Structure Of Original Data Research Articles

Non-negative consistency affinity graph learning for unsupervised feature selection and clustering

Feature selection plays a crucial role in data mining and pattern recognition tasks. This paper proposes an efficient method for robust unsupervised feature selection, called joint local preserving and low-rank representation with nonnegative and symmetric constraint (JLPLRNS). This approach utilizes an indicator matrix instead of the row sparsity of the projection matrix to directly select some significant features from the original data and adaptively preserves the local geometric structure of original data into the low-dimensional embedded feature subspace via learned indicator matrix. By simultaneously imposing nonnegative symmetric and low-rank constraints on the representation coefficient matrix, it cannot only make this matrix discriminative, sparse and weight consistency for each pair of data, but also uncover the global structure of original data. These effectively will improve clustering performance. An algorithm based on the augmented Lagrange multiplier method with an alternating direction strategy is designed to resolve this model. Experimental results on various real datasets show that the proposed method can effectively identify some important features in data and outperforms many state-of-the-art unsupervised feature selection methods in terms of clustering performance.

Optical fiber sensing and tensor AP clustering for high-speed road tunnel vehicle detection

Aiming at the problems of signal acquisition, feature extraction and vehicle recognition in highway tunnel vehicle detection, a new tunnel vehicle detection method is proposed by combining optical time-domain reflectometry distributed fiber sensing technology and tensor affine propagation clustering algorithm. Firstly, the distributed optical fiber system designed by optical time domain reflectometry was used to collect the running signals of tunnel vehicles and obtain the measurement data. Secondly, a high-order tensor sample set is constructed by using the spatial resolution of optical fiber as channel number and combining the feature number, time domain and frequency domain. Finally, tensor affine propagation clustering method and other clustering methods are used to test the accuracy. The test results show that the proposed method can better classify vehicles without destroying the original high-dimensional data structure. Meanwhile, the unsupervised clustering algorithm also reduces manual intervention in the identification process, increases the intelligence level of the whole vehicle detection model, and effectively improves the detection accuracy rate of tunnel vehicles.

Abstract 6126: Understanding disparities in lung cancer using single cell RNA sequencing data transformed by the Gerchberg Saxton algorithm

Abstract Purpose An overarching goal of this study is to reduce disparities in lung cancer research commonly manifested by racial and/or gender under-representations. By employing the Gerchberg-Saxton (GS) algorithm, we aim to diminish data noise/bias in single-cell RNA sequencing (scRNA-seq) gene expression data, thereby enabling more equitable research outcomes. An immediate goal is to remove extensive data bias associated with scRNA-seq data thus enhancing downstream Machine Learning analyses. Background: We have successfully applied the GS algorithm for fairer mortality rate predictions across different racial groups [1]. In this study, we seek to apply the GS algorithm to the scRNA-seq data that interrogate the cellular landscapes of lung cancer and the tumor microenvironment. Methodology: The application of the GS algorithm to single cell RNA data involved a series of steps. Initially, all data frames were transposed, switching columns and rows to represent gene expressions and single cells, respectively. The algorithm was then meticulously applied to each column, allowing for the uniform distribution of certain specific gene expression information across all single cells within the data frame. This column-wise application was crucial to maintain the integrity of individual cell characteristics while ensuring a fair structural distribution of gene expression data. The process aimed to balance the representation of genes across all cells, addressing inherent biases and noises in the original data structure. Data: The study utilizes a previously unpublished scRNA-seq dataset comprising 14 lung cancer patients from Wake Forest Baptist Comprehensive Cancer Center, including 6 African American and 8 Caucasian patients. Results: One out of 14 preliminary clustering results are presented in Figure 2 using ScanPy [2]. The initial findings are based on a restricted set of cell marker genes, which will be further developed with additional markers in subsequent analyses. With the application of the Gerchberg-Saxton algorithm, the Shannon Entropy [3] analysis revealed a more uniform randomness across gene expressions (Figure 1), and the unsupervised clustering indicated a clearer separation of cell types, significantly enhancing the ability for downstream analyses. The data transformation will be applied to the whole dataset for comparative analyses of cellular landscapes between lung cancer from different races and genders, to be reported at the AACR 2024 Annual Meeting. Conclusions: Our preliminary studies showed that the Gerchberg-Saxton algorithm is effectiveness in normalizing data distribution for scRNA-seq data, which has led to enhanced resolution of cell type differentiations in clustering analysis. With this refined methodology, we are better poised to better address lung cancer health disparities revealed by single cell sequencing analysis. Citation Format: Seha Ay, Liang Liu, Elizabeth Forbes, Umit Topaloglu, Wei Zhang. Understanding disparities in lung cancer using single cell RNA sequencing data transformed by the Gerchberg Saxton algorithm [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6126.

