Search NASA⌕ Search

SEARCH · Search NASA

Results for “data classification”

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 1,117 records · Page 62

Using active learning to improve quasar identification for the DESI spectra processing pipeline

The Dark Energy Spectroscopic Instrument (DESI) survey uses an automatic spectral classification pipeline to classify spectra. QuasarNET is a convolutional neural network used as part of this pipeline originally trained using data from the Baryon Oscillation Spectroscopic Survey (BOSS). In this paper we implement an active learning algorithm to optimally select spectra to use for training a new version of the QuasarNET weights file using only DESI data, with the goal of improving classification accuracy. This active learning algorithm includes a novel outlier rejection step using a Self-Organizing Map to ensure we label spectra representative of the larger quasar sample observed in DESI. We perform two iterations of the active learning pipeline, assembling a final dataset of 5600 labeled spectra, a small subset of the approximately 1.3 million quasar targets in DESI's Data Release 1. When splitting the spectra into training and validation subsets we achieve similar performance to the previously trained weights file in completeness and purity calculated on the validation dataset but do so with less than one tenth of the amount of training data. The new weights also more consistently classify objects in the same way when used on unlabeled data compared to the old weights file. In the process of improving QuasarNET's classification accuracy we discovered a systemic error in QuasarNET's redshift estimation and used our findings to improve our understanding of QuasarNET's redshifts.

Machine learning↗

Distributions and motions of nearby stars defined by objective prism surveys and Hipparcos data

Material and objective prism spectral classification work is used to determine the space density distribution of nearby common stars to the limits of objective prism spectral surveys. The aim is to extend the knowledge of the local densities of specific spectral types from a radius of 25 pc from the sun, as limited in the Gliese catalog of nearby stars, to 50 pc or more. Future plans for the application of these results to studies of the kinematic and dynamical properties of stars in the solar neighborhood as a function of their physical properties and ages are described.

Hemenway, P. D.↗

Determining crop residue type and class using satellite acquired data

LANDSAT Thematic Mapper (TM) data for March 23, 1987 with accompanying ground truth data for the study area in Miami County, IN were used to determine crop residue type and class. Principle components and spectral ratioing transformations were applied to the LANDSAT TM data. One graphic information system (GIS) layer of land ownership was added to each original image as the eighth band of data in an attempt to improve classification. Maximum likelihood, minimum distance, and neural networks were used to classify the original, transformed, and GIS-enhanced remotely sensed data. Crop residues could be separated from one another and from bare soil and other biomass. Two types of crop residue and four classes were identified from each LANDSAT TM image. The maximum likelihood classifier performed the best classification for each original image without need of any transformation. The neural network classifier was able to improve the classification by incorporating a GIS-layer of land ownership as an eighth band of data. The maximum likelihood classifier was unable to consider this eighth band of data and thus, its results could not be improved by its consideration.

Zhuang, Xin↗

Developing New Coastal Forest Restoration Products Based on Landsat, ASTER, and MODIS Data

This paper discusses an ongoing effort to develop new geospatial information products for aiding coastal forest restoration and conservation efforts in coastal Louisiana and Mississippi. This project employs Landsat, Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data in conjunction with airborne elevation data to compute coastal forest cover type maps and change detection products. Improved forest mapping products are needed to aid coastal forest restoration and management efforts of State and Federal agencies in the Northern Gulf of Mexico (NGOM) region. In particular, such products may aid coastal forest land acquisition and conservation easement procurements. This region's forests are often disturbed and subjected to multiple biotic and abiotic threats, including subsidence, salt water intrusion, hurricanes, sea-level rise, insect-induced defoliation and mortality, altered hydrology, wildfire, and conversion to non-forest land use. In some cases, such forest disturbance has led to forest loss or loss of regeneration capacity. In response, a case study was conducted to assess and demonstrate the potential of satellite remote sensing products for improving forest type maps and for assessing forest change over the last 25 years. Change detection products are needed for assessing risks for specific priority coastal forest types, such as live oak and baldcypress-dominated forest. Preliminary results indicate Landsat time series data are capable of generating the needed forest type and change detection products. Useful classifications were obtained using 2 strategies: 1) general forest classification based on use of 3 seasons of Landsat data from the same year; and 2) classification of specific forest types of concern using a single date of Landsat data in which a given targeted type is spectrally distinct compared to adjacent forested cover. When available, ASTER data was useful as a complement to Landsat data. Elevation data helped to define areas in which targeted forest types occur, such as live oak forests on natural levees. MODIS Normalized Difference Vegetation Index time series data aided visual assessments of coastal forest damage and recovery from hurricanes. Landsat change detection products enabled change to be identified at the stand level and at 10- year intervals with the earliest date preceding available change detection products from the National Oceanic and Atmospheric Administration and from the U.S. Geological Survey. Additional work is being done in collaboration with State and Federal agency partners in a follow-on NASA ROSES project to refine and validate these new, promising products. The products from the ROSES project will be available for aiding NGOM coastal forest restoration and conservation.

