Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Feature importance”

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.

At least 1,081 records · Page 60
All sourcesNASA NTRSDOE OSTI
NASA NTRS1991

Decision boundary feature selection for non-parametric classifier

Feature selection has been one of the most important topics in pattern recognition. Although many authors have studied feature selection for parametric classifiers, few algorithms are available for feature selection for nonparametric classifiers. In this paper we propose a new feature selection algorithm based on decision boundaries for nonparametric classifiers. We first note that feature selection for pattern recognition is equivalent to retaining 'discriminantly informative features', and a discriminantly informative feature is related to the decision boundary. A procedure to extract discriminantly informative features based on a decision boundary for nonparametric classification is proposed. Experiments show that the proposed algorithm finds effective features for the nonparametric classifier with Parzen density estimation.

Lee, Chulhee↗
NASA NTRS2011

Complete Hard X-Ray Surveys, AGN Luminosity Functions and the X-Ray Background

AGN are believed to make up most of the Cosmic X-Ray Background (CXB) above a few keV, but this background cannot be fully resolved at energies less than 10 keV due to absorption. The Swift/BAT and INTEGRAL missions are performing the first complete hard x-ray surveys with minimal bias due to absorption. The most recent results for both missions will be presented. Although the fraction of the CXB resolved by these surveys is small, it is possible to derive unbiased number counts and luminosity functions for AGN in the local universe. The survey energy range from 15-150 keV contains the important reflection and cutoff spectral features dominate the shape of the AGN contribution to the CXB. Average spectral characteristics of survey detected AGN will be presented and compared with model distributions. The numbers of hard x-ray blazars detected in these surveys are finally sufficient to estimate this important component's contribution the cosmic background. Constraints on CXB models and their significance will be discussed.

Tueller, Jack↗
NASA NTRS1993

Transforming AdaPT to Ada9x

How the concepts of AdaPT can be transformed into programs using the object oriented features proposed in the preliminary mapping for Ada9x are described. Emphasizing, as they do, the importance of data types as units of program, these features match well with the development of partitions as translations into Abstract Data Types which was exploited in the Ada83 translation covered in report R3. By providing a form of polymorphic type, the Ada83 version also gives support for the conformant partition idea which could be achieved in Ada83 only by using UNCHECKED CONVERSIONS. It is assumed that the reader understands AdaPT itself, but the translation into Ada83 is briefly reviewed, by applying it to a small example. This is then used to show how the same translation would be achieved in the 9x version. It is important to appreciate that the distribution features which are proposed in current mapping are not used or discussed in any detail, as those are not well matched to the AdaPT approach. Critical evaluation and comparison of these approaches is given in a separate report.

Goldsack, Stephen J.↗
DOE OSTI2024

Geographical Insights into Suicide Mortality Through Spatial Machine Learning

Suicide mortality is a leading cause of death in the United States, with an upward trend that emphasizes its significance as a public health issue. Previous research has employed global models like ordinary least squares (OLS) regression and local models such as geographically weighted regression (GWR). While local models are useful for analyzing spatial variations in suicide mortality, they share limitations with traditional global models, particularly about their inability to handle multi-collinearity and non-linear relationships. Machine learning approaches, like random forests (RF), can address some of these limitations but often fail to account for spatial variability. This gap highlights the need for spatial ML models specifically designed to tackle suicide mortality. This research seeks to fill this void by using a geographically weighted random forest model (GWRF) to examine the associations between county-level suicide mortality in the U.S. from 2010 to 2020 and various social and environmental determinants of health. A key aspect of our methodology is disciplined feature selection, which reduces the pool of explanatory variables by about 90%. This refinement enhances the explanatory power of both global (R2 improved from 0.59 to 0.67) and local (R2 improved from 0.64 to 0.67) RF models while reducing their run times. An analysis of the importance scores for these selected features reveals that the drivers of suicide mortality vary by context. Thus, to effectively address regional disparities and inform targeted public health interventions, a holistic approach that incorporates multiple county-level characteristics is essential.

Lebakula, Viswadeep [ORNL] (ORCID:0000000152935914↗
NASA NTRS1989

Feature utility in polarimetric radar image classification

The information content in polarimetric SAR images is examined, and the polarimetric image variables containing the information that is important to the classification of terrain features in the images are determined. It is concluded that accurate classification can be done when just over half of the image variables are retained. A reduction in image data dimensionality gives storage savings, and can lead to the improvement of classifier performance. In addition, it is shown that a simplified radar system with only phase-calibrated CO-POL or SINGLE TX channels can give classification performance which approaches that of a fully polarimetric radar.