Bayesian variable selection for matrix autoregressive models

A Bayesian method is proposed for variable selection in high-dimensional matrix autoregressive models which reflects and exploits the original matrix structure of data to (a) reduce dimensionality and (b) foster interpretability of multidimensional relationship structures. A compact form of the model is derived which facilitates the estimation procedure and two computational methods for the estimation are proposed: a Markov chain Monte Carlo algorithm and a scalable Bayesian EM algorithm. Being based on the spike-and-slab framework for fast posterior mode identification, the latter enables Bayesian data analysis of matrix-valued time series at large scales. The theoretical properties, comparative performance, and computational efficiency of the proposed model is investigated through simulated examples and an application to a panel of country economic indicators.

A structural VAR and VECM modeling method for open-high-low-close data contained in candlestick chart

The structural modeling of open-high-low-close (OHLC) data contained within the candlestick chart is crucial to financial practice. However, the inherent constraints in OHLC data pose immense challenges to its structural modeling. Models that fail to process these constraints may yield results deviating from those of the original OHLC data structure. To address this issue, a novel unconstrained transformation method, along with its explicit inverse transformation, is proposed to properly handle the inherent constraints of OHLC data. A flexible and effective framework for structurally modeling OHLC data is designed, and the detailed procedure for modeling OHLC data through the vector autoregression and vector error correction model are provided as an example of multivariate time-series analysis. Extensive simulations and three authentic financial datasets from the Kweichow Moutai, CSI 100 index, and 50 ETF of the Chinese stock market demonstrate the effectiveness and stability of the proposed modeling approach. The modeling results of support vector regression provide further evidence that the proposed unconstrained transformation not only ensures structural forecasting of OHLC data but also is an effective feature-extraction method that can effectively improve the forecasting accuracy of machine-learning models for close prices.

Deep Fuzzy Variable C-Means Clustering Incorporated With Curriculum Learning

End-to-end deep clustering method utilizes deep neural networks to jointly learn representation features and clustering assignments. Although many k-means-friendly deep clustering models have been explored, the existing division-based methods tend to directly implement clustering with a specific number of clusters, which suffers from poor performance resulted from indistinguishable clusters, and contributes to bad local optimum. At the same time, the representation learning of fuzzy <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$c$</tex-math></inline-formula> -means clustering in the feature space still needs more research. In this paper, a Deep Fuzzy Curriculum Clustering (DFC <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) method with the learning strategy of clustering from easy to complex automatically is proposed to tackle above issues. Firstly, considering the soft flexible allocation of fuzzy <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$c$</tex-math></inline-formula> -means and the preservation of local structure of original data, the fuzzy clustering loss and the autoencoder's reconstruction loss are constructed to learn the embedded features and clustering centers simultaneously. Second, curriculum loss is introduced into the constraint to make clusters successively merge in line with implementing clustering from easy to complex, and realize the bottom-up deep aggregative clustering automatically. In addition, novel curriculum information is proposed as constraint to guide the merging of clusters belonging to the same class. Experimental results on four real-world datasets show the superiority of the proposal.

