Variant Of Principal Component Analysis Research Articles

Bridging HIV-1 Cellular Latency and Clinical Long-Term Non-Progressor: An Interactomic View

Development of an effective HIV management is enticed by the fact that long-term non-progressors (LTNP) restrict viral replication spontaneously, but is hindered by HIV-1 latency. Given that the most overlapping characteristics found between HIV-1 LTNP and latency, detailed analysis of the difference would disclose the essentials of latency. In this study, microarray data from our previous study was combined with HIV-1 latency and LTNP data obtained from NCBI GEO database. Principal variance component analysis and hierarchical clustering verified the removal of batch effect across platform. The analysis revealed a total of 456 differential expressed genes with >2-fold change and B-statistic >0. Bayesian inference was used to reconstitute the transcriptional network of HIV-1 latency or LTNP, respectively. Gene regulation was reprogrammed under different disease condition. By network interference, KPNA2 and ATP5G3 were identified as the hubs in latency network which mediate nuclear export and RNA processing. These data offer comparative insights into HIV-1 latency, which will facilitate the understanding of the genetic basis of HIV-1 latency in vivo and serve as a clue for future treatment dealing with key targets in HIV-1 latency.

Neural Network Implementations for PCA and Its Extensions

Many information processing problems can be transformed into some form of eigenvalue or singular value problems. Eigenvalue decomposition (EVD) and singular value decomposition (SVD) are usually used for solving these problems. In this paper, we give an introduction to various neural network implementations and algorithms for principal component analysis (PCA) and its various extensions. PCA is a statistical method that is directly related to EVD and SVD. Minor component analysis (MCA) is a variant of PCA, which is useful for solving total least squares (TLSs) problems. The algorithms are typical unsupervised learning methods. Some other neural network models for feature extraction, such as localized methods, complex-domain methods, generalized EVD, and SVD, are also described. Topics associated with PCA, such as independent component analysis (ICA) and linear discriminant analysis (LDA), are mentioned in passing in the conclusion. These methods are useful in adaptive signal processing, blind signal separation (BSS), pattern recognition, and information compression.

P3‐271: Automated index of semantic verbal fluency in mild cognitive impairment and Alzheimer's disease

Background: Semantic verbal fluency (SVF), widely reported in AD and MCI, often show early and disproportionate decline relative to other language, attention, and executive abilities. Successful performance on the SVF test depends on how well conceptual information is organized into related clusters and whether the patient is able to use an efficient strategy to access these clusters. Current methods for clustering and switching behavior assessment are labor-intensive and subjective. We developed and tested an automated computational linguistic approach to cluster words generated on the SVF test. Methods: Participants were a random sample from the Mayo Clinic Alzheimer’s Disease Registry and the Mayo Clinic Study of Aging (20 probable AD, 21 MCI, 19 controls; controlled for age mean 72 y.o.). All participants underwent a cognitive assessment including the paper-andpencil SVF test subsequently converted to electronic form and a short test of mental status (STMS). The test was analyzed using a novel automatic clustering assessment tool based on semantic relatedness between pairs of words calculated using a variant of principal components analysis. Traditional SVF scores were compared to automatically computed mean cluster size (MCS) and a measure of cumulative relatedness between all pairs of words (CuRel) using logistic regression modeling to distinguish between diagnostic groups and linear regression to predict STMS scores. Results: Classification of MCI vs. controls with SVF score as the only predictor in logistic regression model resulted in smaller area under the curve (AUC) estimates for (MCI vs. controls AUC 1⁄4 0.70; AD vs. controls AUC 1⁄4 0.87) than classification with SVF/CuRel score and MCS as predictors (MCI vs. controls AUC 1⁄4 0.70 and 0.77 for CuRel; AD vs. controls AUC 1⁄4 0.97 and 0.90 for CuRel). Linear regression models that included the SVF score together with MCS resulted in significantly better fit with STMS scores than models with SVF score as the sole predictor (r 2 1⁄4 0.55 vs. r 2 1⁄4 0.38, respectively). Conclusions: Our preliminary study indicates that automatically assessed clustering behavior on the SVF test provides complementary information to traditional SVF scoring. Our computerized approach is objective, reproducible and easily scalable to large numbers of participants.

