Nonlinear Principal Component Analysis Research Articles

Phenotypic diversity of Greek dill (Anethum graveolens L.) landraces

ABSTRACTDill has multiple culinary and medicinal purposes and the use of their landraces into a plant breeding program, requires the analysis of their phenotypic diversity. In this study, 33 Greek dill landraces collected from diverse areas were evaluated using traits based on UPOV descriptor list. Phenotypic diversity was assessed using Shannon-Weaver diversity index (H΄) and non-linear principal component analysis. Grouping of landraces was further performed through hierarchical cluster analysis. The H' index ranged from 0.32 (stem waxiness) to 0.98 (density of foliage) with a mean value of 0.68 indicating a high level of phenotypic diversity. High H' values were recorded for the foliage width, stem color and anthocyanin coloration. Multivariate analysis revealed three common genetic groups: 1) North mainland Greece, 2) Aegean islands and 3) Central mainland Greece. The landraces’ heterogeneity was attributed to various traits linked to specific geographic origin, such as early time of flowering and high stem waxiness allied with the landraces originated from the Aegean islands. Greek dill landraces revealed useful variation on yield component traits related to fresh herb weight and to seed production, such as high number of leaves/plant and large diameter of main umbel that can be promptly exploited in breeding programs.

Evaluation of effect of unsupervised dimensionality reduction techniques on automated arrhythmia classification

Abstract Automation in cardiac arrhythmia classification helps medical professionals to make accurate decisions upon the patient’s health. The aim of this work is to evaluate the performance of five different linear and nonlinear unsupervised dimensionality reduction (DR) techniques namely principal component analysis (PCA), fast independent component analysis (fastICA) with tangential, kurtosis and Gaussian contrast functions, kernel PCA (KPCA) with polynomial kernel, hierarchical nonlinear PCA (hNLPCA) and principal polynomial analysis (PPA) on classification of cardiac arrhythmias using probabilistic neural network classifier (PNN). The design phase of the classification model comprises of the following stages: preprocessing of the cardiac signal by eliminating detail coefficients that contain noise, feature extraction through Daubechies wavelet transform, dimensionality reduction through unsupervised DR techniques, and arrhythmia classification using PNN. PCA is a widely used DR technique for mapping high dimensional data to its low dimensional representation. But real world data like electrocardiogram (ECG) signals are complex and nonlinear in nature. This work concentrates on performance analysis of four nonlinear DR techniques and conventional linear PCA technique on classification of cardiac arrhythmias. Entire MIT-BIH arrhythmia database is used for experimentation. The experimental results demonstrate that the combination of PNN classifier (at spread parameter, σ = 0.4) and fastICA DR technique with tangential contrast function exhibit highest F score of 99.83% with a minimum of 10 dimensions. hNLPCA and KPCA requires more computation time for low dimensional mapping. PPA performs about 10% better than PCA and serves intermediate between linear and nonlinear techniques.

A Deterministic Feedback Model for Safe Driving based on Nonlinear Principal Analysis Scheme

For intelligent driver assistant service, research works have been conducted using sensors to analyze drivers’ behavior and score for safe driving. There have been many studies on safe driving, but there was a limit on providing the driver with a comprehensive solution for safe driving. The studies mainly focused on a method to quantify the driving characteristics or the indicators that can be used to assess the driving risk. In order to practically utilize those indicators, it is necessary to provide the driver with the feedback on how to change which controllable factors to fix the indicators to desired values. However, there has been no study to provide feedback on interpretable factors. To improve the problem, in this paper, we propose a model that provides comprehensive feedback to the driver by offering quantitative driving improvement instructions. In the proposed model, K-means clustering is used to classify the safe driving level and the non-linear principal component analysis (NLPCA) model is trained by the classified low-risk data to analyze arbitrary driving data and provide feedback. To evaluate proposal model, we collected sensor data while driving vicinity of Daejeon Metropolitan City in Korea, and analyzed the principal component extracted using the NLPCA. We evaluated the classification accuracy of the principal components to verify the validity of the proposed model, and showed that the characteristics of safe driving can be represented through the proposal feedback model.