Human Activity Recognition based on Local Linear Embedding and Geodesic Flow Kernel on Grassmann manifolds

Human Activity Recognition (HAR) plays a crucial role in various applications(e.g., medical treatment, video surveillance and sports monitoring). Transfer learning is a promising solution to cross-domain identification problems in HAR. However, existing methods usually ignore the negative transfer caused by using the features of each source domain in equal proportions, as well as the distribution difference between the source and target domains. In this paper, an HAR method based on manifold learning is proposed. Firstly, the similarity between the domain and multiple source domains is calculated using the Multi-Kernel-Maximum Mean Difference (MK-MMD), and the source domain most similar to the target domain is selected as the optimal source domain in the transfer task. Secondly, Locally Linear Embedding (LLE) is leveraged to reduce the dimensionality of both optimal source domain and target domain data to remove redundant information, and the Geodesic Flow Kernel (GFK) is utilized to project low-dimensional data into the Grassmann manifold space and reduce the distribution difference between the two domains. Finally, the source domain action training model is applied to the target domain. Three public datasets (i.e., PAMAP2, OPPORTUNITY and UCI DSADS) are utilized to validate the effectiveness of the proposed approach. Experimental results are presented to demonstrate that the proposed HAR method can predict a large number of unlabeled samples in the target domain while preserving the original data structure.

μ- PBWT: a lightweight r-indexing of the PBWT for storing and querying UK Biobank data.

The Positional Burrows-Wheeler Transform (PBWT) is a data structure that indexes haplotype sequences in a manner that enables finding maximal haplotype matches in h sequences containing w variation sites in O(hw) time. This represents a significant improvement over classical quadratic-time approaches. However, the original PBWT data structure does not allow for queries over Biobank panels that consist of several millions of haplotypes, if an index of the haplotypes must be kept entirely in memory. In this article, we leverage the notion of r-index proposed for the BWT to present a memory-efficient method for constructing and storing the run-length encoded PBWT, and computing set maximal matches (SMEMs) queries in haplotype sequences. We implement our method, which we refer to as μ-PBWT, and evaluate it on datasets of 1000 Genome Project and UK Biobank data. Our experiments demonstrate that the μ-PBWT reduces the memory usage up to a factor of 20% compared to the best current PBWT-based indexing. In particular, μ-PBWT produces an index that stores high-coverage whole genome sequencing data of chromosome 20 in about a third of the space of its BCF file. μ-PBWT is an adaptation of techniques for the run-length compressed BWT for the PBWT (RLPBWT) and it is based on keeping in memory only a succinct representation of the RLPBWT that still allows the efficient computation of set maximal matches (SMEMs) over the original panel. Our implementation is open source and available at https://github.com/dlcgold/muPBWT. The binary is available at https://bioconda.github.io/recipes/mupbwt/README.html.

Binary multi-view clustering with spectral embedding

Techniques for large-scale multi-view clustering have attracted increasing attention in recent years. To explore complementary and consensus information across multiple views, mapping multi-view data into a hamming space for binary multi-view learning is highly effective. However, for binary multi-view clustering, it is difficult to adopt the local manifold structure of original data into the consensus representation. Therefore, a novel algorithm named Binary Multi-View Clustering With Spectral Embedding (BMVC_SE) is proposed in this study. This algorithm unifies binary consensus representation learning and multi-view spectral learning into a joint framework. Specifically, this algorithm exploits the shareable and individual information across multi-view data to learn a binary consensus representation. Furthermore, spectral rotation technique is employed to embed manifold information from original multi-view data into binary consensus representation. Note that a multi-view GPI algorithm is proposed in this study to obtain the better local manifold information. We also provide an effective optimization method to solve the existing method. Extensive experimental results show that our method outperforms some state-of-the-art methods.

Unsupervised band selection of medical hyperspectral images guided by data gravitation and weak correlation

