ISI Digital Commons (Indian Statistical Institute )
Not a member yet
7571 research outputs found
Sort by
Carbonaceous aerosol in the Brahmaputra plains: Sources, and influence from the hotspot Indo-Gangetic plains, India
No full textOrganic carbon (OC) and elemental carbon (EC) play a significant role in aerosol mass and atmospheric processes. This study is focused on the eastern part of the Great Northern Plains of India, namely the Brahmaputra Plains, to understand the influence of regional and local contribution on the carbonaceous fraction of PM2.5. Mean annual PM2.5 concentrations exceeded the National Ambient Air Quality Standards (NAAQS), with values of 46.6 ± 30.0 μg/m3 in the rural area and 50.4 ± 34.4 μg/m3 in the semi-urban area. The range in monsoon-winter was found to be 22.7–71.9 μg/m3. OC and EC contribute 44–50% of the PM2.5 mass concentration. The OC/EC ratios ranged from 3.3 to 9.3 in the rural area and from 4.3 to 6.9 in the semi-urban area, indicating significant secondary organic aerosol (SOA) formation, especially during the high photochemical period of the pre-monsoon season. Lower δ13C values were observed during winter (-27.5‰ rural, -27.3‰ semi-urban), pre-monsoon (-28.1‰ rural, -27.6‰ semi-urban), and post-monsoon (-28.2‰ rural, -28.1‰ semi-urban) periods, suggesting influences from biomass burning, fossil fuel combustion, and aged aerosols. The study employs cluster analysis of air mass trajectory, and Moderate Resolution Imaging Spectroradiometer (MODIS) fire data to determine the influence of the hotspot Indo-Gangetic Plain (IGP) and long-range transport on aerosol carbonaceous content during most seasons except the monsoon period June–September in the Brahmaputra Plains
CIRCUMCENTER EXTENSION OF MOEBIUS MAPS TO CAT(−1) SPACES
No full text— Given a Moebius homeomorphism f : ∂X → ∂Y between boundaries of proper, geodesically complete CAT(−1) spaces X, Y , we describe an extension fb: X → Y of f, called the circumcenter map of f, which is constructed using circumcenters of expanding sets. The extension fb is shown to coincide with the (1, log 2)-quasi-isometric extension constructed in a previous paper of the author, and is locally 1/2-Holder continuous. When X, Y are complete, simply connected manifolds with sectional curvatures K satisfying −b2 ≼ K ≼ −1 for some b ≽ 1 then the extension fb: X → Y is a (1, (1− 1b ) log 2)-quasi-isometry, and is surjective. Circumcenter extension of Moebius maps is natural with respect to composition with isometries
Classification of cancer microarray data using a two-step feature selection framework with moth-flame optimization and extreme learning machine
No full textAnalysis of microarray gene expression data for the detection/classification of cancer is one of the common approaches adopted worldwide. However, many genes (features) with correlated and irrelevant information in these data sets become the bottleneck for a classification model and significantly deteriorate its performance. A large number of features with fewer samples further make the classification task more cumbersome. Several feature selection methods (both filter and wrapper) have been proposed individually to address this issue, but choosing the best one among them is an open challenge. Our objective in the present study is to simplify the search for the best feature selection method without relying completely on individual methods and propose a two-step hybrid approach. In the first step, we use an ensemble of filter-based heterogeneous feature selection methods. These selected features then undergo the second step of wrapper-based selection. We propose to use the bio-inspired method called Moth-flame optimization (MFO) with an extreme learning machine (ELM) as its fitness function in this step. The motivation for using ELM is to leverage its learning strategy with one-pass processing of samples. Using this hybrid feature selection method, we proposed a classification model for Cancer Micraoarray data, where ELM is also considered as a classifier. The work demonstrates the superiority of the proposed model over other state-of-the-art methods in classifying cancer data from four different microarray gene expression datasets. Several measurement indexes are used for the performance evaluation of models
Comparisons of coherent systems with active redundancy and component lifetimes following the proportional odds model
No full textThe use of redundancies or spares in a system is a widely adopted technique to enhance system reliability and reduce the risk of system failure. Redundancies are typically incorporated into systems at the component or system levels. It is a significant problem to allocate appropriate redundancies into a system from a set of available options for the same. In this paper, we establish sufficient conditions to compare the reliability of coherent systems of dependent components with different sets of active redundancy, whether at the component level or the system level, based on some stochastic orders. We have obtained the results for the component lifetimes following the proportional odds (PO) model (the Marshall–Olkin family of distributions) for any lifetime distribution as a baseline distribution. We have studied the problem in the most general setup, with the consideration of coherent system that includes most of the common system structures, the consideration of non-matching spares, the consideration of dependencies of the components with different associated parameters of the copulas, and the consideration of general distribution as the baseline distribution of the PO model. We provide examples satisfying the sufficient conditions of the theoretical results. Additionally, we illustrate some of the results using real-world data
Correction to: Preface (Proceedings of the Indian National Science Academy, (2024), 90, 2, (161-165), 10.1007/s43538-024-00314-w)
