Canonical Correlation Analysis Technique Research Articles

Local Discriminant Hyperalignment for Multi-Subject fMRI Data Alignment

Multivariate Pattern (MVP) classification can map different cognitive states to the brain tasks. One of the main challenges in MVP analysis is validating the generated results across subjects. However, analyzing multi-subject fMRI data requires accurate functional alignments between neuronal activities of different subjects, which can rapidly increase the performance and robustness of the final results. Hyperalignment (HA) is one of the most effective functional alignment methods, which can be mathematically formulated by the Canonical Correlation Analysis (CCA) methods. Since HA mostly uses the unsupervised CCA techniques, its solution may not be optimized for MVP analysis. By incorporating the idea of Local Discriminant Analysis (LDA) into CCA, this paper proposes Local Discriminant Hyperalignment (LDHA) as a novel supervised HA method, which can provide better functional alignment for MVP analysis. Indeed, the locality is defined based on the stimuli categories in the train-set, where the correlation between all stimuli in the same category will be maximized and the correlation between distinct categories of stimuli approaches to near zero. Experimental studies on multi-subject MVP analysis confirm that the LDHA method achieves superior performance to other state-of-the-art HA algorithms.

Motion artifact removal from single channel electroencephalogram signals using singular spectrum analysis

Abstract In ambulatory electroencephalogram (EEG) health care systems, recorded EEG signals often contaminated by motion artifacts. In this paper, we proposed a singular spectrum analysis (SSA) technique with new grouping criteria to remove the motion artifact from a single channel EEG signal. In order to remove the motion artifact from a single channel EEG signal, we considered the eigenvectors (basis vectors) corresponding to motion artifact are grouped or identified based on their local mobility, which is a signal complexity measure. However, as the local mobility of eigenvectors associated to the motion artifact are small, a threshold of 0.1 is set to identify them. The motion artifact signal is estimated using the identified eigenvectors and subtracted from the contaminated EEG signal to obtain the corrected EEG signal. The proposed technique is tested on 21 single channel real EEG signals contaminated by motion artifact and compared the results with the existing combined ensemble empirical mode decomposition and canonical correlation analysis (EEMD-CCA) technique. The simulation results show that the proposed modified SSA enjoys an improvement in the signal to noise ratio and the percentage reduction in artifact. Moreover, as the ambulatory EEG systems are battery operated, use of high computational signal processing techniques will reduce the battery lifetime. Hence, low computational signal processing techniques are greatly demanded in such applications. Thus, we have also evaluated the computational complexity of the proposed technique and compared with EEMD-CCA. We found that the proposed modified SSA technique significantly reduces the computational complexity and thereby lower power consumption compared to the EEMD-CCA.

MetaCCA: summary statistics-based multivariate meta-analysis of genome-wide association studies using canonical correlation analysis.

Motivation: A dominant approach to genetic association studies is to perform univariate tests between genotype-phenotype pairs. However, analyzing related traits together increases statistical power, and certain complex associations become detectable only when several variants are tested jointly. Currently, modest sample sizes of individual cohorts, and restricted availability of individual-level genotype-phenotype data across the cohorts limit conducting multivariate tests.Results: We introduce metaCCA, a computational framework for summary statistics-based analysis of a single or multiple studies that allows multivariate representation of both genotype and phenotype. It extends the statistical technique of canonical correlation analysis to the setting where original individual-level records are not available, and employs a covariance shrinkage algorithm to achieve robustness.Multivariate meta-analysis of two Finnish studies of nuclear magnetic resonance metabolomics by metaCCA, using standard univariate output from the program SNPTEST, shows an excellent agreement with the pooled individual-level analysis of original data. Motivated by strong multivariate signals in the lipid genes tested, we envision that multivariate association testing using metaCCA has a great potential to provide novel insights from already published summary statistics from high-throughput phenotyping technologies.Availability and implementation: Code is available at https://github.com/aalto-ics-kepacoContacts: anna.cichonska@helsinki.fi or matti.pirinen@helsinki.fiSupplementary information: Supplementary data are available at Bioinformatics online.

Canonical correlation analysis-based fault detection methods with application to alumina evaporation process

In this paper, canonical correlation analysis (CCA)-based fault detection methods are proposed for both static and dynamic processes. Different from the well-established process monitoring and fault diagnosis systems based on multivariate analysis techniques like principal component analysis and partial least squares, the core of the proposed methods is to build residual signals by means of the CCA technique for the fault detection purpose. The proposed methods are applied to an alumina evaporation process, and the achieved results show that both methods are applicable for fault detection, while the dynamic one delivers better detection performance.