Background and ObjectiveMedical hyperspectral images (MHSIs) are used for a contact-free examination of patients without harmful radiation. However, high-dimensionality images contain large amounts of data that are sparsely distributed in a high-dimensional space, which leads to the “curse of dimensionality” (called Hughes’ phenomenon) and increases the complexity and cost of data processing and storage. Hence, there is a need for spectral dimensionality reduction before the clinical application of MHSIs. Some dimensionality-reducing strategies have been proposed; however, they distort the data within MHSIs. MethodsTo compress dimensionality without destroying the original data structure, we propose a method that involves data gravitation and weak correlation-based ranking (DGWCR) for removing bands of noise from MHSIs while clustering signal-containing bands. Band clustering is done by using the connection centre evolution (CCE) algorithm and selecting the most representative bands in each cluster based on the composite force. The bands within the clusters are ranked using the new entropy-containing matrix, and a global ranking of bands is obtained by applying an S-shaped strategy. The source code is available at https://www.github.com/zhangchenglong1116/DGWCR. ResultsUpon feeding the reduced-dimensional images into various classifiers, the experimental results demonstrated that the small number of bands selected by the proposed DGWCR consistently achieved higher classification accuracy than the original data. Unlike other reference methods (e.g. the latest deep-learning-based strategies), DGWCR chooses the spectral bands with the least redundancy and greatest discrimination. ConclusionIn this study, we present a method for efficient band selection for MHSIs that alleviates the “curse of dimensionality”. Experiments were validated with three MHSIs in the human brain, and they outperformed several other band selection methods, demonstrating the clinical potential of DGWCR.

Metabolite-disease interaction prediction based on logistic matrix factorization and local neighborhood constraints.

Increasing evidence indicates that metabolites are closely related to human diseases. Identifying disease-related metabolites is especially important for the diagnosis and treatment of disease. Previous works have mainly focused on the global topological information of metabolite and disease similarity networks. However, the local tiny structure of metabolites and diseases may have been ignored, leading to insufficiency and inaccuracy in the latent metabolite-disease interaction mining. To solve the aforementioned problem, we propose a novel metabolite-disease interaction prediction method with logical matrix factorization and local nearest neighbor constraints (LMFLNC). First, the algorithm constructs metabolite-metabolite and disease-disease similarity networks by integrating multi-source heterogeneous microbiome data. Then, the local spectral matrices based on these two networks are established and used as the input of the model, together with the known metabolite-disease interaction network. Finally, the probability of metabolite-disease interaction is calculated according to the learned latent representations of metabolites and diseases. Extensive experiments on the metabolite-disease interaction data were conducted. The results show that the proposed LMFLNC method outperformed the second-best algorithm by 5.28 and 5.61% in the AUPR and F1, respectively. The LMFLNC method also exhibited several potential metabolite-disease interactions, such as "Cortisol" (HMDB0000063), relating to "21-Hydroxylase deficiency," and "3-Hydroxybutyric acid" (HMDB0000011) and "Acetoacetic acid" (HMDB0000060), both relating to "3-Hydroxy-3-methylglutaryl-CoA lyase deficiency." The proposed LMFLNC method can well preserve the geometrical structure of original data and can thus effectively predict the underlying associations between metabolites and diseases. The experimental results show its effectiveness in metabolite-disease interaction prediction.

Survey on attribute and concept reduction methods in formal concept analysis

Formal concept analysis (FCA) is now widely recognized as a useful approach for extracting, representing, and analyzing knowledge in various domains. The high computational cost of knowledge processing and the difficulty of visualizing the lattice are two key challenges in practical FCA implementations. Moreover, assessing the finalized built-up lattice may be problematic due to the enormous number of formal concepts and the complexity of their connections. The challenge of constructing concept lattices of adequate size and structure to convey high-importance context features remains a significant FCA aim. In the literature, various strategies for concept lattice reduction have been presented. In this work, we suggest a categorization of reduction methods for concept lattice based on three main categories: context pre-processing, non-essential distinctions elimination, and concept filtration, whereby using FCA-based analysis, the most important methods in the literature are analyzed and compared based on six pillars: the preliminary step of the reduction process, domain expert, changing the original data structure, final concept lattice, quality of reduction, and category of reduction method.

Joint range, angle and polarization estimation in polarimetric FDA-MIMO radar based on Tucker tensor decomposition

Frequency diverse array multiple-input multiple-output (FDA-MIMO) radar is an emerging technology to offer range-angle-dependent beampattern. Polarimetric FDA-MIMO radar can sense additional polarization information to improve target identification capability. In this article, we investigate the problem of joint range, angle and polarization parameter estimation in a monostatic polarimetric FDA-MIMO radar with an FDA at transmitter and a cross-dipole array at receiver. Unlike the conventional methods in which the multidimensional data structure is rearranged into vectors or matrices by stacking operation, we propose a Tucker tensor decomposition-based scheme, which can reserve the original data structure and avoid spoiling the inherent characteristics of interest, especially when the number of snapshots is small. The third-order tensor model of the observed data is constructed. Two approaches named as Tucker covariance reconstruction and Tucker signal subspace are presented using the fourth-order covariance tensor decomposition. The Cramér–Rao bound for range, angle and polarization estimation is also provided. Numerical experiments demonstrate the superiorities of the proposed approaches. Specifically, two targets with identical range and close angles are effectively distinguished.

