Varied statistical problems and tests, volume 2
Multivariate statistical analysis including divergence, probability of misclassification, information theory, and matrix analysis
Source record↗
SEARCH · Search NASA
Results for “Multivariate statistical analysis”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
A comparison of limited-area energetic processes between observations and primitive equation model predictions
Energetic analyses of the NMC initial conditions and NMC six-layer primitive equation operational prediction model 12-hr forecast for a developing cyclone are presented. Consideration is given to the total kinetic energy, the energetics of the divergent and nondivergent flows and the baroclinic (vertical shear flow) and barotropic (vertical mean flow) components of the kinetic energy. It is found that the model initial conditions lose 10-15% of the kinetic energy at various levels compared to a limited-area multivariate statistical analysis of the observational data, leading to a decrease in the horizontal kinetic energy flux, a misrepresentation of the synoptic scale wave system in the 12-hr forecast. Similar results are obtained for the nondivergent flow, while the divergent flow energetics are not reproduced accurately by the model. The horizontal flux terms of the vertical mean and vertical shear energetics are also not found to be reproduced in the upper levels, although horizontal flux contributions to the baroclinic component are improved at middle and lower levels. Finally, vertical shear kinetic energy generation is found to be well represented in the model prediction, however kinetic energy conversion between vertical shear and mean flow is not reproduced in the lower layer.
Alpert, J. C.↗
The role of multispectral scanners as data sources for EPA hydrologic models
An estimated cost savings of 30% to 50% was realized from using LANDSAT-derived data as input into a program which simulates hydrologic and water quality processes in natural and man-made water systems. Data from the satellite were used in conjunction with EPA's 11-channel multispectral scanner to obtain maps for characterizing the distribution of turbidity plumes in Flathead Lake and to predict the effect of increasing urbanization in Montana's Flathead River Basin on the lake's trophic state. Multispectral data are also being studied as a possible source of the parameters needed to model the buffering capability of lakes in an effort to evaluate the effect of acid rain in the Adirondacks. Water quality in Lake Champlain, Vermont is being classified using data from the LANDSAT and the EPA MSS. Both contact-sensed and MSS data are being used with multivariate statistical analysis to classify the trophic status of 145 lakes in Illinois and to identify water sampling sites in Appalachicola Bay where contaminants threaten Florida's shellfish.
Slack, R.↗
Identifying environmental features for land management decisions
Multivariate statistical analysis and imaging processing techniques are being applied to the study of arid/semiarid environments, with emphasis on desertification. Field level indicators of land-soil biota degradation are being sifted out with staging up to the low aircraft reconnaissance level, to LANDSAT TM & MSS, and even to the AVHRR level. Three completed projects are reviewed: riparian habitat on the Humboldt River floodplain, Salt Lake County Urban expansion detection, and salinization/desertification detection in the delta area. Beginning projects summarized include: comparative condition of rangeland in Rush Valley; modeling a GIS/remote sensing data base for Cache County; universal soil loss equation applied to Pinyon-Juniper; relating MSS to ground radiometry near Battle Mountain; and riparian habitat mapping on Mary's River, Nevada.
Source record↗
Survey of chondrule average properties in H-, L-, and LL-group chondrites - Are chondrules the same in all unequilibrated ordinary chondrites?
The petrogenetic properties of chondrules in different unequilibrated ordinary chondrites (UOCs) are compared to averaged chondrule-suite values obtained from recent analyses of several H-group, L-group, and LL-group chondrites. The purpose of the study was to develop a data base for future statistical analyses of chondrite characteristics. Mean end-member compositions of olivine (mol percent Fa) and pyroxene (mol percent Fs) were used as indices of the relative degree of 'equilibration' of each chondrule suite. It is found that the bulk chondrule geometric-mean abundances of Na, Mg, and Ni are the same from one UOC to another, and show no major systematic trends related to the H-group, L-group, of LL-group parentage of the host chondrites. The patterns of rare-earth element abundances in the chondrules are also examined, and the results are compared with statistical analyses. It is concluded that multivariate statistical analysis of pooled UOC chondrule data is justified for chondrule bulk compositions, as long as the statistical results are not misinterpreted as the primary petrogenetic features of chondrules.
Gooding, J. L.↗
A statistical study of Mercurian crater classes applied to the emplacement of the intercrater plains
Two principal processes have been invoked for emplacement of the intercrater plains: (1) basin ejecta and autochthonous surface materials reworked by that ejecta or (2) volcanic flows. A new approach was taken: the multivariate statistical analysis of crater class. The results of the analysis constrain the formative process of the intercrater plains units to have affected craters on both terrains in a similar manner. Any emplacement process brief in temporal extent (compared to the period of late heavy bombardment), such as emplacement of ejecta from a single event or a brief period of extensive volcanism, violates this constraint. This analysis is briefly discussed.