Sea surface temperature patterns in the Tropical Atlantic: Principal component analysis and nonlinear principal component analysis

Sea surface temperature patterns in the Tropical Atlantic: Principal component analysis and nonlinear principal component analysis

Interactive Pattern Discovery in High-Dimensional, Multimodal Data Using Manifolds

Data mining to discover patterns and aid decisions is the key to utilizing massive data for process automation and optimization. An especially challenging data mining problem is kriging, i.e., prediction of multiple, related variables from latent patterns in the data. We present a manifold based machine learning approach to discover patterns in massive, correlated, high-dimensional data. Dimensionality reduction using a manifold is a type of non-linear principal component analysis (PCA). The manifold captures the underlying data structure of the inputs and corresponding outputs by way of projecting the data onto a set of basis functions defined by the manifold. These bases ensure that any future adjustments affect the model with respect to the natural geometry of the data. We chose the manifold learning technique for its robustness against unbalanced data. Our contribution, described in this paper, enables interactive learning and incremental learning, i.e., incremental adjustment of the manifold (and its predictions) based on new observations and also user corrections to the predicted values, rerun the analysis on the full data set. Our experiments demonstrate that prediction performance remains equivalent to Multi-kernel Gaussian Processes on standard data sets despite these practically useful enhancements.

Nonlinear system monitoring using multiscaled principal components analysis based on neural network

In this paper, we propose a new method based on multiscaled principal component analysis for nonlinear systems analysis. We introduce nonlinear PCA based on neural networks and discrete wavelet transform. The data matrix describing a nonlinear process is decomposed into five wavelet resolution levels. The neural PCA is applied to each coefficient of details and approximations; we select only the scales having a defect to reconstruct the data matrix. Neural PCA is again applied to the new matrix to determine the defective variables, which are detected using the square predictive error (SPE) statistic and identified using the contributions calculation method. This method is applied to a biological process and shows efficient results.

Multivariate Analysis of MRI Biomarkers for Predicting Neurologic Impairment in Cervical Spinal Cord Injury.

Acute markers of spinal cord injury are essential for both diagnostic and prognostic purposes. The goal of this study was to assess the relationship between early MR imaging biomarkers after acute cervical spinal cord injury and to evaluate their predictive validity of neurologic impairment. We performed a retrospective cohort study of 95 patients with acute spinal cord injury and preoperative MR imaging within 24 hours of injury. The American Spinal Injury Association Impairment Scale was used as our primary outcome measure to define neurologic impairment. We assessed several MR imaging features of injury, including axial grade (Brain and Spinal Injury Center score), sagittal grade, length of injury, maximum canal compromise, and maximum spinal cord compression. Data-driven nonlinear principal component analysis was followed by correlation and optimal-scaled multiple variable regression to predict neurologic impairment. Nonlinear principal component analysis identified 2 clusters of MR imaging variables related to 1) measures of intrinsic cord signal abnormality and 2) measures of extrinsic cord compression. Neurologic impairment was best accounted for by MR imaging measures of intrinsic cord signal abnormality, with axial grade representing the most accurate predictor of short-term impairment, even when correcting for surgical decompression and degree of cord compression. This study demonstrates the utility of applying nonlinear principal component analysis for defining the relationship between MR imaging biomarkers in a complex clinical syndrome of cervical spinal cord injury. Of the assessed imaging biomarkers, the intrinsic measures of cord signal abnormality were most predictive of neurologic impairment in acute spinal cord injury, highlighting the value of axial T2 MR imaging.

Smart, open, user innovation and competitive advantage: a model for museums and heritage sites

ABSTRACTThe aim of this paper is to analyze the meanings and the contents of innovation and its role in museums and heritage sites’ management to create higher levels of competitiveness. Through a complex framework, we investigate the key role of innovation in determining the competitive advantage. We propose a theoretical model to study smart and open innovation in cultural services, with specific reference to museums and heritage sites, building up an ‘innovation indicator’, in order to measure the nature and the intensity of innovation as well as to identify how it can be leveraged to sustain a business-oriented approach and to get competitive advantage. This latter is conceived in strategic terms, using a resource-based theory approach: competitive advantage is therefore considered with respect to both economic and social performance, market and competitive positioning. We apply our index on a group of innovative selected cases of museums and heritage sites using nonlinear principal component analysis and hierarchical clustering. Empirical results show that smart innovation seems to be a necessary path to achieve competitive parity but it is not sufficient for competitive advantage.

Distinct Variation Pattern Discovery Using Alternating Nonlinear Principal Component Analysis.