No full textIn this article the date for the 37th IGC in Busan was incorrectly written as August 2-24 . The correct is August 2024 . The original article has been corrected
Current status data with two competing risks and time-dependent missing failure types
No full textIn competing risks data, in practice, there may be lack of information or uncertainty about the true failure type, termed as ‘missing failure type’, for some subjects. We consider a general pattern of missing failure type in which we observe, if not the true failure type, a set of possible failure types containing the true one. In this work, we focus on both parametric and non-parametric estimation based on current status data with two competing risks and the above-mentioned missing failure type. Here, the missing probabilities are assumed to be time-dependent, that is, dependent on both failure and monitoring time points, in addition to being dependent on the true failure type. This makes the missing mechanism non-ignorable. We carry out maximum likelihood estimation and obtain the asymptotic properties of the estimators. Simulation studies are conducted to investigate the finite sample properties of the estimators. Finally, the methods are illustrated through a data set on hearing loss
Degradable strong entanglement breaking maps
No full textIn this paper, we provide a structure theorem and various characterizations of degradable strong entanglement breaking maps on separable Hilbert spaces. In the finite-dimensional case, we prove that unital degradable entanglement breaking maps are precisely the C⁎-extreme points of the convex set of unital entanglement breaking maps on matrix algebras. Consequently, we get a structure for unital degradable positive partial transpose (PPT) maps
Distance-decay equations of antibiotic resistance genes across freshwater reservoirs
No full textDistance-decay (DD) equations can discern the biogeographical pattern of organisms and genes in a better way with advanced statistical methods. Here, we developed a data Compilation, Arrangement, and Statistics framework to advance quantile regression (QR) into the generation of DD equations for antibiotic resistance genes (ARGs) across various spatial scales using freshwater reservoirs as an illustration. We found that QR is superior at explaining dissemination potential of ARGs to the traditionally used least squares regression (LSR). This is because our model is based on the ‘law of limiting factors’, which reduces influence of unmeasured factors that reduce the efficacy of the LSR method. DD equations generated from the 99th QR model for ARGs were ‘Sall = 90.03e−0.01Dall’ in water and ‘Sall = 92.31e−0.011Dall’ in sediment. The 99th QR model was less impacted by uneven sample sizes, resulting in a better quantification of ARGs dissemination. Within an individual reservoir, the 99th QR model demonstrated that there is no dispersal limitation of ARGs at this smaller spatial scale. The QR method not only allows for construction of robust DD equations that better display dissemination of organisms and genes across ecosystems, but also provides new insights into the biogeography exhibited by key parameters, as well as the interactions between organisms and environment
DN3MF: deep neural network for non-negative matrix factorization towards low rank approximation
No full textDimension reduction is one of the most sought-after methodologies to deal with high-dimensional ever-expanding complex datasets. Non-negative matrix factorization (NMF) is one such technique for dimension reduction. Here, a multiple deconstruction multiple reconstruction deep learning model (DN3MF) for NMF targeted towards low rank approximation, has been developed. Non-negative input data has been processed using hierarchical learning to generate part-based sparse and meaningful representation. The novel design of DN3MF ensures the non-negativity requirement of the model. The use of Xavier initialization technique solves the exploding or vanishing gradient problem. The objective function of the model has been designed employing regularization, ensuring the best possible approximation of the input matrix. A novel adaptive learning mechanism has been developed to accomplish the objective of the model. The superior performance of the proposed model has been established by comparing the results obtained by the model with that of six other well-established dimension reduction algorithms on three well-known datasets in terms of preservation of the local structure of data in low rank embedding, and in the context of downstream analyses using classification and clustering. The statistical significance of the results has also been established. The outcome clearly demonstrates DN3MF’s superiority over compared dimension reduction approaches in terms of both statistical and intrinsic property preservation standards. The comparative analysis of all seven dimensionality reduction algorithms including DN3MF with respect to the computational complexity and a pictorial depiction of the convergence analysis for both stages of DN3MF have also been presented
Domains where the uniform rule is well behaved
No full textWe consider the problem of dividing one unit of an infinitely divisible object among a finite number of agents. We provide a characterization of all single-peaked domains on which the uniform rule is the unique division rule satisfying efficiency, strategy-proofness, and equal treatment of equals (ETE). Next, we consider non-single-peaked domains and provide a characterization of all such domains on which the uniform rule satisfies efficiency, strategy-proofness, and ETE. We also show that under some mild richness conditions the uniform rule is the unique rule satisfying the mentioned properties on these domains. Finally, we provide a wide range of applications to justify the usefulness of our results