crisis and Asian financial crisis lead to different effects in stock markets. It is indicated that the markets are more correlated during global financial crisis compared to Asian financial crisis. The cross-correlations, especially for the markets within different geographic region, increased significantly during the crisis. It is shown that stock markets can be described by the first and second principal component during normal and crisis period, respectively. The results presented in this paper are useful to interpret the multiscale properties of stock markets and describe the dynamics of interactions among considered stock markets.

Effect of multiscale PCA de-noising on EMG signal classification for diagnosis of neuromuscular disorders

Different approaches have been applied for quantitative analysis of EMG signals. This study introduces the effect of Multiscale Principal Component Analysis (MSPCA) denoising method in ElectroMyoGram (EMG) signal classification. The effect of the MSPCA denoising method discussed on EMG signal classification. In addition, effect of Multiple Single Classification (MUSIC) feature extraction method presented and compared for the classification of EMG signals. The results were accomplished on the basis of EMG signal data to classify into normal, ALS or myopathic. Furthermore, total accuracy of classifiers such as k-Nearest Neighbor (k-NN), Artificial Neural Network (ANN) and Support Vector Machines (SVMs) were discussed. Significant results were found by using MSPCA denoising method. The comparisons between the developed classifiers were based on a number of scalar performances such as sensitivity, specificity, accuracy, F-measure and area under ROC curve (AUC). The results show that MSPCA de-noising has considerably increased the accuracy as compared to EMG data without MSPCA de-noising.

Multiscale principal component analysis

Principal component analysis (PCA) is an important tool in exploring data. The conventional approach to PCA leads to a solution which favours the structures with large variances. This is sensitive to outliers and could obfuscate interesting underlying structures. One of the equivalent definitions of PCA is that it seeks the subspaces that maximize the sum of squared pairwise distances between data projections. This definition opens up more flexibility in the analysis of principal components which is useful in enhancing PCA. In this paper we introduce scales into PCA by maximizing only the sum of pairwise distances between projections for pairs of datapoints with distances within a chosen interval of values [l,u]. The resulting principal component decompositions in Multiscale PCA depend on point (l,u) on the plane and for each point we define projectors onto principal components. Cluster analysis of these projectors reveals the structures in the data at various scales. Each structure is described by the eigenvectors at the medoid point of the cluster which represent the structure. We also use the distortion of projections as a criterion for choosing an appropriate scale especially for data with outliers. This method was tested on both artificial distribution of data and real data. For data with multiscale structures, the method was able to reveal the different structures of the data and also to reduce the effect of outliers in the principal component analysis.

Spline Curve Modeling Based Gait Recognition

Gait recognition has become an active research area with increasing demand for effective video surveillance systems. This paper deals with an innovative method of modelling human gait with spline curves. The method proposed involves finding the locations of several human joints namely, coxa joint, a pair of knee joints and a pair of ankle joints. The five joints located are used as control points to construct spline curve. Instead of comparing the gait models constructed, for which time complexity is high, we consider the area under the spline curve constructed, which is a linear metric, as our gait feature and construct feature vector containing area signals of the sequence of images considered. DCT (Discrete Cosine Transform) is applied to the feature vector to obtain the feature matrix. The dimensional reduction of the constructed feature matrix is achieved by adopting the method of MSPCA (Multi-scale Principal Component Analysis). The classification of the feature vectors is done using K-NN and Neuro-Fuzzy classifiers, for the subjects considered in CASIA datasets A, B and DTW (dynamic time warping) for the subjects in CASIA dataset C.

Malfunction and Bad Behavior Diagnosis on Domestic Environment

Greenhouse gas emissions from homes arise primarily from fossil fuels burned for heat, the use of products that contain greenhouse gases, and the handling of waste. Human activities are responsible for almost all of the increase in greenhouse gases in the atmosphere over the last 150 years. The household sector is one of the biggest aggregate consumers and this is the reason why increasingly policies have been considering it. One of the key factors in curbing energy consumption in this sector is widely recognized to be due to erroneous behaviors and systems malfunctioning, mainly explained by the lack of awareness of the final user; so, training the final user to energy awareness can be more effective and cheaper than other policies. In this context, energy management in homes is playing, and will play even more in future, a key role in increasing the final consumer awareness towards its own energy consumption and consequently in bursting its active role in smart grids. The aim of this paper is to highlight the economic benefits of low cost intelligent control domestic devices, to identify energy behavior, system status and improve energy efficiency. The scope is to develop interaction between final users to create a network of energy consumption efficiency. The paper presents an application of Multi-scale Principal Component Analysis to diagnose inefficient occupant behavior and systems malfunctioning and suggest good practices of energy conservation.