Autoassociative neural networks (ANNs) have been proposed as a nonlinear extension of principal component analysis (PCA), which is commonly used to identify linear variation patterns in high-dimensional data. While principal component scores represent uncorrelated features, standard backpropagation methods for training ANNs provide no guarantee of producing distinct features, which is important for interpretability and for discovering the nature of the variation patterns in the data. Here, we present an alternating nonlinear PCA method, which encourages learning of distinct features in ANNs. A new measure motivated by the condition of orthogonal loadings in PCA is proposed for measuring the extent to which the nonlinear principal components represent distinct variation patterns. We demonstrate the effectiveness of our method using a simulated point cloud data set as well as a subset of the MNIST handwritten digits data. The results show that standard ANNs consistently mix the true variation sources in the low-dimensional representation learned by the model, whereas our alternating method produces solutions where the patterns are better separated in the low-dimensional space.

Voice assessments for detecting patients with Parkinson’s diseases using PCA and NPCA

In this study, we wanted to discriminate between two groups of people. The database used in this study contains 20 patients with Parkinson's disease and 20 healthy people. Three types of sustained vowels (/a/, /o/ and /u/) were recorded from each participant and then the analyses were done on these voice samples. Firstly, an initial feature vector extracted from time, frequency and cepstral domains. Then we used linear and nonlinear feature extraction techniques, principal component analysis (PCA), and nonlinear PCA. These techniques reduce the number of parameters and choose the most effective acoustic features used for classification. Support vector machine with its different kernel was used for classification. We obtained an accuracy up to 87.50 % for discrimination between PD patients and healthy people.

A framework for data compression and damage detection in structural health monitoring applied on a laboratory three-story structure

Structural Health Monitoring (SHM) is an important technique used to preserve many types of structures in the short and long run, using sensor networks to continuously gather the desired data. However, this causes a strong impact in the data size to be stored and processed. A common solution to this is using compression algorithms, where the level of data compression should be adequate enough to allow the correct damage identification. In this work, we use the data sets from a laboratory three-story structure to evaluate the performance of common compression algorithms which, then, are combined with damage detection algorithms used in SHM. We also analyze how the use of Independent Component Analysis, a common technique to reduce noise in raw data, can assist the detection performance. The results showed that Piecewise Linear Histogram combined with Nonlinear PCA have the best trade-off between compression and detection for small error thresholds while Adaptive PCA with Principal Component Analysis perform better with higher values.

Non-linear optimal multivariate spatial design using spatial vine copulas

A multivariate spatial sampling design that uses spatial vine copulas is presented that aims to simultaneously reduce the prediction uncertainty of multiple variables by selecting additional sampling locations based on the multivariate relationship between variables, the spatial configuration of existing locations and the values of the observations at those locations. Novel aspects of the methodology include the development of optimal designs that use spatial vine copulas to estimate prediction uncertainty and, additionally, use transformation methods for dimension reduction to model multivariate spatial dependence. Spatial vine copulas capture non-linear spatial dependence within variables, whilst a chained transformation that uses non-linear principal component analysis captures the non-linear multivariate dependence between variables. The proposed design methodology is applied to two environmental case studies. Performance of the proposed methodology is evaluated through partial redesigns of the original spatial designs. The first application is a soil contamination example that demonstrates the ability of the proposed methodology to address spatial non-linearity in the data. The second application is a forest biomass study that highlights the strength of the methodology in incorporating non-linear multivariate dependence into the design.

Internal defect inspection in magnetic tile by using acoustic resonance technology

This paper focuses on the validity of a nondestructive methodology for magnetic tile internal defect inspection based on acoustic resonance. The principle of this methodology is to analyze the acoustic signal collected from the collision of magnetic tile with a metal block. To accomplish the detection process, the separating part of the detection system is designed and discussed in detail in this paper. A simplified mathematical model is constructed to analyze the characteristics of the impact of magnetic tile with a metal block. The results demonstrate that calculating the power spectrum density (PSD) can diagnose the internal defect of magnetic tile. Two different data-driven multivariate algorithms are adopted to obtain the feature set, namely principal component analysis (PCA) and hierarchical nonlinear principal component analysis (h-NLPCA). Three different classifiers are then performed to deal with magnetic tile classification problem based on features extracted by PCA or h-NLPCA. The classifiers adopted in this paper are fuzzy neural networks (FNN), variable predictive model based class discrimination (VPMCD) method and support vector machine (SVM). Experimental results show that all six methods are successful in identifying the magnetic tile internal defect. In this paper, the effect of environmental noise is also considered, and the classification results show that all the methods have high immunity to background noise, especially PCA-SVM and h-NLPCA-SVM. Considering the accuracy rate, computation cost problem and the ease of implementation, PCA-SVM turns out to be the best method for this purpose.