Sparse canonical correlation analysis from a predictive point of view.

Canonical correlation analysis (CCA) describes the associations between two sets of variables by maximizing the correlation between linear combinations of the variables in each dataset. However, in high-dimensional settings where the number of variables exceeds the sample size or when the variables are highly correlated, traditional CCA is no longer appropriate. This paper proposes a method for sparse CCA. Sparse estimation produces linear combinations of only a subset of variables from each dataset, thereby increasing the interpretability of the canonical variates. We consider the CCA problem from a predictive point of view and recast it into a regression framework. By combining an alternating regression approach together with a lasso penalty, we induce sparsity in the canonical vectors. We compare the performance with other sparse CCA techniques in different simulation settings and illustrate its usefulness on a genomic dataset.

Quantitative Monthly Precipitation Forecasting Using Cyclostationary Empirical Orthogonal Function and Canonical Correlation Analysis

AbstractAn empirical statistical system for quantitative forecasting of monthly precipitation in Korea has been developed using the cyclostationary empirical orthogonal function (CSEOF) and the canonical correlation analysis (CCA) with sea surface temperature (SST) data as the predictor. Monthly Korean precipitation and SST data are comprehensively analyzed using the empirical orthogonal function (EOF) technique and the CSEOF technique, respectively, and the CSEOF technique can exhibit the spatial distribution and temporal evolution characteristics of variability along with recurrent seasons of precipitation in Korea. Through a multivariate regression method, the CCA technique is used to forecast precipitation with different lead times, and the forecasting results indicate that the CSEOF-CCA forecasting model agrees well with the observation data and is particularly useful in forecasting seasonal precipitation variations in Korea.

基于半监督学习的黄萎病互作基因对的预测

黄萎病(Greensickness)属于不可治愈性病害，每年均造成巨大的经济损失。为了培育抗病植株、得到更多的互作基因对，逐次进行生物实验排除是不现实的。为了在已知少量关联基因的情况下挖掘更多的可靠基因对，本文主要使用统计技术典型相关分析(Canonical Correlation Analysis, CCA)和数据挖掘技术半监督学习(Semi-Supervised Learning, SSL)等生物信息技术对相关基因进行学习，最终实现对关联基因的预测。研究结果能够有效地指导黄萎病抗病研究的方向、精确研究范围、提高研究速度。 Greensickness is an incurable disease, which can cause huge economic losses every year. Using a series of biological experiments to develop resistant plants and get more interaction gene is un-realistic. In order to dig more reliable genes under the premise of only having very few interaction gene, the paper mainly uses bio-IT, such as statistical techniques of Canonical Correlation Analysis and data mining techniques of Semi-Supervised Learning to study the related gene and gets the prediction of the interaction gene at the end. The result of the study can effectively guide the direction of the greensickness research, narrow the scope of the research and improve the research speed.

Secondary analysis of two environmental practice studies. Do empirical variables represent expressed theoretical constructs?

Abstract Many research questions in studies on environmental practices have used variables with little regard as to whether they represent theoretical constructs. This validation is fundamental to testing and interpreting theories using variables. This study examines whether environmental disclosure, environmental performance, and economic performance are different constructs. It uses canonical correlation analysis technique, and reviews two published studies as case presentations. Analyzing the study by Al-Tuwaijri et al., the environmental disclosure variable is found to represent the disclosure construct, while the environmental performance and economic performance variables are found to represent two dimensions of the performance construct. In the Hasseldine et al. study, environmental performance is found to represent the performance construct. Results indicate that the two disclosure variables are uni-dimensional of the disclosure construct. These findings are supported by the main analysis and additional analysis undertaken for each study. This research demonstrates how the canonical correlation analysis technique can be used to test whether empirical variables represent expressed theoretical constructs before publishing findings. It also shows a unique way to review published studies to develop and analyze research questions not examined in the original work.