Multivariate statistics for the differentiation of erythropoietin preparations based on intact glycoforms determined by CE-MS

Owing to the increasing number of erythropoietin biosimilars being approved, the comparison of different erythropoietin preparations in the pharmaceutical area is gaining in importance. Erythropoietin has a distinct natural heterogeneity arising from its glycosylation, which shows strong composition variations. This heterogeneity increases the complexity of the analysis of erythropoietin considerably, but may also be used to distinguish different preparations. Here, a method is presented for the differentiation of various erythropoietin preparations by capillary electrophoresis-mass spectrometry and the subsequent application of multivariate statistics. Relative peak areas of selected intact erythropoietin isoforms were used as variables in principal component analysis and hierarchical agglomerative clustering. Both of these strategies were suited for the clear differentiation of all erythropoietin preparations, including marketed products and preproduction preparations, which differ in the manufacturer, the production cell line, and the batch number. By this means, even closely related preparations were distinguished on the basis of the combined information on the antennarity, the sialoform, and the acetylation of the observed isoforms.

Airway obstruction lability helps distinguish levels of disease activity in asthma

Classifying disease activity in asthma relies on clinical and physiological variables, but these variables do not capture all aspects of asthma that distinguish levels of disease activity. We used data from two pivotal trials of montelukast in asthma to classify disease activity as "high" or "low". We performed a principal component analysis (PCA) of disease activity using 21 efficacy outcome variables, including several novel derived outcome variables reflecting clinical and airway obstruction lability. Then we performed discriminant analysis (DA) based on disease activity classification. PCA revealed 6 factors (daytime asthma control, nighttime-predominant asthma control, airway obstruction, exacerbations, clinical lability, airway obstruction lability) that explained 76% of the variance between outcome variables. Although airway obstruction lability (comprising both diurnal variability in peak expiratory flow and diurnal variability in β-agonist use) accounted for only 6% of the explained variance in PCA, in DA it was more accurate (canonical coefficient 0.75) than traditional measures of asthma severity such as obstruction (-0.54) and daytime control (-0.56) in distinguishing between high and low disease activity. We conclude that airway obstruction lability, a parameter not typically captured in clinical trials, may contribute to more complete assessment of asthma disease activity and may define an emerging clinical target of future therapy.

Defence response of Sitka spruce before and after inoculation with Heterobasidion annosum: 1H NMR fingerprinting of bark and sapwood metabolites

Metabolite fingerprinting of Sitka spruce (Picea sitchensis) bark and sapwood was carried out by (1)H nuclear magnetic resonance after wounding and artificial inoculation with the white rot fungus Heterobasidion annosum sensu stricto. The aim was to determine whether metabolites would differ in clones showing differing levels of susceptibility to H. annosum, in the fungal as compared with the control treatment (wounding, no fungus) and the reference (healthy sample collected at 0 days), at two different locations on the host, and at different sampling times (3 and 43 days after treatment). The results suggested that different metabolic processes occur in bark and sapwood after wounding and fungal inoculation, compared with healthy samples collected before treatment: In bark, greater peaks were elicited in the aromatic region whereas, in sapwood, lower amounts of all metabolites were observed in inoculated samples, compared with healthy samples. Multivariate statistical analysis carried out with analysis of variance-principal component analysis showed highly significant effects of reference, location, and time (PC1), and significant effects of clone and fungus. Differences between clones were apparent in sapwood but not in bark and were due to peaks in the aliphatic and carbohydrate regions. Over time, in bark, there was a decrease in carbohydrate peaks, followed by an increase in aliphatic and aromatic peaks. Sapwood, by contrast, showed a decrease in all peaks, followed by an increase in carbohydrate and aliphatic peaks. Changes in carbohydrate levels were observed within the lesion compared with the more distal location in both bark and sapwood.