A nonlinear principal component analysis to study archeometric data

Statistical techniques, when applied to data obtained by chemical investigations on ancient artworks, are usually expected to recognize groups of objects to classify the archeological finds, to attribute the provenance of items compared with earlier investigated ones, or to determine whether an archaelogical attribution is possible or not. The statistical technique most frequently used in archeometry is the principal component analysis (PCA), because of its simplicity in theory and implementation. However, the application of PCA to archeometric data showed severe limitations because of its linear feature. Indeed, PCA is inadequate to classify data whose behavior describe a curve or a curved subspace of the original data space. As a consequence of it, an amount of information is lost because the multi‐dimensional data space is compressed into a lower‐dimensional subspace including principal components. The aim of this work is then to test a novel statistical technique for archeometry. We propose a nonlinear PCA method to extract maximum chemical information by plotting data on the smallest number of principal components and to answer archeological questions. The higher accuracy and effectiveness of nonlinear PCA approach with respect to standard PCA for the analysis of archeometric data are shown through the study of Apulian red figured pottery (fifth–fourth century BC) coming from some of the most relevant archeological sites of ancient Apulia (Monte Sannace (Gioia del Colle), Egnatia (Fasano), Canosa, Altamura, Conversano, and Arpi(Foggia)). Copyright © 2016 John Wiley & Sons, Ltd.

An evaluation of regionalization and watershed classification schemes for continuous daily streamflow prediction in ungauged watersheds

Regionalization – the process of transferring hydrological information from gauged to ungauged watersheds – has the potential to perform significantly better if these watersheds are classified in advance. In this study, we demonstrate the benefits of classification by a systematic combination of watershed classification techniques, regionalization methods, and rainfall-runoff models. Basins are first classified, then regionalization methods are applied, for continuous daily streamflow estimation at ungauged watersheds in Ontario, Canada. Nonlinear data-driven methods are used as regionalization and watershed classification schemes to transfer the parameters of two conceptual hydrologic models – namely McMaster University Hydrologiska Byråns Vattenbalansavdelning (MAC-HBV) and Sacramento Soil Moisture Accounting (SAC-SMA) – from gauged to ungauged watersheds. Our results suggest that a certain combination of watershed classification technique, regionalization method and hydrologic model can significantly improve the estimation of continuous streamflow at ungauged basins by improving the accuracy of estimated daily mean, low and peak flows. However, some combinations do not provide a clear improvement when compared to the scenario of unclassified basins. For example, the MAC-HBV model coupled with a counter propagation neural network as a regionalization technique provides a clear improvement in estimated daily mean, low and peak flow when the watersheds are first classified using a nonlinear principal component analysis method. Interestingly, a higher improvement is achieved for low flow as well, which is usually difficult to estimate in ungauged basins.

Route effect on the perception of public transport services quality

User satisfaction sets a basic example for public service quality, especially those considered as basic necessities. The quality in transport service related to conceptualization and measurement, as a decisive demand key, presents challenges for both economic and mobility policies.Several operator companies are involved in the transport sector. Therefore this report aims to design a model of overall satisfaction based on the level of satisfaction with a specific set of factors, considering individual characteristics of users and the differential effect of different bus lines.The current presentation has applied a combined method, using Nonlinear Principal Component Analysis (NLPCA) and a Logit Multilevel Model (LMLM) in two-steps. DOI: http://dx.doi.org/10.4995/CIT2016.2016.3515

Examining the impact of leadership style and coaching on employees’ commitment and trust