An Application of Fuzzy Canonical Correlation and Fuzzy DEA for Ranking Bank Branches

Performance evaluation and efficiency analysis of economic units are of great importance. Measuring the efficiency of the banking industry has been one of the most interesting areas of research for the past few years. There are literally various techniques for measuring the relative performance of similar units such as banks including Data Envelopment Analysis. Data Envelopment Analysis method is a fact based mathematical programming which is used to measure and analyze the efficiency of decision making units. In addition, the canonical correlation analysis technique is one of the multivariate statistical methods to analyze and rank units. However, the observed values of the input and output data in real- world problems are sometimes imprecise or vague. Many researchers have proposed various fuzzy methods for dealing with the imprecise and Original Research Article

Modeling the global NmF2 from the GNSS‐derived TEC‐GIMs

Based on canonical correlation analysis (CCA), we propose a technique to map the peak electron density of the ionospheric F2‐layer (NmF2) from the global observations of the total electron content (TEC). We first statistically analyze the canonical correlation between NmF2 and TEC using the NmF2 observations of the worldwide ionosonde stations and global ionospheric maps (GIMs) of TEC (TEC‐GIMs) provided by the Jet Propulsion Laboratory. The obtained CCA modes consist of the CCA patterns and corresponding amplitudes, which separately reflect the short‐term (e.g., diurnal variation) and long‐term (e.g., annual, semiannual variations, and solar‐cycle variations, etc.) tendencies of NmF2. With the obtained CCA patterns and corresponding amplitudes, we then construct two TEC‐derived models of the monthly median NmF2, a single station model at Boulder (105.3°W, 40°N) and a global model. The CCA results are further compared with the observations and corresponding NmF2 predictions of the International Reference Ionosphere 2007 (IRI‐07). It is found that the CCA NmF2 is in better agreement with the observations at Boulder than the IRI‐07 NmF2. On the other hand, the temporal and spatial structures of NmF2 are successfully reproduced by the global model. Furthermore, the correlation coefficient (root mean square error) of the CCA NmF2 versus the observed NmF2 is relatively higher (lower) than that of IRI‐07 NmF2. The CCA NmF2 from the global model is more reliable. Thus, we conclude that the CCA technique provides an effective approach to map the global NmF2 from the TEC‐GIMs.

Using the Canonical Correlation Analysis Technique for Imaging Dimensionality Reduction in Multisource Land sat Images

The Canonical Correlations Analysis technique (CCA) was suggested in the dimensionality reduction images for the multivariate multisource data applied in remote sensing . These techniques transform multivariate multiset data into new orthogonal variables called Canonical Variates (CVs) . This research uses the LANDSAT-5 TM data for the set of multivariate multispectral correlation at fixed points in time . The results show maximum similarity for the low- order canonical variates and minimum similarity for the high- order canonical variates .

Using the Canonical Correlation Analysis Technique for Imaging Dimensionality Reduction in Multisource Land sat Images

Using the Canonical Correlation Analysis Technique for Imaging Dimensionality Reduction in Multisource Land sat Images

Integrating Principle Component Analysis and Canonical Correlation Analysis for Monitoring Water Quality in Storage Reservoir

Water quality monitoring is of great importance in managing drinking water lakes or reservoirs. However, developing such an effective monitoring system used in evaluating water quality is difficult, due to spatial and temporal variation in water quality that are usually hard to interpret. Principle component analysis (PCA) and canonical correlation analysis (CCA) techniques were used in this study to determine the principle water parameters, and explore the relationship between the physical and chemical water parameters, respectively in Macau Main Storage Reservoir (MSR) that is experiencing algal bloom in recent years. Twenty-eight water parameters including physical and chemical parameters were monitored each month from Jan 2001 to Dec 2010. The results showed that using PCA, six of the water quality parameters were found less important in explaining the monthly variation of the water quality and thus excluded from the further CCA, while applying CCA, the first six canonical correlations were 0.998, 0.988, 0.937, 0.778, 0.606 and 0.453, indicating that electro-conductivity and chloride compounds were the dominant (highly scored) variables of the physical parameters and chemical parameters, respectively. These results are helpful in understanding the main physical and chemical parameters involved in MSR, thus improving the water quality monitoring system in drinking water reservoirs.