Use of a (MSPCA) and (SVM) Method for Diagnosis of Motor Faults

This paper uses the combination between support vector machine and multi-scale principal component analysis. For motor fault detection, the principal component model can be established in various scales. Through T2 and Q statistic judgment whether motor can run normally. The experimental results show that the method of combination vector machine and multi-scale principal component analysis is supported to diagnose motor fault. This offers a new method and idea to diagnose motor. This method improves the accuracy of motor fault detection and practical significance.

Robust Extraction of Respiratory Activity From PPG Signals Using Modified MSPCA

The pulse oximeter's photoplethysmographic (PPG) signals can be well utilized for extracting the vital respiratory activity, in addition to saturation and heart rate estimations, avoiding the usage of additional sensor for recording respiratory signal, in turn reducing the number of sensors connected to the patient's body for recording vital signals. In this paper, we present a robust algorithm called modified multi scale principal component analysis (MMSPCA), for extraction of respiratory activity embedded in the PPG signals. The PPG signals are more commonly corrupted by motion artifacts (MA) due to voluntary or involuntary movements of the patients, making it difficult for the algorithms to extract respiratory signals. The problem of extracting respiratory signals from PPGs in the presence of MA is addressed for the first time in this paper. The problem gets aggravated when PPGs are severely afflicted with MAs in situations such as the MA frequency band (usually below 0.2 Hz) overlapping on to the band of respiratory frequencies (0.2-0.4 Hz). In the presented algorithm, the kurtosis and energy contribution levels (ECL) of approximate and detail coefficients are calculated for each wavelet sub-band matrix, generating a modified wavelet sub-band matrix. This makes the presented algorithm based on MMSPCA more robust in the sense that it is made motion resistant by suitably modifying the MSPCA. Functioning of the proposed algorithm is tested on the data recorded from 15 healthy subjects. Each data set consists of intentionally created possible MA noises, viz., vertical, horizontal, waving, and pressing MAs with different breathing patterns. The method is also applied on the recordings available with MIMIC database of Physionet archive. The statistical and error analysis, performed to test the efficacy of the presented MMSPCA algorithm, revealed a very good acceptance for derived respiratory signal, when compared with the originally recorded respiratory signals using classical method. The MMSPCA method clearly outperformed the conventional MSPCA-based method in the presence of MA.

MSPCA with KDE thresholding to support QC in electrical motors production line

In this paper a Fault Detection and Isolation (FDI) procedure is applied for the defects detection and analysis of electrical motors at the end of production line in hood factories. The objective consists of developing a fast and robust methodology to detect defective motors and to identify defects for quality analysis on production line. Using a signal based FDI procedure, an end of a line bench system is designed, which is able to analyze the defects of produced motors. Multi-Scale Principal Component Analysis (MSPCA) is used for defect detection and a Kernel Density Estimation (KDE) algorithm is used for fault isolation on the PCA residual contributions. Also a method to choose the WT levels is adopted. MSPCA with KDE thresholding advantage is demonstrated by experimentations on test bench, using vibration measurements. Experiments show that the stochastic method used to compute thresholds on PCA residuals is robust and at the same time accurate.

Data-Driven Residual-Based Fault Detection for Condition Monitoring in Rolling Mills

We propose a residual-based approach for fault detection in rolling mills which is based on data-driven soft computing techniques. The basic idea is to transform original measurement signals into a feature space by (i) identifying multi-dimensional relationships in the system, (ii) representing the nominal fault-free case, and (iii) analyzing residuals with incremental/decremental statistical techniques. Model identification and fault detection are conducted in a completely unsupervised manner, that is, solely based on the data streams recorded online. Thus, neither annotated samples nor fault patterns/models, which are often very time-intensive and costly to obtain, must be available a priori. We use purely linear models, a new genetic variant of Box-Cox models (termed Genetic Box-Cox) that consider weak non-linearities, and Takagi-Sugeno fuzzy models, which are able to express more complex non-linearities, trained with an extended version of SparseFIS. Using three typical scenarios from rolling mill production, we compare our method to state-of-the-art approaches that are based on principal components analysis and multi scale principal components analysis. The results show that our method outperforms these state-of-the-art approaches.

Feature extraction and classification of imagined motor movement electroencephalogram signals

Brain–Computer Interface (BCI) establishes a communication channel between brain and external world. BCI can control numerous applications, such as controlling a cursor on computer screen, movement of a robotic arm or a wheel chair and many more. The efficiency and accuracy of BCI systems completely relies on efficient preprocessing and classification algorithms. In present work, the reliability of a BCI has been analysed, which is implemented using Electroencephalogram (EEG) signals, recorded from motor imagery, for imagination of three different mental tasks: left fist blink, right fist blink and both fists blink. The recorded EEG signals were primarily filtered out by a low pass Butterworth filter and further preprocessed using multiscale principal component analysis. Statistical parameters have been calculated from preprocessed EEG signals as features. For classification of the EEG signals, support vector machine model is employed.