PurposeThe purpose of this study is to analytically extend the understanding of leadership styles and organisational coaching and their influence on organisational outcomes and workplace counterproductive behaviours within a non-Western context, i.e. the United Arab Emirates (UAE).Design/methodology/approachIn this study, a sample size of 656 participants from 14 different business sectors in the Emirates was used. The meta-analysis concept of latent variables and non-linear principal components analysis, along with the corresponding methodology of structural equation modelling, were implemented.FindingsThe study finds that organisational coaching has a significant positive effect on transactional leadership and has a significant influence on job alienation. Interestingly, coaching has a significant effect on commitment and counterproductive workplace behaviours. The detailed data analysis usingFtests and independentt-tests, when applicable, indicated that there was a tendency for older employees to have more favourable attitudes towards transformational leadership or commitment but not towards coaching.Originality/valueDespite the popularity of the presented topic in today's organisations, research in a Middle East context has not kept pace with its counterpart in Western areas of the world. The present study attempts to bridge the gap between Western theories in developed countries and under-researched Eastern countries, namely, the UAE, and to test the impact of leadership styles and organisational coaching and their influence on employee commitment and trust mediated by job bullying and job alienation.

Hydrogeochemical analysis of multiple aquifers in a coal mine based on nonlinear PCA and GIS

Hydrogeochemical characteristic analysis of multiple categorical aquifers, including the delineation of hydrogeochemical spatial distributions, can provide a basic understanding of the groundwater flow system necessary for the prevention and control of groundwater damage associated with coal mines. Nonlinear principal component analysis (PCA) was used to analyse the groundwater chemical characteristics of water samples from five aquifers in Gubei Coal Mine, Anhui Province, China. In particular, Ca2+, Mg2+, Na+ + K+, HCO3−, Cl−, SO42−, and pH were analysed. To analyse the spatial distribution of groundwater chemistry in this Permian fractured aquifer, interpolation of principal component 1 (PC1) scores obtained from the nonlinear PCA was performed using the inverse distance weighting (IDW) method. Additionally, concentrations of the hydrogeochemical variables were plotted against depth. The relationships between water samples, variables, and water samples with variables were simultaneously analysed by nonlinear PCA. The results show: (1) nonlinear PCA explains more variability than standard PCA in the first two principal components (PCs); (2) the Cenozoic top aquifer has low concentrations of Na+ + K+ and Cl−; the Cenozoic middle aquifer and Cenozoic bottom aquifer are characterised by high concentrations of Ca2+ and Mg2+ and low concentrations of HCO3−; the shallow groundwater of the Permian fractured aquifer is recharged by the Cenozoic bottom aquifer, while other parts have high concentrations of HCO3− and low concentrations of Ca2+ and Mg2+. The Taiyuan limestone aquifer has similar groundwater chemistry to that of the Cenozoic bottom aquifer, and thus it may be recharged by the Cenozoic bottom aquifer in outcropping Taiyuan Formation strata; (3) the wide south-western and north-western areas of the Permian fractured aquifer may be recharged by the Cenozoic bottom aquifer; and (4) concentrations of Ca2+ and Mg2+ in the Permian fractured aquifer are negatively correlated with sampling depth, while the concentrations of Na+ + K+ and HCO3− are positively correlated with sampling depth, which verifies that the shallow groundwater of this Permian fractured aquifer is recharged by the Cenozoic bottom aquifer. The findings demonstrate that combining nonlinear PCA with GIS can be a powerful approach for conducting spatial nonlinear analysis of hydrogeochemical characteristics.

Geodesic Flow Kernel Support Vector Machine for Hyperspectral Image Classification by Unsupervised Subspace Feature Transfer

In order to deal with scenarios where the training data, used to deduce a model, and the validation data have different statistical distributions, we study the problem of transformed subspace feature transfer for domain adaptation (DA) in the context of hyperspectral image classification via a geodesic Gaussian flow kernel based support vector machine (GFKSVM). To show the superior performance of the proposed approach, conventional support vector machines (SVMs) and state-of-the-art DA algorithms, including information-theoretical learning of discriminative cluster for domain adaptation (ITLDC), joint distribution adaptation (JDA), and joint transfer matching (JTM), are also considered. Additionally, unsupervised linear and nonlinear subspace feature transfer techniques including principal component analysis (PCA), randomized nonlinear principal component analysis (rPCA), factor analysis (FA) and non-negative matrix factorization (NNMF) are investigated and compared. Experiments on two real hyperspectral images show the cross-image classification performances of the GFKSVM, confirming its effectiveness and suitability when applied to hyperspectral images.

Nonlinear Principal Component Analysis Approach to Pattern Recognition

Nonlinear Principal Component Analysis Approach to Pattern Recognition