Cumming, Ian G.↗
NASA NTRS1994

Supported transition metal catalysts for para- to ortho-hydrogen conversion

The main goal of this study was to develop and improve on existing catalysts for the conversion of ortho- to para-hydrogen. Starting with a commercially available Air Products nickel silicate, which had a beta value of 20, we were trying to synthesize catalysts that would be an improvement to AP. This was accomplished by preparing silicates with various metals as well as different preparation methods. We also prepared supported ruthenium catalysts by various techniques using several metal precursors to improve present technology. What was also found was that the activation conditions prior to catalytic testing was highly important for both the silicates and the supported ruthenium catalysts. While not the initial focus of the research, we made some interesting observations into the adsorption of H2 on ruthenium. This helped us to get a better understanding of how ortho- to para-H2 conversion takes place, and what features in a catalyst are important to optimize activity. Reactor design was the final area in which some interesting conclusions were drawn. As discussed earlier, the reactor catalyst bed must be constructed using straight 1/8 feet OD stainless steel tubing. It was determined that the use of 1/4 feet OD tubing caused two problems. First, the radius from the center of the bed to the wall was too great for thermal equilibrium. Since the reaction of ortho- to para-H2 is exothermic, the catalyst bed center was warmer than the edges. Second, the catalyst bed was too shallow using a 1/4 feet tube. This caused reactant blow-by which was thought to decrease the measured activity when the flow rate was increased. The 1/8 feet tube corrected both of these concerns.

Brooks, Christopher J.↗
DOE OSTI2024

Human limits in machine learning: prediction of potato yield and disease using soil microbiome data

Abstract Background The preservation of soil health is a critical challenge in the 21st century due to its significant impact on agriculture, human health, and biodiversity. We provide one of the first comprehensive investigations into the predictive potential of machine learning models for understanding the connections between soil and biological phenotypes. We investigate an integrative framework performing accurate machine learning-based prediction of plant performance from biological, chemical, and physical properties of the soil via two models: random forest and Bayesian neural network. Results Prediction improves when we add environmental features, such as soil properties and microbial density, along with microbiome data. Different preprocessing strategies show that human decisions significantly impact predictive performance. We show that the naive total sum scaling normalization that is commonly used in microbiome research is one of the optimal strategies to maximize predictive power. Also, we find that accurately defined labels are more important than normalization, taxonomic level, or model characteristics. ML performance is limited when humans can’t classify samples accurately. Lastly, we provide domain scientists via a full model selection decision tree to identify the human choices that optimize model prediction power. Conclusions Our study highlights the importance of incorporating diverse environmental features and careful data preprocessing in enhancing the predictive power of machine learning models for soil and biological phenotype connections. This approach can significantly contribute to advancing agricultural practices and soil health management.

Aghdam, Rosa↗
NASA NTRS2010

Exploring Martian Impact Craters: Why They are Important for the Search for Life

Fluvial features and evidence for aqueous alteration indicate that Mars was wet, at least partially and/or periodically, in the Noachian. Also, impact cratering appears to have been the dominant geological process [1] during that epoch. Thus, investigation of Noachian craters will further our understanding of this geologic process, its effects on the water-bearing Martian crust, and any life that may have been present at the time. Impact events disturbed and heated the water- and/or ice-bearing crust, likely initiated long-lived hydrothermal systems [2-4], and formed crater lakes [5], creating environments suitable for life [6]. Thus, Noachian impact craters are particularly important exploration targets because they provide a window into warm, water-rich environments of the past which were possibly conducive to life. In addition to the presence of lake deposits, assessment of the presence of hydrothermal deposits in the walls, floors and uplifts of craters is important in the search for life on Mars. Impact craters are also important for astrobiological exploration in other ways. For example, smaller craters can be used as natural excavation pits, and so can provide information and samples that would otherwise be inaccessible (e.g., [7]). In addition, larger (> ~75 km) craters can excavate material from a potentially habitable region, even on present-day Mars, located beneath a >5-km deep cryosphere.

Schwenzer, S. P.↗
NASA NTRS2018

Vortex Shedding Characteristics of the Wake of a Thin Flat Plate with a Circular Trailing Edge