Feature Extraction by Multi-Scale Principal Component Analysis and Classification in Spectral Domain

Feature extraction of signals plays an important role in classification problems because of data dimension reduction property and potential improvement of a classification accuracy rate. Principal component analysis (PCA), wavelets transform or Fourier transform methods are often used for feature extraction. In this paper, we propose a multi-scale PCA, which combines discrete wavelet transform, and PCA for feature extraction of signals in both the spatial and temporal domains. Our study shows that the multi-scale PCA combined with the proposed new classification methods leads to high classification accuracy for the considered signals.

Fault diagnosis of Tennessee Eastman process with multi-scale PCA and ANFIS

Fault diagnosis in industrial processes are challenging tasks that demand effective and timely decision making procedures under the extreme conditions of noisy measurements, highly interrelated data, large number of inputs and complex interaction between the symptoms and faults. The purpose of this study is to develop an online fault diagnosis framework for a dynamical process incorporating multi-scale principal component analysis (MSPCA) for feature extraction and adaptive neuro-fuzzy inference system (ANFIS) for learning the fault-symptom correlation from the process historical data. The features extracted from raw measured data sets using MSPCA are partitioned into score space and residual space which are then fed into multiple ANFIS classifiers in order to diagnose different faults. This data-driven based method extracts fault-symptom correlation from the data eliminating the use of process model. The use of multiple ANFIS classifiers for fault diagnosis with each dedicated to one specific fault, reduces the computational load and provides an expandable framework to incorporate new fault identified in the process. Also, the use of MSPCA enables the detection of small changes occurring in the measured variables and the proficiency of the system is improved by monitoring the subspace which is most sensitive to the faults. The proposed MSPCA-ANFIS based framework is tested on the Tennessee Eastman (TE) process and results for the selected fault cases, particularly those which exhibit highly non-linear characteristics, show improvement over the conventional multivariate PCA as well as the conventional PCA-ANFIS based methods.

Multichannel ECG Data Compression Based on Multiscale Principal Component Analysis

In this paper, multiscale principal component analysis (MSPCA) is proposed for multichannel electrocardiogram (MECG) data compression. In wavelet domain, principal components analysis (PCA) of multiscale multivariate matrices of multichannel signals helps reduce dimension and remove redundant information present in signals. The selection of principal components (PCs) is based on average fractional energy contribution of eigenvalue in a data matrix. Multichannel compression is implemented using uniform quantizer and entropy coding of PCA coefficients. The compressed signal quality is evaluated quantitatively using percentage root mean square difference (PRD), and wavelet energy-based diagnostic distortion (WEDD) measures. Using dataset from CSE multilead measurement library, multichannel compression ratio of 5.98:1 is found with PRD value 2.09% and the lowest WEDD value of 4.19%. Based on, gold standard subjective quality measure, the lowest mean opinion score error value of 5.56% is found.

Signal feature extraction by multi-scale PCA and its application to respiratory sound classification

Respiratory sound (RS) signals carry significant information about the underlying functioning of the pulmonary system by the presence of adventitious sounds. Although many studies have addressed the problem of pathological RS classification, only a limited number of scientific works have focused in multi-scale analysis. This paper proposes a new signal classification scheme for various types of RS based on multi-scale principal component analysis as a signal enhancement and feature extraction method to capture major variability of Fourier power spectra of the signal. Since we classify RS signals in a high dimensional feature subspace, a new classification method, called empirical classification, is developed for further signal dimension reduction in the classification step and has been shown to be more robust and outperform other simple classifiers. An overall accuracy of 98.34% for the classification of 689 real RS recording segments shows the promising performance of the presented method.

Multiscale PCA based Quality Controlled Denoising ofMultichannel ECG Signals

In this article, it is suggested to apply Multiscale PCA to multichannel ECG signals, for quality controlled denoising. PCA is applied at wavelet scales after forming multivariate data matrices. The quality controlled denoising is a two steps process, (a) selection of multivariate matrices at wavelet scale for PCA based dimension reduction and (b) the choice of number of Principal Component (PC) at each wavelet matrices. The selection of effective PCs is normally based on the cumulative percentage of total variation shown by eigenvalues. For conventional PCA, it is taken 70% to 99% [3]. The diagnostic quality of denoised signal is qualitatively compared with the original signal. The quantitative distortion measures like PRD and WEDD are evaluated [10], [12], [13]. The enhanced multichannel signals contain all the clinical components with required 'PQRST' morphologies of ECG signals while reducing the noise.