Spruce, Joseph P.↗

Automated solar collector installation design including version management

Embodiments may include systems and methods to create and edit a representation of a worksite, to create various data objects, to classify such objects as various types of pre-defined “features” with attendant properties and layout constraints. As part of or in addition to classification, an embodiment may include systems and methods to create, associate, and edit intrinsic and extrinsic properties to these objects. A design engine may apply of design rules to the features described above to generate one or more solar collectors installation design alternatives, including generation of on-screen and/or paper representations of the physical layout or arrangement of the one or more design alternatives. Some embodiments may provide viewing, creating, and manipulating of multiple versions of a solar collector layout design for a particular installation worksite. The use of versions may allow analysis of alternative layouts, alternative feature classifications, and cost and performance data corresponding to alternative design choices. Version summary information providing a representative comparison between versions across a number of dimensions may be provided.

14 SOLAR ENERGY↗

Logistic classification for tool life modeling in machining

This paper describes the application of logistic classification for tool life modeling and prediction in an industrial setting using shop floor data. Tool life is treated as a classification problem since tool wear can only be measured at the time of tool replacement in a production environment. Laboratory tool wear experiments are used to simulate shop floor wear data by two states: not worn (class 0); and worn (class 1). To incorporate non-linearity in logistic classification, a log-transformation of input features is performed. The logistic classification approach, results, and interpretability of the logistic model are presented.

Karandikar, Jaydeep↗

Fast Image Texture Classification Using Decision Trees

Texture analysis would permit improved autonomous, onboard science data interpretation for adaptive navigation, sampling, and downlink decisions. These analyses would assist with terrain analysis and instrument placement in both macroscopic and microscopic image data products. Unfortunately, most state-of-the-art texture analysis demands computationally expensive convolutions of filters involving many floating-point operations. This makes them infeasible for radiation- hardened computers and spaceflight hardware. A new method approximates traditional texture classification of each image pixel with a fast decision-tree classifier. The classifier uses image features derived from simple filtering operations involving integer arithmetic. The texture analysis method is therefore amenable to implementation on FPGA (field-programmable gate array) hardware. Image features based on the "integral image" transform produce descriptive and efficient texture descriptors. Training the decision tree on a set of training data yields a classification scheme that produces reasonable approximations of optimal "texton" analysis at a fraction of the computational cost. A decision-tree learning algorithm employing the traditional k-means criterion of inter-cluster variance is used to learn tree structure from training data. The result is an efficient and accurate summary of surface morphology in images. This work is an evolutionary advance that unites several previous algorithms (k-means clustering, integral images, decision trees) and applies them to a new problem domain (morphology analysis for autonomous science during remote exploration). Advantages include order-of-magnitude improvements in runtime, feasibility for FPGA hardware, and significant improvements in texture classification accuracy.

Thompson, David R.↗

Vegetation classification based on Advanced Very High Resolution Radiometer /AVHRR/ satellite imagery

Data from the NOAA-6 spacecraft Advanced Very High Resolution Radiometer (AVHRR) were tested for effectiveness for vegetation classification. Vegetation, climatological, and meteorological data were gathered for three days over 12 locations, and the normalized differences between the AVHRR bands 1 and 2 were determined. A vegetative greenness index was compared with a hydrologic factor and vegetation characteristics as measured by ground truth. A multivariate vegetation gradient model was formulated, incorporating AVHRR and climatological data. The hydrologic factor was calculated in terms of the precipitation, evaporation, maximum and minimum temperatures, and the hydrologic capacity. The observations were taken over Texas, which has a wide range of climates. A high correlation was found in the vegetation-HF index. The AVHRR data are concluded to be an effective tool for analysis of vegetation/climate relationships.