Multi-view clustering via matrix factorization assisted k-means

Matrix factorization based multi-view clustering algorithms has attracted much attention in recent years due to the strong interpretability and efficient implementation. In general, these approaches firstly compute the coefficient matrices of each data views, and learn a consensus matrix simultaneously. By applying the classical clustering techniques, such as k-means, on the consensus matrix, the final partition can be easily obtained. However, most of previous models work in a “step-by-step” manner, which cannot perform multi-view matrix factorization and clustering label generation simultaneously, leading to degenerated performance. In this paper, we propose a novel “one-pass” method, which integrates matrix factorization and k-means into a unified framework, named multi-view clustering via matrix factorization assisted k-means (MFK). In MFK, the generation of cluster indicator matrix and coefficient matrix learning can boost each other, leading to final improved clustering performance. Furthermore, we adopt a graph Laplacian regularization on the indicator matrix in order to capture the intrinsic geometric structure of original data. An alternating optimization strategy is designed to solve the resultant optimization problem and extensive experiments on six publicly datasets are conducted to demonstrate the superiority and effectiveness of the proposed MFK model.

Sparse Local Fisher Discriminant Analysis for Gas-Water Two-Phase Flow Status Monitoring With Multisensor Signals

Gas-water two-phase flow has typically stable flow statuses and constantly changing transition flow statuses. Accurate identification and real-time monitoring of flow status are conducive to the in-depth study of two-phase flow and the safe operation of industrial process. A monitoring strategy based on sparse local Fisher discriminant analysis (SLFDA) is proposed in this article. First, multisensor signals are obtained to reflect flow process information. Second, the least absolute shrinkage and selection operator is used to find the sparse discriminant directions to determine the key variables relevant to the flow process from multiple sensor signals. Then, the weight coefficient matrixes of SLFDA keep the original structure of the same flow status data and make the data of different flow statuses more separated, which distinguish different flow statuses to the maximum extent. Finally, two monitoring indexes including the discriminant index and the stability index are established to analyze the dynamic flow process, which enable a concurrent monitoring of both flow evolution and instability to realize fine-scale description of flow process. SLFDA can monitor various flow statuses through only one projection discriminant matrix by transforming high-dimensional signals into features representing the flow characteristics, which avoids model traversal and improves monitoring efficiency. Further study of flow status features provides meaningful physical interpretation and in-depth process analysis with consideration of actual flow process. The application on the data of gas-water two-phase flow in horizontal pipe demonstrates the feasibility and efficacy of the method.

Light Field Image Super-Resolution Network via Joint Spatial-Angular and Epipolar Information

This paper proposes a novel efficient convolutional neural network (CNN) for light field (LF) spatial super-resolution (SR). Due to the high dimensionality of LF data, it is important to decompose the original 4D LF data structure into a lower data subspace helping to reduce memory and computational costs. In this context, we propose a novel hybrid CNN structure working on different 2D data representations to fully handle the 4D structure. Our proposed network can simultaneously exploit information from three different types of 4D LF representations: spatial, angular, and epipolar plane image (EPI) information. Based on pixel similarities of each data subspace, we propose an inter-intra spatial feature extractor (II-SFE) and an inter-intra angular feature extractor (II-AFE) that can exploit the pixel correlations among the spatial-angular domain. Moreover, based on the correlated epipolar geometry among pixels, we also propose a multi-orientation epipolar feature extractor (MO-EFE) by decomposing an LF image into four orientation stacks of views and feeding them into different network branches to learn rich sub-pixel information in the EPI domain. The extracted features of each 2D subspace are integrated by our proposed attention fusion module (AFM), which adaptively rescales the embedding weights based on their correlations. Experimental results on both small- and wide-baseline benchmark LF datasets confirm the effectiveness of the proposed method compared to the well-known state-of-the-art methods in terms of objective and subjective performance.