Kernel Principal Component Analysis for dimensionality reduction in fMRI-based diagnosis of ADHD

This study explored various feature extraction methods for use in automated diagnosis of Attention-Deficit Hyperactivity Disorder (ADHD) from functional Magnetic Resonance Image (fMRI) data. Each participant's data consisted of a resting state fMRI scan as well as phenotypic data (age, gender, handedness, IQ, and site of scanning) from the ADHD-200 dataset. We used machine learning techniques to produce support vector machine (SVM) classifiers that attempted to differentiate between (1) all ADHD patients vs. healthy controls and (2) ADHD combined (ADHD-c) type vs. ADHD inattentive (ADHD-i) type vs. controls. In different tests, we used only the phenotypic data, only the imaging data, or else both the phenotypic and imaging data. For feature extraction on fMRI data, we tested the Fast Fourier Transform (FFT), different variants of Principal Component Analysis (PCA), and combinations of FFT and PCA. PCA variants included PCA over time (PCA-t), PCA over space and time (PCA-st), and kernelized PCA (kPCA-st). Baseline chance accuracy was 64.2% produced by guessing healthy control (the majority class) for all participants. Using only phenotypic data produced 72.9% accuracy on two class diagnosis and 66.8% on three class diagnosis. Diagnosis using only imaging data did not perform as well as phenotypic-only approaches. Using both phenotypic and imaging data with combined FFT and kPCA-st feature extraction yielded accuracies of 76.0% on two class diagnosis and 68.6% on three class diagnosis—better than phenotypic-only approaches. Our results demonstrate the potential of using FFT and kPCA-st with resting-state fMRI data as well as phenotypic data for automated diagnosis of ADHD. These results are encouraging given known challenges of learning ADHD diagnostic classifiers using the ADHD-200 dataset (see Brown et al., 2012).

The chemometrics approach applied to FTIR spectral data for the analysis of rice bran oil in extra virgin olive oil

Among eleven studied vegetable oils, rice bran oil (RBO) has the close similarity to extra virgin olive oil (EVOO) in terms of FTIR spectra, as shown in the score plot of first and second principal components. The peak intensities at 18 frequency regions were used as matrix variables in principal component analysis (PCA). Consequently, the presence of RBO in EVOO is difficult to detect. This study aimed to use the chemometrics approach, namely discriminant analysis (DA) and multivariate calibrations of partial least square and principle component regression to analyze RBO in EVOO. DA was used for the classification of EVOO and EVOO mixed with RBO. Multivariate calibrations were exploited for the quantification of RBO in EVOO. The combined frequency regions of 1200–900 and 3020–3000 cm − 1 were used for such analysis. The results showed that no misclassification was reported for the classification of EVOO and EVOO mixed with RBO. Partial least square regression either using normal or first derivative FTIR spectra can be successfully used for the quantification of RBO in EVOO. In addition, analysis of fatty acid composition can complement the results obtained from FTIR spectral data.

Comparison of two robust PCA methods for damage detection in presence of outliers

Statistical methods such as Principal Component Analysis (PCA) are suffering from contaminated data. For instance, variance and covariance as vital parts of PCA method are sensitive to anomalous observation called outliers. Outliers, who are usually, appear due to experimental errors, are observations that lie at a considerable distance from the bulk of the observations. An effective way to deal with this problem is to apply a robust, i.e. not sensitive to outliers, variant of PCA. In this work, two robust PCA methods are used instead of classical PCA in order to construct a model using data in presence of outliers to detect and distinguish damages in structures. The comparisons of the results shows that, the use of the mentioned indexes based on the robust models, distinguish the damages much better than using classical one, and even in many cases allows the detection where classic PCA is not able to discern between damaged and non-damaged structure. In addition, two robust methods are compared with each other and their features are discussed. This work involves experiments with an aircraft turbine blade using piezoelectric transducers as sensors and actuators and simulated damages.

