UMD Theses and Dissertations

Permanent URI for this collectionhttp://hdl.handle.net/1903/3

New submissions to the thesis/dissertation collections are added automatically as they are received from the Graduate School. Currently, the Graduate School deposits all theses and dissertations from a given semester after the official graduation date. This means that there may be up to a 4 month delay in the appearance of a given thesis/dissertation in DRUM.

More information is available at Theses and Dissertations at University of Maryland Libraries.

Browse

Subject: search.filters.subject.data analysis

Search Results

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    AN ANALYSIS OF VARIABILITY IN NEWSPAPER REPORTING OF NATIONALLY COVERED TOPICS
    (2019) Marciano, Fernando Santos; Butler, Brian; Master in Information Management; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    This study looks into variability in newspaper reporting of important national topics. Using a sample of seven hundred articles from ten newspapers, a quantitative analysis using analysis of variance (ANOVA) found that variability in the values of reported numerical facts was statistically significant. Further, a qualitative analysis was performed to identifying potential reasons for the variability. Variability is unique to each topic being covered; however, some follow predictable patterns, such as updates on developing stories.
  • Thumbnail Image
    Item
    Exploiting Data-Dependent Structure for Improving Sensor Acquisition and Integration
    (2014) Cloninger, Alexander; Czaja, Wojciech; Benedetto, John J; Applied Mathematics and Scientific Computation; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    This thesis deals with two approaches to building efficient representations of data. The first is a study of compressive sensing and improved data acquisition. We outline the development of the theory, and proceed into its uses in matrix completion problems via convex optimization. The aim of this research is to prove that a general class of measurement operators, bounded norm Parseval frames, satisfy the necessary conditions for random subsampling and reconstruction. We then demonstrate an example of this theory in solving 2-dimensional Fredholm integrals with partial measurements. This has large ramifications in improved acquisition of nuclear magnetic resonance spectra, for which we give several examples. The second part of this thesis studies the Laplacian Eigenmaps (LE) algorithm and its uses in data fusion. In particular, we build a natural approximate inversion algorithm for LE embeddings using L1 regularization and MDS embedding techniques. We show how this inversion, combined with feature space rotation, leads to a novel form of data reconstruction and inpainting using a priori information. We demonstrate this method on hyperspectral imagery and LIDAR. We also aim to understand and characterize the embeddings the LE algorithm gives. To this end, we characterize the order in which eigenvectors of a disjoint graph emerge and the support of those eigenvectors. We then extend this characterization to weakly connected graphs with clusters of differing sizes, utilizing the theory of invariant subspace perturbations and proving some novel results.
  • Thumbnail Image
    Item
    Proton and Helium Spectra from the First Flight of the CREAM Balloon-Borne Experiment
    (2010) Yoon, Young Soo; Seo, EunSuk; Physics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    Cosmic-ray proton and helium spectra have been measured with the balloon-borne Cosmic Ray Energetics And Mass (CREAM) experiment flown for 42 days in Antarctica in the 2004-2005 austral summer season. High-energy cosmic-ray data were collected at an average altitude of ∼38.5 km with an average atmospheric overburden of ∼3.9 g/cm2. Individual elements are clearly separated with a charge resolution of ∼0.15e (in charge units) and ∼0.2e, respectively, for protons and helium nuclei. The measured spectra at the top of the atmosphere are represented by a power law with a spectral index of -2.66 ± 0.02 for protons from 2.5 TeV to 250 TeV and -2.58 ± 0.02 for helium nuclei from 630 GeV/nucleon to 63 TeV/nucleon. The measured proton and helium spectra are harder than previous measurements at a few tens of GeV/nucleon. Possible explanations of this spectral hardening could be the effect of a relatively nearby source or the effect of spectral concavity caused by interactions of cosmic rays with the accelerating shock. The helium flux is higher than that expected from extrapolation of a power-law fit to the lower-energy data. The relative abundance of protons to helium nuclei is about 8.8 ± 0.5 in the range from 2.5 TeV/nucleon to 63 TeV/nucleon. In this thesis, the analysis of proton and helium spectra will be discussed.
  • Thumbnail Image
    Item
    Improving analytical templates and searching for gravitational waves from coalescing black hole binaries
    (2010) Ochsner, Evan Lee; Buonanno, Alessandra; Physics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    The Laser Interferometer Gravitational-wave Observatory (LIGO) and Virgo are taking data at design sensitivity. They will be upgraded to Advanced LIGO and Virgo within the next 5 years and the detection of gravitational waves will be very likely. Binaries of two compact objects which inspiral and coalesce are one of the most promising sources for LIGO and Virgo. Most searches have focused solely on the inspiral portion of the waveform, and are consequently limited to low total mass. Recent breakthroughs in numerical relativity allow one to construct complete inspiral-merger-ringdown waveforms and search for the whole signal. This thesis will review some of the basic characteristics of gravitational waves from compact binaries and methods of searching for them. Analytical template waveforms for such systems will be presented including a comparison of different families of analytical waveforms, a study on the inclusion of spin effects in such waveforms, and a study of inspiral-merger-ringdown waveforms with amplitude corrections and the importance of these effects for parameter estimation. The thesis will culminate with a presentation of the first gravitational wave search to use inspiral-merger-ringdown templates, which was performed on data from the fifth science run of LIGO.
  • Thumbnail Image
    Item
    APPLYING NUMERICAL RELATIVITY TO GRAVITATIONAL WAVE ASTRONOMY
    (2008-03-12) McWilliams, Sean Thomas; Shawhan, Peter; Physics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)
    General relativity predicts the existence of gravitational waves produced by the motion of massive objects. The inspiral, merger, and ringdown of black hole binaries is expected to be one of the brightest sources in the gravitational wave sky. Interferometric detectors, such as the current ground-based Laser Interferometer Gravitational Wave Observatory (LIGO) and the future space-based Laser Interferometer Space Antenna (LISA), measure the influx of gravitational radiation from the whole sky. Each physical process that emits gravitational radiation will have a unique waveform, and prior knowledge of these waveforms is needed to distinguish a signal from the noise inherent in the interferometer. In the strong field regime of the merger, only numerical relativity, which solves the full set of Einstein's equations numerically, has been able to provide accurate waveforms. We present a comprehensive study of the nonspinning portion of parameter space for which we have generated accurate simulations of the late inspiral through merger and ringdown, and a comparison of those results with predictions from the adiabatic Taylor-expanded post-Newtonian (PN) and effective-one-body (EOB) PN approximations. We then focus on data analysis questions using the equal-mass nonspinning as well as the 2:1, 4:1, and 6:1 mass ratio nonspinning black hole binary (BHB) waveforms. We construct a full waveform by combining our results from numerical relativity with a highly accurate Taylor PN approximation, and use it to calculate signal-to-noise ratios (SNRs) for several detectors. We measure the mass ratio scaling of the waveform amplitude through the inspiral and merger, and compare our observations with predictions from PN. Lastly, we turn our focus to parameter estimation with LISA, and investigate the increased accuracy with which parameters can be measured by including both the merger and inspiral waveforms, compared to estimates without numerical waveforms which can only incorporate the inspiral. We use the equal mass, nonspinning waveform as a test case and assess the parameter uncertainty by means of the Fisher matrix formalism.