Woronow, Alex↗
On the origin of globular clusters in elliptical galaxies
Multivariate statistical analysis is used to correlate the estimated total numbers of globular clusters in elliptical galaxies (N sub gc), and the numbers per unit luminosity (S sub N), with a variety of observed properties of the host galaxies. Both N sub gc and S sub N are correlated with the so-called fundamental plane of elliptical galaxies, suggesting that the processes which determine global properties of ellipticals also largely determine their globular cluster content, although a considerable residual scatter remians. The redder (presumably more metal-rich) galaxies were found to have more globular clusters, implying that a dissipative formation mechanism was involved. Galaxies which have more anisotropic velocity distributions were also found to have more clusters, suggesting that mergers may have been important in the production of globular systems.
Djorgovski, S.↗
The manifold of globular clusters
A multivariate statistical analysis of a large data set on the Galactic globular clusters indicates that at least 4 (possibly 5 or 6) parameters are needed in order to describe fully the global properties of these objects. Some nontrivial correlations are present in the data, and are described in the text. One of them is the scaling relation between the luminosity and the central velocity dispersion, whose possible physical origin is briefly discussed. Many of the systematic trends found in the data suggest that dynamical effects played an important or even a dominant role in determining the present-day global properties of globular clusters.
Djorgovski, S.↗
Extracting galactic structure parameters from multivariated density estimation
Multivariate statistical analysis, including includes cluster analysis (unsupervised classification), discriminant analysis (supervised classification) and principle component analysis (dimensionlity reduction method), and nonparameter density estimation have been successfully used to search for meaningful associations in the 5-dimensional space of observables between observed points and the sets of simulated points generated from a synthetic approach of galaxy modelling. These methodologies can be applied as the new tools to obtain information about hidden structure otherwise unrecognizable, and place important constraints on the space distribution of various stellar populations in the Milky Way. In this paper, we concentrate on illustrating how to use nonparameter density estimation to substitute for the true densities in both of the simulating sample and real sample in the five-dimensional space. In order to fit model predicted densities to reality, we derive a set of equations which include n lines (where n is the total number of observed points) and m (where m: the numbers of predefined groups) unknown parameters. A least-square estimation will allow us to determine the density law of different groups and components in the Galaxy. The output from our software, which can be used in many research fields, will also give out the systematic error between the model and the observation by a Bayes rule.
Chen, B.↗
Multivariable Parametric Cost Model for Ground Optical Telescope Assembly
A parametric cost model for ground-based telescopes is developed using multivariable statistical analysis of both engineering and performance parameters. While diameter continues to be the dominant cost driver, diffraction-limited wavelength is found to be a secondary driver. Other parameters such as radius of curvature are examined. The model includes an explicit factor for primary mirror segmentation and/or duplication (i.e., multi-telescope phased-array systems). Additionally, single variable models Based on aperture diameter are derived.
Stahl, H. Philip↗
Experiments with a three-dimensional statistical objective analysis scheme using FGGE data
A three-dimensional (3D), multivariate, statistical objective analysis scheme (referred to as optimum interpolation or OI) has been developed for use in numerical weather prediction studies with the FGGE data. Some novel aspects of the present scheme include: (1) a multivariate surface analysis over the oceans, which employs an Ekman balance instead of the usual geostrophic relationship, to model the pressure-wind error cross correlations, and (2) the capability to use an error correlation function which is geographically dependent. A series of 4-day data assimilation experiments are conducted to examine the importance of some of the key features of the OI in terms of their effects on forecast skill, as well as to compare the forecast skill using the OI with that utilizing a successive correction method (SCM) of analysis developed earlier. For the three cases examined, the forecast skill is found to be rather insensitive to varying the error correlation function geographically. However, significant differences are noted between forecasts from a two-dimensional (2D) version of the OI and those from the 3D OI, with the 3D OI forecasts exhibiting better forecast skill. The 3D OI forecasts are also more accurate than those from the SCM initial conditions. The 3D OI with the multivariate oceanic surface analysis was found to produce forecasts which were slightly more accurate, on the average, than a univariate version.
Baker, Wayman E.↗
A method of using cluster analysis to study statistical dependence in multivariate data
A technique is presented that uses both cluster analysis and a Monte Carlo significance test of clusters to discover associations between variables in multidimensional data. The method is applied to an example of a noisy function in three-dimensional space, to a sample from a mixture of three bivariate normal distributions, and to the well-known Fisher's Iris data.