잡음 민감성이 향상된 주성분 분석 기법의 비선형 변형

Principal component analysis (PCA) is a well-known method for dimensionality reduction and feature extraction while maintaining most of the variation in data. Although PCA has been applied in many areas successfully, it is sensitive to outliers and only valid for Gaussian distributions. Several variants of PCA have been proposed to resolve noise sensitivity and, among the variants, improved robust fuzzy PCA (RF-PCA2) demonstrated promising results. RF-PCA, however, is still a linear algorithm that cannot accommodate non-Gaussian distributions. In this paper, a non-linear algorithm that combines RF-PCA2 and kernel PCA (K-PCA), called improved robust kernel fuzzy PCA (RKF-PCA2), is introduced. The kernel methods make it to accommodate non-Gaussian distributions. RKF-PCA2 inherits noise robustness from RF-PCA2 and non-linearity from K-PCA. RKF-PCA2 outperforms previous methods in handling non-Gaussian distributions in a noise robust way. Experimental results also support this.

잡음 민감성이 개선된 변형 퍼지 주성분 분석 기법

주성분 분석(PCA)은 차원 축소와 특징 추출을 위해 널리 사용되는 기법 중의 하나이지만 자승 오류의 사용으로 인해 잡음에 민감한 단점이 있다. 이러한 잡음 민감성을 개선하기 위해 다양한 방법이 소개되었고 그 중 improved robust fuzzy PCA(RF-PCA2)는 퍼지 소속도를 이용한 반복적 최적화 기법으로 다른 방법에 비해 우수한 성능을 보였다. 하지만 RF-PCA2 역시 국부적인 최적해에 빠질 수 있으며 그 원인 중 하나는 RF-PCA2 알고리듬이 소속도를 균일한 값으로 초기화시키기 때문이다. 또한 퍼지 소속도를 사용하고 있지만 여전히 목적함수가 자승 오류 최소화에 기초하고 있다는 사실도 그 원인이 된다. 이 논문에서는 RF-PCA2의 이러한 문제점을 개선한 RF-PCA3를 제안한다. 제안하는 알고리듬은 RF-PCA2의 목적 함수를 바탕으로 하고 있다. 여기에 PCA의 목적 함수를 추가하고 초기 소속도 값을 데이터의 분포로부터 계산함으로써 전역 최적해에 가까운 해를 얻을 수 있는 가능성을 높여준다. 이러한 사실들은 실험 결과를 통해 확인할 수 있다. Principal component analysis (PCA) is a well-known method for dimensionality reduction and feature extraction. Although PCA has been applied in many areas successfully, it is sensitive to outliers due to the use of sum-square-error. Several variants of PCA have been proposed to resolve the noise sensitivity and, among the variants, improved robust fuzzy PCA (RF-PCA2) demonstrated promising results. RF-PCA2, however, still can fall into a local optimum due to equal initial membership values for all data points. Another reason comes from the fact that RF-PCA2 is based on sum-square-error although fuzzy memberships are incorporated. In this paper, a variant of RF-PCA2 called RF-PCA3 is proposed. The proposed algorithm is based on the objective function of RF-PCA2. RF-PCA3 augments RF-PCA2 with the objective function of PCA and initial membership calculation using data distribution, which make RF-PCA3 to have more chance to converge on a better solution than that of RF-PCA2. RF-PCA3 outperforms RF-PCA2, which is demonstrated by experimental results.