Sparse Canonical Correlation Analysis from a Predictive Point of View

Canonical correlation analysis (CCA) describes the associations between two sets of variables by maximizing the correlation between linear combinations of the variables in each data set. However, in high-dimensional settings where the number of variables exceeds the sample size or when the variables are highly correlated, traditional CCA is no longer appropriate. This paper proposes a method for sparse CCA. Sparse estimation produces linear combinations of only a subset of variables from each data set, thereby increasing the interpretability of the canonical variates. We consider the CCA problem from a predictive point of view and recast it into a multivariate regression framework. By combining a multi-variate alternating regression approach together with a lasso penalty, we induce sparsity in the canonical vectors. We compare the performance with other sparse CCA techniques in different simulation settings and illustrate its usefulness on a genomic data set.

Linkages between global sea surface temperatures and decadal rainfall variability over Eastern Africa region

AbstractLinkages between dominant spatio‐temporal decadal rainfall variability modes and the global sea surface temperature (SST) modes are investigated over East Africa region for the period 1950–2008. Singular value decomposition (SVD) and canonical correlation analysis (CCA) techniques are employed to examine potential linkages and predictability of decadal rainfall variability over the region. When the ten‐year periodicity is filtered out from the observed monthly rainfall data, distinct decadal rainfall regimes are exhibited in the time series of mean seasonal rainfall anomalies. Spectral density analysis of rainfall time series showed dominance of a ten‐year periodicity, significant at 95% confidence level. The Principal Component Analysis (PCA) results yielded nine and seven homogeneous decadal rainfall zones for long rains; March–May (MAM), and the short rains: October–December (OND) seasons, respectively. The third season of June–August (JJA) which is mainly experienced in western and coastal sub‐regions had eight homogenous zones delineated. Results show that the leading three SVD‐coupled modes explain greater than 75% of the squared covariance between the two fields. The first SVD mode for Indian, Atlantic and Pacific Oceans contributed to 50, 43 and 38% of the total square covariance for MAM season, respectively. The same mode accounted for 65, 48 and 40% for OND rainfall season, respectively. For the JJA season, mode one contributed to about 61, 39 and 42% of the variance. The study showed that forcing of decadal rainfall over the region is associated with El Niño mode that is prominent over the Pacific Ocean, while Indian Ocean dipole is the leading mode over the Indian Ocean basin. An inter‐hemispheric dipole mode that is common during ENSO was a prominent feature in the Atlantic Ocean forcing regional decadal rainfall. The high variability of these modes highlighted the significant roles of all the global oceans in forcing decadal rainfall variability over the region. In addition, results from multiple linear regression model showed substantial variation of the model prediction skill of the decadal rainfall variability modes within various homogenous zones and for different seasons. Copyright © 2012 Royal Meteorological Society

A Least-Squares Framework for Component Analysis

Over the last century, Component Analysis (CA) methods such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Canonical Correlation Analysis (CCA), Locality Preserving Projections (LPP), and Spectral Clustering (SC) have been extensively used as a feature extraction step for modeling, classification, visualization, and clustering. CA techniques are appealing because many can be formulated as eigen-problems, offering great potential for learning linear and nonlinear representations of data in closed-form. However, the eigen-formulation often conceals important analytic and computational drawbacks of CA techniques, such as solving generalized eigen-problems with rank deficient matrices (e.g., small sample size problem), lacking intuitive interpretation of normalization factors, and understanding commonalities and differences between CA methods. This paper proposes a unified least-squares framework to formulate many CA methods. We show how PCA, LDA, CCA, LPP, SC, and its kernel and regularized extensions correspond to a particular instance of least-squares weighted kernel reduced rank regression (LS--WKRRR). The LS-WKRRR formulation of CA methods has several benefits: 1) provides a clean connection between many CA techniques and an intuitive framework to understand normalization factors; 2) yields efficient numerical schemes to solve CA techniques; 3) overcomes the small sample size problem; 4) provides a framework to easily extend CA methods. We derive weighted generalizations of PCA, LDA, SC, and CCA, and several new CA techniques.

Comparing Canonical Correlation Analysis with Partial Least Squares Regression in Estimating Forest Leaf Area Index with Multitemporal Landsat TM Imagery

The leaf area index (LAI) of plant canopies is an important structural variable for assessing terrestrial ecosystems. This research examined the use of multitemporal Landsat TM imagery to estimate and map LAI in mixed natural forests in the southeastern USA. The performances of canonical correlation analysis (CCA) and partial least squares (PLS) regression techniques were evaluated for feature extraction to estimate forest LAI. The experimental results indicate that use of multitemporal TM imagery can improve the accuracy of estimating the forest LAI, and that CCA analysis outperforms PLS regression for feature extraction.