The near and very near wake of a thin flat plate with a circular trailing edge are investigated with direct numerical simulations (DNS). Data obtained for two different Reynolds numbers (based on plate thickness, D) are the main focus of this study. The separating boundary layers are turbulent in both cases. An earlier investigation of one of the cases (Case F) showed shed vortices in the wake that were about 1.0 D to 4.0 D in spanwise length. Considerable variation in both the strength and frequency of these shed vortices was observed. One objective of the present investigation is to determine the important contributors to this variability in strength and frequency of shed vortices and their finite spanwise extent. Analysis of the data shows that streamwise vortices in the separating boundary layer play an important role in strengthening/weakening of the shed vortices and that high/low-speed streaks in the boundary layer are important contributors to variability in shedding frequency. Both these features of the boundary layer contribute to the finite extent of the vortices in the spanwise direction. The second plate DNS (Case G, with 40 percent of the plate thickness of Case F) shows that while shedding intensity is weaker than obtained in Case F, many of the wake features are similar to that of Case F. This is important in understanding the path to the wake of the thin plate with a sharp trailing edge where shedding is absent. Here we also test the efficacy of a functional relationship between the shedding frequency and the Reynolds numbers based on the boundary layer momentum thickness (Re (sub theta) and D (Re (sub D)); data for developing this behavioral model is from Cases F & G and five earlier DNSs of the flat plate wake.

Turbulent wakes↗
NASA NTRS2000

Interactive Computer Graphics

Aerospace data analysis tools that significantly reduce the time and effort needed to analyze large-scale computational fluid dynamics simulations have emerged this year. The current approach for most postprocessing and visualization work is to explore the 3D flow simulations with one of a dozen or so interactive tools. While effective for analyzing small data sets, this approach becomes extremely time consuming when working with data sets larger than one gigabyte. An active area of research this year has been the development of data mining tools that automatically search through gigabyte data sets and extract the salient features with little or no human intervention. With these so-called feature extraction tools, engineers are spared the tedious task of manually exploring huge amounts of data to find the important flow phenomena. The software tools identify features such as vortex cores, shocks, separation and attachment lines, recirculation bubbles, and boundary layers. Some of these features can be extracted in a few seconds; others take minutes to hours on extremely large data sets. The analysis can be performed off-line in a batch process, either during or following the supercomputer simulations. These computations have to be performed only once, because the feature extraction programs search the entire data set and find every occurrence of the phenomena being sought. Because the important questions about the data are being answered automatically, interactivity is less critical than it is with traditional approaches.

Kenwright, David↗
NASA NTRSDate not supplied

Geometric Acoustics for Aircraft Noise Scattering

This paper discusses aircraft noise scattering by geometric acoustics, which consists of the basic features of sound propagation and reflection in rays or ray tubes, diffraction by smooth geometry in terms of surface creeping waves, and diffraction by abrupt geometry features such as the wing trailing edges. Based on the classic theories of these features, prediction methodologies can be constructed for aircraft noise scattering. The methodologies, however, require important modifications and extensions to account for unique features in aircraft noise applications, for both current conventional aircraft and future unconventional designs. These modifications and extensions include the derivation of a general reflection coefficient that contains the effects of surface geometry curvature, surface impedance, and mean flow. Corrections to the basic formulation of diffraction, both smooth geometry and sharp edges, are formulated to account for the finite dimensions of practical applications. For smooth geometry diffraction, a second order correction is used to continuously transition from insolified to shadow zones and an approach is presented to compute the properties of the geodesic path of surface wave propagation. A model is developed to make use of the ray propagation formulation for incoherent and partially coherent propagation and scattering, which is an important phenomenon in aircraft noise. Mean flow effect is also included in the methodology for low Mach number flows. Examples of calculations based on these theoretical developments are presented to illustrate the unique features in aircraft noise applications.

Yueping Guo↗
NASA NTRS1994

Live from Antarctica: the Coldest, Windiest Place on Earth

In this first part of a four part 'Passport to Knowledge Special', hosted by Camille Jennings from Maryland Public Television, children from Maryland and Texas schools had the opportunity to directly interact with and ask questions of scientists and researchers in Antarctica live. The physical characteristics of Antarctica are featured, along with their effects on the human and microbiological organisms living in the region. The reasons behind the clothing worn in the Antarctic and the importance of the meteorological station are featured. Interviews with Professor Ian Dolziel (U of Texas) and Lt. commander John Joseph, NSFA (the head of the Navy Meteorology Center) occur with the school children, along with actual video footage of the surrounding geological features and geography. The 'Weatherops' is located at McMurdo Station, Antarctica.

Source record↗
NASA NTRS2008

GIDEP Batching Tool

This software provides internal, automated search mechanics of GIDEP (Government- Industry Data Exchange Program) Alert data imported from the GIDEP government Web site. The batching tool allows the import of a single parts list in tab-delimited text format into the local JPL GIDEP database. Delimiters from every part number are removed. The original part numbers with delimiters are compared, as well as the newly generated list without the delimiters. The two lists run against the GIDEP imports, and output any matches. This feature only works with Netscape 2.0 or greater, or Internet Explorer 4.0 or greater. The user selects the browser button to choose a text file to import. When the submit button is pressed, this script will import alerts from the text file into the local JPL GIDEP database. This batch tool provides complete in-house control over exported material and data for automated batch match abilities. The batching tool has the ability to match capabilities of the parts list to tables, and yields results that aid further research and analysis. This provides more control over GIDEP information for metrics and reports information not provided by the government site. This software yields results quickly and gives more control over external data from the government site in order to generate other reports not available from the external source. There is enough space to store years of data. The program relates to risk identification and management with regard to projects and GIDEP alert information encompassing flight parts for space exploration.