AUV-based bed roughness mapping over a tropical reef

Identifying fixed bed roughness scales of hydrodynamic relevance to waves and currents is challenging around coral reefs due to their highly inhomogeneous bathymetry. In order to characterize the spatial variability in reef roughness, a quantitative analysis of high-resolution sidescan sonar backscatter is performed for the identification of distinct substrates around a tropical reef and is related to echo sounder-based roughness measurements. Data were collected in the vicinity of the Kilo Nalu Observatory on the south shore of Oahu using sidescan sonar and a narrow beam echo sounder incorporated in a REMUS-100 (Remote Environmental Monitoring UnitS) autonomous underwater vehicle (AUV). With basic statistics and principal component analysis of variables derived from the backscatter data, it is possible to discriminate between areas of rough reef, bare reef, and rippled sand. Echo sounder-derived spectral analysis did not reveal dominant length scales. However, by combining the seabed classification obtained from sidescan measurements with echo sounder data, spectral root mean square (RMS) height values of approximately 3.3 cm and 7.3 cm are assigned to the bare reef and rough reef areas, respectively, for roughness with wavelengths between 0.2 and 6 m.

잡음 민감성이 개선된 퍼지 주성분 분석

주성분 분석(PCA)은 데이터의 차원을 줄이면서 최대의 데이터 변이를 보존하는 기법으로 차원 축소나 피처 추출을 위해 널리 사용되고 있다. 하지만 PCA는 잡음에 민감한 단점이 있으며, 이러한 잡음 민감성을 해결하기 위해 여러 가지 PCA 변형이 제안되었다. 그 중 robust fuzzy PCA(RF-PCA)는 퍼지 소속도를 사용하여 잡음의 영향을 효과적으로 줄일 수 있음이 입증되었다. 하지만 RF-PCA 역시 몇 가지 문제점이 있고, 수렴성이 그 중 하나이다. RF-PCA는 소속도와 주성분을 갱신할 때 서로 다른 목적 함수를 사용하므로 수렴 속도가 느리고 구해지는 해가 국부 최적 해임을 보장하지 않는다. 이 논문에서는 RF-PCA의 문제점을 해결하기 위해 하나의 목적 함수를 이용해 소속도와 주성분을 갱신할 수 있는 방법을 제안한다. 제안한 방법, RF-PCA2는 반복 최적화를 이용함으로써 국부 최적해에 수렴함을 보장하며, RF-PCA에 비해 빠른 수렴 속도를 가지고, 잡음 민감성이 줄어든다. 이러한 사실들은 실험 결과를 통해 확인할 수 있다. Principal component analysis (PCA) is a well-known method for dimension reduction while maintaining most of the variation in data. Although PCA has been applied to many areas successfully, it is sensitive to outliers. Several variants of PCA have been proposed to resolve the problem and, among the variants, robust fuzzy PCA (RF-PCA) demonstrated promising results. RF-PCA uses fuzzy memberships to reduce the noise sensitivity. However, there are also problems in RF-PCA and the convergence property is one of them. RF-PCA uses two different objective functions to update memberships and principal components, which is the main reason of the lack of convergence property. The difference between two functions also slows the convergence and deteriorates the solutions of RF-PCA. In this paper, a variant of RF-PCA, called RF-PCA2, is proposed. RF-PCA2 uses an integrated objective function both for memberships and principal components. By using alternating optimization, RF-PCA2 is guaranteed to converge on a local optimum. Furthermore, RF-PCA2 converges faster than RF-PCA and the solutions found are more similar to the desired solutions than those of RF-PCA. Experimental results also support this.

Analysing population characteristics using geographically weighted principal components analysis: A case study of Northern Ireland in 2001

The analysis of quantitative sources of data on multiple population characteristics is often conducted through the use of some form of multivariate statistical procedure such as principal components analysis (PCA). Such approaches assist in the identification of characteristics that group different populations or that vary between these groups. This paper focuses on two particular problems that are rarely considered in the analysis of multivariate population data. Firstly, it makes use of methods for dealing properly with compositional data (i.e., sets of values that sum to a constant such as 1 (proportions) or 100 (percentages)). Secondly, it uses a geographically weighted variant of PCA (GWPCA) (in addition to standard PCA) to explore variations in the characteristics in the population across the study area. In cases where data are spatially-referenced (and are therefore mappable), standard PCA, like any global summary, captures general trends but may mask marked local variations. GWPCA allows assessment of the representivity of standard PCA by providing locally-derived sets of principal components at all data locations. The methods are illustrated through a case study with focuses on selected characteristics of the population of Northern Ireland as represented in data released from the 2001 Census of Population. Key substantive findings include that the characteristics which most strongly differentiate members of the population are geographically variable.

