A linearly convergent algorithm for distributed principal component analysis

Gang, Arpita; Bajwa, Waheed U.

doi:10.1016/j.sigpro.2021.108408

Cited by 19 publications

(7 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Principal Component Analysis (PCA) is a statistical analysis that reduces the dimensionality of a data set by linear transformation and performs a comprehensive analysis of the sample ( 44 ). Correlation analysis is a straightforward analytical technique for determining the relationship between two sets of quantitative data.…”

Section: Resultsmentioning

confidence: 99%

Physical properties, chemical composition, and nutritional evaluation of common salad dressings

et al. 2022

View full text Add to dashboard Cite

Salad dressings (SDs), a subcategory of flavored sauces with more than 20% fat content and less than 30% moisture content, is favored by consumers due to its delicacy. The physical properties, chemical composition and nutritional evaluation of common SDs on the Chinese market needs to be systematically investigated. This study compared the quality (physical properties, proximate composition, amino acids, fatty acids, and minerals) of six commercially available sweet SDs (SD1, SD2, SD3, SD4, SD5, and SD6) from the Chinese market. The results indicated that the water activity of six SDs was less than 0.60 (0.35–0.41), the moisture content was less than 65% (24.0–60.0%), and the crude fat content was between 23.25 and 64.15%. The essential amino acid index (EAAI) of SD3, SD1, SD2, and SD4 was greater than the FAO/WHO standard (EAAI = 75). Numerous fatty acids were detected, mainly linoleic and oleic acids; n-3 polyunsaturated fatty acids were ranged from 1,090 mg/100 g to 2,520 mg/100 g. In addition, SDs were rich in minerals such as Mg, Ca, and Fe and the atherogenic index and thrombogenic index were 0.03–0.09 and 0.77–0.91, respectively. In summary, this work helps to provide key nutritional information on the composition of common SDs. The availability of this data may help purchasers with different nutritional needs to make informed choices about the use of SD and encourage more moderate consumption of pre-packaged sauces.

show abstract

Section: Resultsmentioning

confidence: 99%

Physical properties, chemical composition, and nutritional evaluation of common salad dressings

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The work in this paper is an extension of the preliminary work in [1] that proposed two fast and efficient algorithms for distributed PCA in the case of sample-wise distributed data. A distributed algorithm for PCA based on generalized Hebbian algorithm was developed and analyzed in our previous work [30]. Though it is a linearly convergent one-time scale algorithm that converges with any random initial point, it only reached to a neighborhood of the optimal solution.…”

Section: A Relation To Prior Workmentioning

confidence: 99%

“…Estimation of the eigenvectors of C at every node without sharing raw data C i would require some form of collaboration among the nodes of the network. To this end, we proposed a combine-and-adapt strategy in our earlier work [30] that is used widely in distributed optimization literature for convex and strongly convex problems. Using that approach and an extensive analysis, we could show that even for the non-convex PCA problem, each node converges linearly and globally, i.e., starting from any random initial point, but only to a neighborhood of the true eigenvectors of the global covariance matrix C. Even though we used the generalized Hebbian algorithm, some straightforward calculations and manipulations can show similar results for Krasulina's method.…”

Section: Proposed Algorithm: Fast-pcamentioning

confidence: 99%

FAST-PCA: A Fast and Exact Algorithm for Distributed Principal Component Analysis

Gang¹,

Bajwa

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Principal Component Analysis (PCA) is a fundamental data preprocessing tool in the world of machine learning.While PCA is often reduced to dimension reduction, the purpose of PCA is actually two-fold: dimension reduction and feature learning. Furthermore, the enormity of the dimensions and sample size in the modern day datasets have rendered the centralized PCA solutions unusable. In that vein, this paper reconsiders the problem of PCA when data samples are distributed across nodes in an arbitrarily connected network. While a few solutions for distributed PCA exist those either overlook the feature learning part of the purpose, have communication overhead making them inefficient and/or lack exact convergence guarantees. To combat these aforementioned issues, this paper proposes a distributed PCA algorithm called FAST-PCA (Fast and exAct diSTributed PCA). The proposed algorithm is efficient in terms of communication and can be proved to converge linearly and exactly to the principal components that lead to dimension reduction as well as uncorrelated features. Our claims are further supported by experimental results.

show abstract

“…In local descriptors, the methodology is deployed on small image patches to develop its feature size. Extracting features from different image regions make local descriptors much effective than global descriptors [4][5][6]. In global descriptors, methodology is deployed on the entire image.…”

Section: Introductionmentioning

confidence: 99%

Radial Orthogonal Median LBP (ROM-LBP): A Discriminant Local Descriptor in Light Variations for Face Recognition

Karanwal

Diwakar

2023

JGEU

View full text Add to dashboard Cite

LBP and majority of its variants performs extremely well in front of moderate light variations. But when light variations becomes severe then performance of LBP and its variants is not satisfactory. Therefore there is a need of the more promising and impressive descriptor which performs well in harsh light variations. To complement these LBP based descriptors the proposed work launches the novel descriptor for Face Recognition (FR) in harsh lightning variations. This proposed descriptor is called as Radial Orthogonal Median LBP (ROM-LBP). The main demerit of these LBP based descriptors is that they all consider the uniform coordination between the neighbors and center pixel. Which mean raw pixel intensity is used for the comparison with the center pixel. The proposed work eliminates this problem in the introduced descriptor ROM-LBP, by replacing the raw pixels intensity with the median of the radial points in each orthogonal position of the two separate groups. The generated median is then used for comparison with the center pixel. The respective codes obtained from both the groups are concatenated to form the ROM-LBP size. As region feature extraction is done therefore ROM-LBP develops the large feature size. To make more effective descriptor, the services of FLDA is used and then classification was conducted by SVMs. Experiments conducted on EYB and YB datasets demonstrates the ability of the proposed ROM-LBP against various LBP and non-LBP based descriptors.

show abstract

A linearly convergent algorithm for distributed principal component analysis

Cited by 19 publications

References 24 publications

Physical properties, chemical composition, and nutritional evaluation of common salad dressings

Physical properties, chemical composition, and nutritional evaluation of common salad dressings

FAST-PCA: A Fast and Exact Algorithm for Distributed Principal Component Analysis

Radial Orthogonal Median LBP (ROM-LBP): A Discriminant Local Descriptor in Light Variations for Face Recognition

Contact Info

Product

Resources

About