Fourier Transform Infrared and Raman Spectroscopies for the Rapid Detection, Enumeration, and Growth Interaction of the Bacteria Staphylococcus aureus and Lactococcus lactis ssp. cremoris in Milk

Staphylococcus aureus is one of the main pathogenic microorganisms found in milk and dairy products and has been involved in bacterial foodborne outbreaks in the past. Current enumeration techniques for bacteria are very time-consuming, typically taking 24 h or longer, and bacterial antagonism in the form of lactic acid bacteria (LAB) may inhibit the growth of S. aureus . Therefore, the aim of this investigation was to establish the accuracy and sensitivity of rapid nondestructive metabolic fingerprinting techniques, such as Fourier transform infrared (FT-IR) spectroscopy and Raman spectroscopy (RS), in combination with multivariate analysis techniques, for the detection and enumeration of S. aureus in milk, as well as to study the growth interaction between S. aureus and Lactococcus lactis ssp. cremoris , a common LAB. The two bacterial species were investigated both in a pure monoculture and in a combined inoculated coculture after inoculation into ultraheated milk during the first 24 h of growth at 37 °C. Plating techniques were used to obtain primary reference data for viable bacteria counts. Principal component discriminant function analysis, canonical correlation analysis, partial least-squares (PLS), and kernel PLS (KPLS) multivariate statistical techniques were employed to analyze the data. FT-IR provided very reasonable quantification results both with PLS and KPLS, the latter providing marginally better predictions, with correlation coefficients in the test set (Q(2)) and training set (R(2)) varying from 0.64 to 0.76 and from 0.78 to 0.88 for different bacterial sample combinations. RS results were less encouraging with high degrees of error and poor correlation to viable bacterial counts. S. aureus growth was not inhibited by the presence of the LAB, but metabolic fingerprinting of the coculture indicated that the phenotype of this dual bacterial culture was closer to that of pure LAB cultures. In conclusion, FT-IR spectroscopy in combination with the above multivariate techniques appears to be a promising discrimination and enumeration analytical technique for the two bacterial species. In addition, it has been demonstrated that the L. cremoris metabolic effect in milk dominates that of S. aureus even though there was no growth antagonism observed.

ApproxCCA: An approximate correlation analysis algorithm for multidimensional data streams

Correlation analysis is regarded as a significant challenge in the mining of multidimensional data streams. Great emphasis is generally placed on one-dimensional data streams with the existing correlation analysis methods for the mining of data streams. Therefore, the identification of underlying correlation among multivariate arrays (e.g. Sensor data) has long been ignored. The technique of canonical correlation analysis (CCA) has rarely been applied in multidimensional data streams. In this study, a novel correlation analysis algorithm based on CCA, called ApproxCCA, is proposed to explore the correlations between two multidimensional data streams in the environment with limited resources. By introducing techniques of unequal probability sampling and low-rank approximation to reduce the dimensionality of the product matrix composed by the sample covariance matrix and sample variance matrix, ApproxCCA successfully improves computational efficiency while ensuring the analytical precision. Experimental results of synthetic and real data sets have indicated that the computational bottleneck of traditional CCA can be overcome with ApproxCCA, and the correlations between two multidimensional data streams can also be detected accurately.

IMAGE QUALITY FEEDBACK-BASED ADAPTIVE VIDEO DEFINITION IMPROVEMENT FOR THE SPACE MANIPULATION TASK

A novel video definition enhancement technique for space manipulation task is presented in this paper. In contrast to other techniques, we implement our algorithm by considering both the feedback of the image quality (IQ) and the correction of the visual fatigue factor simultaneously. First, when designing the objective IQ evaluation metrics, we employ both the subjective evaluation result of the IQ and the information of the attention region to improve their performances. The canonical correlation analysis technique is used to build the connection between the subjective evaluation results and the objective ones. The eye tracking system is employed to search the location of the visual attention. Second, a denoising and enhancement technique in wavelet domain is developed to improve the imaging definition. The whole calculation process is monitored by the feedback of the objective IQ evaluation metrics mentioned above. The evaluation result of IQ is also corrected by some visual fatigue factors since the space tasks are always a long-time mission. The face recognition technique will be included in our system in the future to detect and analyze the visual fatigue, although it is not developed currently. Extensive experimental results have shown the effectivity of our proposed technique.