No evidence for dark energy dynamics from a global analysis of cosmological data

We use a variant of principal component analysis to investigate the possible temporal evolution of the dark energy equation of state, w(z). We constrain w(z) in multiple redshift bins, utilizing the most recent data from Type Ia supernovae, the cosmic microwave background, baryon acoustic oscillations, the integrated Sachs-Wolfe effect, galaxy clustering, and weak lensing data. Unlike other recent analyses, we find no significant evidence for evolving dark energy; the data remains completely consistent with a cosmological constant. We also study the extent to which the time-evolution of the equation of state would be constrained by a combination of current- and future-generation surveys, such as Planck and the Joint Dark Energy Mission.

Annual growth rings of the shrub Anarthrophyllum rigidum across Patagonia: Interannual variations and relationships with climate

Annual growth rings are well-defined in some shrubby species distributed along the Patagonian steppe and provide useful information to identify the environmental factors that influence the radial growth of shrubs in the region. However, little is known about variations in ring width from these species and their relationships with local environment. In this paper we present 18 growth ring width chronologies covering the last 2–3 decades from the shrub Anarthrophyllum rigidum (Fabaceae) encompassing most of its natural range of distribution in Argentina. Interannual variations in growth were first analyzed collectively to identify common regional patterns. In addition, shrub growth at each site was compared with local climate records and site-specific characteristics. Based on a principal component analysis (PCA), sites were joined in four major groups related to the site proximity to the Andes (foothills versus Patagonian plateau) and latitude (north versus south Patagonia). The relatively similar percentage of PCA variance associated with each group reflects environmental differences among sites. Consistent with this observation, we noted large variability in the responses of the A. rigidum woody growth to regional climatic fluctuations. At most sites, this species is favored by abundant winter precipitation, warmer spring and autumn temperatures (controlling the duration of the growing season), cool summers, or a combination of these factors. However, deviations from these patterns are also recorded. Our study represents the first large-scale dendroclimatological analysis of a shrubby species in southern South America. The results presented here provide information on the main patterns and factors affecting the growth of A. rigidum in Patagonia, and may be useful for determining well-suited practices for management and conservation of this ubiquitous species.

Variable Neighborhood Search Heuristics for Selecting a Subset of Variables in Principal Component Analysis

The selection of a subset of variables from a pool of candidates is an important problem in several areas of multivariate statistics. Within the context of principal component analysis (PCA), a number of authors have argued that subset selection is crucial for identifying those variables that are required for correct interpretation of the components. In this paper, we adapt the variable neighborhood search (VNS) paradigm to develop two heuristics for variable selection in PCA. The performances of these heuristics were compared to those obtained by a branch-and-bound algorithm, as well as forward stepwise, backward stepwise, and tabu search heuristics. In the first experiment, which considered candidate pools of 18 to 30 variables, the VNS heuristics matched the optimal subset obtained by the branch-and-bound algorithm more frequently than their competitors. In the second experiment, which considered candidate pools of 54 to 90 variables, the VNS heuristics provided better solutions than their competitors for a large percentage of the test problems. An application to a real-world data set is provided to demonstrate the importance of variable selection in the context of PCA.