Analysis-ready satellite data mosaics from Landsat and Sentinel-2 imagery

Today's enormous amounts of freely available high-resolution satellite imagery provide the demand for effective preprocessing methods. One such preprocessing method needed in many applications utilizing optical satellite imagery from the Landsat and Sentinel-2 archives is mosaicking. Merging hundreds of single scenes into a single satellite data mosaic before conducting analysis such as land cover classification, change detection, or modelling is often a prerequisite. Maintaining the original data structure and preserving metadata for further modelling or classification would be advantageous for many applications. Furthermore, in other applications, e.g., connected to land cover classification creating the mosaic for a specific period matching the phenological state of the phenomena in nature would be beneficial. In addition, supporting in-house and computing centers not directly connected to a specific cloud provider could be a requirement for some institutions or companies. In the current work, we present a method called Geomosaic that meets these criteria and produces analysis-ready satellite data mosaics from Landsat and Sentinel-2 imagery.•The method described produces analysis-ready satellite data mosaics.•The satellite data mosaics contain pixel metadata usable for further analysis.•The algorithm is available as an open-source tool coded in Python and can be used on multiple platforms.

Algebra of Data Reconciliation

With distributed computing and mobile applications becoming ever more prevalent, synchronizing diverging replicas of the same data is a common problem. Reconciliation – bringing two replicas of the same data structure as close as possible without overriding local changes – is investigated in an algebraic model. Our approach is to consider two sequences of simple commands that describe the changes in the replicas compared to the original structure, and then determine the maximal subsequences of each that can be propagated to the other. The proposed command set is shown to be functionally complete, and an update detection algorithm is presented which produces a command sequence transforming the original data structure into the replica while traversing both simultaneously. Syntactical characterization is provided in terms of a rewriting system for semantically equivalent command sequences. Algebraic properties of sequence pairs that are applicable to the same data structure are investigated. Based on these results the reconciliation problem is shown to have a unique maximal solution. In addition, syntactical properties of the maximal solution allow for an efficient algorithm that produces it.

Latent Multi-View Semi-Nonnegative Matrix Factorization with Block Diagonal Constraint

Multi-view clustering algorithms based on matrix factorization have gained enormous development in recent years. Although these algorithms have gained impressive results, they typically neglect the spatial structures that the latent data representation should have, for example, the ideal data representation owns a block structure just like the indicator matrix has. To address this issue, a new algorithm named latent multi-view semi-nonnegative matrix factorization with block diagonal constraint (LMSNB) is proposed. First, latent representation learning and Semi-NMF are combined to get a lower-dimensional representation with consistent information from different views. Second, the block diagonal constraint is able to capture the global structure of original data. In addition, the graph regularization is considered in our model to preserve the local structure. LMSNB can deal with negative data matrix and be applied to more fields. Although the low dimensional representation from semi-nonnegative matrix factorization loses some valuable information, it still has same structure as original data with the help of block diagonal constraint and graph regularization. Finally, an iterative optimization algorithm is proposed for our objective problem. Experiments on several multi-view benchmark datasets demonstrate the effectiveness of our approach against other state-of-the-art methods.

Concurrent size

The size of a data structure (i.e., the number of elements in it) is a widely used property of a data set. However, for concurrent programs, obtaining a correct size efficiently is non-trivial. In fact, the literature does not offer a mechanism to obtain a correct (linearizable) size of a concurrent data set without resorting to inefficient solutions, such as taking a full snapshot of the data structure to count the elements, or acquiring one global lock in all update and size operations. This paper presents a methodology for adding a concurrent linearizable size operation to sets and dictionaries with a relatively low performance overhead. Theoretically, the proposed size operation is wait-free with asymptotic complexity linear in the number of threads (independently of data-structure size). Practically, we evaluated the performance overhead by adding size to various concurrent data structures in Java−a skip list, a hash table and a tree. The proposed linearizable size operation executes faster by orders of magnitude compared to the existing option of taking a snapshot, while incurring a throughput loss of 1%−20% on the original data structure’s operations.