Borucki, W. J.↗
An Integrated Analysis of the Physiological Effects of Space Flight: Executive Summary
A large array of models were applied in a unified manner to solve problems in space flight physiology. Mathematical simulation was used as an alternative way of looking at physiological systems and maximizing the yield from previous space flight experiments. A medical data analysis system was created which consist of an automated data base, a computerized biostatistical and data analysis system, and a set of simulation models of physiological systems. Five basic models were employed: (1) a pulsatile cardiovascular model; (2) a respiratory model; (3) a thermoregulatory model; (4) a circulatory, fluid, and electrolyte balance model; and (5) an erythropoiesis regulatory model. Algorithms were provided to perform routine statistical tests, multivariate analysis, nonlinear regression analysis, and autocorrelation analysis. Special purpose programs were prepared for rank correlation, factor analysis, and the integration of the metabolic balance data.
Leonard, J. I.↗
Data analysis techniques
A large and diverse number of computational techniques are routinely used to process and analyze remotely sensed data. These techniques include: univariate statistics; multivariate statistics; principal component analysis; pattern recognition and classification; other multivariate techniques; geometric correction; registration and resampling; radiometric correction; enhancement; restoration; Fourier analysis; and filtering. Each of these techniques will be considered, in order.
Park, Steve↗
External Influences on Modeled and Observed Cloud Trends
Understanding the cloud response to external forcing is a major challenge for climate science. This crucial goal is complicated by intermodel differences in simulating present and future cloud cover and by observational uncertainty. This is the first formal detection and attribution study of cloud changes over the satellite era. Presented herein are CMIP5 (Coupled Model Intercomparison Project - Phase 5) model-derived fingerprints of externally forced changes to three cloud properties: the latitudes at which the zonally averaged total cloud fraction (CLT) is maximized or minimized, the zonal average CLT at these latitudes, and the height of high clouds at these latitudes. By considering simultaneous changes in all three properties, the authors define a coherent multivariate fingerprint of cloud response to external forcing and use models from phase 5 of CMIP (CMIP5) to calculate the average time to detect these changes. It is found that given perfect satellite cloud observations beginning in 1983, the models indicate that a detectable multivariate signal should have already emerged. A search is then made for signals of external forcing in two observational datasets: ISCCP (International Satellite Cloud Climatology Project) and PATMOS-x (Advanced Very High Resolution Radiometer (AVHRR) Pathfinder Atmospheres - Extended). The datasets are both found to show a poleward migration of the zonal CLT pattern that is incompatible with forced CMIP5 models. Nevertheless, a detectable multivariate signal is predicted by models over the PATMOS-x time period and is indeed present in the dataset. Despite persistent observational uncertainties, these results present a strong case for continued efforts to improve these existing satellite observations, in addition to planning for new missions.
clouds (meteorology)↗
Characterization of exopolymers of aquatic bacteria by pyrolysis-mass spectrometry
Exopolymers from a diverse collection of marine and freshwater bacteria were characterized by pyrolysis-mass spectrometry (Py-MS). Py-MS provides spectra of pyrolysis fragments that are characteristic of the original material. Analysis of the spectra by multivariate statistical techniques (principal component and canonical variate analysis) separated these exopolymers into distinct groups. Py-MS clearly distinguished characteristic fragments, which may be derived from components responsible for functional differences between polymers. The importance of these distinctions and the relevance of pyrolysis information to exopolysaccharide function in aquatic bacteria is discussed.
NASA Program CELSS↗
Mercurian crater-filling classes constrain the emplacement process of the intercrater plains material
The multivariate method for statistical analysis of crater-filling classes, as presently applied to craters of more than 10-km diameter on Mercury, is noted to be superior to current techniques due to the obviation of any amalgamation of data from regions with different histories in order to arrive at the pertinent counting statistics. In this application to the Mercury intercrater plains, the process responsible for emplacement is constrained to be one that leaves the crater class distribution unaltered from that observed on the densely cratered terrain. The analysis explicitly eliminates the viability of the pulse emplacement of volcanics or basin ejecta.
Woronow, Alex↗
Orthostatic function during a stand test before and after head-up or head-down bedrest
The effects of head-down or head-up bedrest at -5, +10, +20, or +42 deg (simulating 0, 1/6, 1/3, and 2/3 g, respectively) for 6 hrs on four different days on the orthostatic tolerance were investigated by measuring relevant physiological reactions to orthostatic test taken before and after bedrest sessions. The multivariate analysis of variance statistical analyses indicates that there was no angle effect on any of the cardiovascular parameters monitored during the last 3 min of the stand test, suggesting that partial gravity loads would have no effect on the cardiovascular deconditioning exhibited postflight. There was, however, a significant elevation in the heart rate post-bedrest, and the heart rate increased on standing. Results from the stand test pre- and post-bedrest at -5 deg (but not at +10, +20, and +42 deg) were similar to those observed after space flight.
Lathers, Claire M.↗