Fong, Danny↗
NASA NTRS2011

Titan's Ammonia Feature

NH3 has long been considered an important component in the formation and evolution of the outer planet satellites. NH3 is particularly important for Titan, since it may serve as the reservoir for atmospheric nitrogen. A brightening seen on Titan starting in 2004 may arise from a transient low-lying fog or surface coating of ammonia. The spectral shape suggests the ammonia is anhydrous, a molecule that hydrates quickly in the presence of water.

ammonia↗
NASA NTRS2005

E-Nose Vapor Identification Based on Dempster-Shafer Fusion of Multiple Classifiers

Electronic nose (e-nose) vapor identification is an efficient approach to monitor air contaminants in space stations and shuttles in order to ensure the health and safety of astronauts. Data preprocessing (measurement denoising and feature extraction) and pattern classification are important components of an e-nose system. In this paper, a wavelet-based denoising method is applied to filter the noisy sensor measurements. Transient-state features are then extracted from the denoised sensor measurements, and are used to train multiple classifiers such as multi-layer perceptions (MLP), support vector machines (SVM), k nearest neighbor (KNN), and Parzen classifier. The Dempster-Shafer (DS) technique is used at the end to fuse the results of the multiple classifiers to get the final classification. Experimental analysis based on real vapor data shows that the wavelet denoising method can remove both random noise and outliers successfully, and the classification rate can be improved by using classifier fusion.

Li, Winston↗
DOE OSTI2024

The Importance of the Social Aspects of Agrivoltaics

This presentation features a high-level synthesis of international social science research on agrivoltaics. The main takeaways from this body of scholarship - namely the importance of farmer engagement, cross-sector collaboration, and legal frameworks - is discussed in the context of NREL's "5 C's of Agrivoltaic Success Factors" framework. A nested model approach to the 5 C's is proposed to emphasize the importance of the social aspects of agrivoltaics (Compatibility and Collaboration).

agrivoltaics↗
NASA NTRS1991

IRAS low-resolution spectral observations of the 10 and 18 micron silicate emission feature

The IRAS LRS Atlas contains 8-23 micron spectra of 1816 stars with silicate emission features. In order to study the shapes of the silicate dust features, the 117 stars whose emission features appear to be optically thin and which have the best signal-to-noise ratio at the longest wavelengths were analyzed. Simple spherical dust shell models were calculated in both the optically thin and the slightly optical thick approximations. From the comparison of the predicted spectra of the dust and stellar continuum, the emissivity function Kappa(lambda) was derived. In the different stars, the shape of the 10-micron feature is either narrow or broad and it is peaked either at about 9.7 microns or at 10 microns. Either particle size effects (particles above 0.75 microns in radius) or optical depth effects (central optical depth about 1) could broaden the 10-micron feature. Chemical composition differences no doubt are also important, particularly as regards the position of the peak of the 10-micron feature. The stars with the peak at 10 microns are more closely confined to the Galactic plane than the stars with the peak at 9.7 microns. The shape of the 18-micron feature is essentially the same in all stars, and can be used to extend the interstellar extinction curve past 13 to 22 microns.

Simpson, Janet P.↗
DOE OSTI2021

Perovskite Solar Cells Go Bifacial—Mutual Benefits for Efficiency and Durability

Bifacial solar cells hold the potential to achieve a higher power output per unit area than conventional monofacial devices without significantly increasing manufacturing costs. However, efficient bifacial designs are challenging to implement in inorganic thin-film solar cells because of their short carrier lifetimes and high rear surface recombination. The emergence of perovskite photovoltaic (PV) technology creates a golden opportunity to realize efficient bifacial thin-film solar cells, owing to their outstanding optoelectronic properties and unique features of device physics. More importantly, transparent conducting oxide electrodes can prevent electrode corrosion by halide ions, mitigating one major instability issue of the perovskite devices. Here, the theory of bifacial PV devices is summarized and the advantages of bifacial perovskite solar cells, such as high power output, enhanced device durability, and low economic and environmental costs, are reviewed. The limitations and challenges for bifacial perovskite solar cells are also discussed. Lastly, the awareness of bifacial solar cells as a feasible commercialization pathway of perovskite PV for mainstream solar power generation and building-integrated PV is advocated and future research directions are suggested.

14 SOLAR ENERGY↗