Comparative analyses of anthropometry associated with overweight and obesity: PCA and ICA approaches

This study undertakes to explore the co-varying structure in anthropometric variables that might be affected by the recent surge of overweight and obesity. The increase of overweight and obesity makes the distribution of body dimensions asymmetric by the long tail in distribution (skewness, kurtosis). Principal component analysis (PCA) has been well applied to understand the co-varying body dimensions. However, because PCA decomposes covariance/correlation matrix, the effects of overweight and obesity may not be well captured. Independent component analysis (ICA) is a variant of PCA with the additional assumptions of components being non-Gaussian and independent, in which kurtosis is decomposed. PCA and ICA are applied on the anthropometric data from the North American portion of the Civilian American and European Surface Anthropometry Resource (CAESAR) project. ICA yields more interpretable results by visual inspection than corresponding PCA results. The first independent component (IC 1) is associated with hip/thigh circumferences and chest/waist circumferences and has the largest correlation coefficients with body mass index (BMI). Only the second IC shows the overall size factor that reveals gender difference while principal components 1, 2 and 3 show gender difference. The ICs 3 (torso length) and 4 (arm and leg lengths) are associated with individual differences in body dimensions. The ranges of 38 body dimensions are identified in order to satisfactorily meet the anthropometric variations for both males and females. The ICA gives promise of becoming a valuable tool in the field of ergonomics.

SEGMENT CLASSIFICATION OF ECG DATA AND CONSTRUCTION OF SCATTER PLOTS USING PRINCIPAL COMPONENT ANALYSIS

In many medical applications, feature selection is obvious; but in medical domains, selecting features and creating a feature vector may require more effort. The wavelet transform (WT) technique is used to identify the characteristic points of an electrocardiogram (ECG) signal with fairly good accuracy, even in the presence of severe high-frequency and low-frequency noise. Principal component analysis (PCA) is a suitable technique for ECG data analysis, feature extraction, and image processing — an important technique that is not based upon a probability model. The aim of the paper is to derive better diagnostic parameters for reducing the size of ECG data while preserving morphology, which can be done by PCA. In this analysis, PCA is used for decorrelation of ECG signals, noise, and artifacts from various raw ECG data sets. The aim of this paper is twofold: first, to describe an elegant algorithm that uses WT alone to identify the characteristic points of an ECG signal; and second, to use a composite WT-based PCA method for redundant data reduction and better feature extraction. PCA scatter plots can be observed as a good basis for feature selection to account for cardiac abnormalities. The study is analyzed with higher-order statistics, in contrast to the conventional methods that use only geometric characteristics of feature waves and lower-order statistics. A new algorithm — viz. PCA variance estimator — is developed for this analysis, and the results are also obtained for different combinations of leads to find correlations for feature classification and useful diagnostic information. PCA scatter plots of various chest and augmented ECG leads are obtained to examine the varying orientations of the ECG data in different quadrants, indicating the cardiac events and abnormalities. The efficacy of the PCA algorithm is tested on different leads of 12-channel ECG data; file no. 01 of the Common Standards for Electrocardiography (CSE) database is used for this study. Better feature extraction is obtained for some specific combinations of leads, and significant improvement in signal quality is achieved by identifying the noise and artifact components. The quadrant analysis discussed in this paper highlights the filtering requirements for further ECG processing after performing PCA, as a primary step for decorrelation and dimensionality reduction. The values of the parameters obtained from the results of PCA are also compared with those of wavelet methods.

Decomposition and Analysis of Process Variability Using Constrained Principal Component Analysis

Process-induced variability has become a predominant limiter of performance and yield of IC products especially in a deep submicron technology. However, it is difficult to accurately model systematic process variability due to the complicated and interrelated nature of physical mechanisms of variation. In this paper, a simple and practical method is presented to decompose process variability using statistics of the measurements from manufacturing inline test structures without assuming any underlying model for process variation. The decomposition method utilizes a variant of principal component analysis and is able to reveal systematic variation signatures existing on a die-to-die and wafer-to-wafer scale individually. Experimental results show that the most dominant die-to-die variation and wafer-to-wafer variation represent 31% and 25% of the total variance of a large set of manufacturing inline parameters in 65-nm SOI CMOS technology. The process variation in RF circuit performance is also analyzed and shown to contain 66% of process variation obtained with manufacturing inline parameters.