Norwine, J.↗

Evaluating the Computational Requirements of using SVM software to train Data-Intensive Problems

Support vector machine (SVM) technology is a growing class of computational algorithms for solving classification problems using supervised learning. There are many freely available SVM codes with implementations of varying granularity in the core optimization task. The performance of SVM implementations is related to four key elements; (1) the size and dimension of the data set on which they are operating, (2) the granularity of their core SVM optimization implementation, (3) the kernel transformation applied to the data, and the (4) underlying hardware on which the implementation is running. To assess the performance of different SVM implementations, several freely available codes representing the spectrum of optimization granularity were built on a variety of hardware—two linux clusters, and a shared memory machine. Binary classifiers were trained on datasets of varying size containing anywhere from a few hundred to over 50,000 vectors. Performance of the methods was measured in terms of wall-clock time for training, statistical quality of the resulting classifier, hardware performance (memory footprint, memory bandwidth, floating point unit usage, I/O bandwidth, and instruction stalls), robustness and portability.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Final Comparison of TM and MSS Data for Surface Mine Assessment in Logan County, West Virginia

A variety of classifications during both raw and transformed MSS and TM data sets from 4 September 1982 were performed for the Logan County, West Virginia study area. The object was to compare the utility of TM and MSS data for delineating small, irregular ground features, particularly surface mines, and also to test data reduction/transformation techniques (band selection, canonical analysis, and principal components) in relation to a traditional means of unsupervised classification. Statistical results demonstrate that, on the average, the TM classifications yielded an overall .53 factor of improvement relative to the MSS classifications. When the accuracies for only three minor (in terms of areal extent) land use categories are examined, the factor of improvement for TM over MSS increases to 1.48; i.e., the TM is nearly one and one-half times better than the MSS for delineating small and irregular ground features such as contour strip mines.

Witt, R. G.↗

Question-answering system extracts information on injection drug use from clinical notes

Background. Injection drug use (IDU) can increase mortality and morbidity. Therefore, identifying IDU early and initiating harm reduction interventions can benefit individuals at risk. However, extracting IDU behaviors from patients’ electronic health records (EHR) is difficult because there is no other structured data available, such as International Classification of Disease (ICD) codes, and IDU is most often documented in unstructured free-text clinical notes. Although natural language processing can efficiently extract this information from unstructured data, there are no validated tools. Methods. Here, to address this gap in clinical information, we design a question-answering (QA) framework to extract information on IDU from clinical notes for use in clinical operations. Our framework involves two main steps: (1) generating a gold-standard QA dataset and (2) developing and testing the QA model. We use 2323 clinical notes of 1145 patients curated from the US Department of Veterans Affairs (VA) Corporate Data Warehouse to construct the gold-standard dataset for developing and evaluating the QA model. We also demonstrate the QA model’s ability to extract IDU-related information from temporally out-of-distribution data. Results. Here, we show that for a strict match between gold-standard and predicted answers, the QA model achieves a 51.65% F1 score. For a relaxed match between the gold-standard and predicted answers, the QA model obtains a 78.03% F1 score, along with 85.38% Precision and 79.02% Recall scores. Moreover, the QA model demonstrates consistent performance when subjected to temporally out-of-distribution data. Conclusions. Our study introduces a QA framework designed to extract IDU information from clinical notes, aiming to enhance the accurate and efficient detection of people who inject drugs, extract relevant information, and ultimately facilitate informed patient care.

60 APPLIED LIFE SCIENCES↗

PHOTOMETRIC SUPERNOVA CLASSIFICATION WITH MACHINE LEARNING

Automated photometric supernova classification has become an active area of research in recent years in light of current and upcoming imaging surveys such as the Dark Energy Survey (DES) and the Large Synoptic Survey Telescope, given that spectroscopic confirmation of type for all supernovae discovered will be impossible. Here, we develop a multi-faceted classification pipeline, combining existing and new approaches. Our pipeline consists of two stages: extracting descriptive features from the light curves and classification using a machine learning algorithm. Our feature extraction methods vary from model-dependent techniques, namely SALT2 fits, to more independent techniques that fit parametric models to curves, to a completely model-independent wavelet approach. We cover a range of representative machine learning algorithms, including naive Bayes, k -nearest neighbors, support vector machines, artificial neural networks, and boosted decision trees (BDTs). We test the pipeline on simulated multi-band DES light curves from the Supernova Photometric Classification Challenge. Using the commonly used area under the curve (AUC) of the Receiver Operating Characteristic as a metric, we find that the SALT2 fits and the wavelet approach, with the BDTs algorithm, each achieve an AUC of 0.98, where 1 represents perfect classification. We find that a representative training set is essential for good classification, whatever the feature set or algorithm, with implications for spectroscopic follow-up. Importantly, we find that by using either the SALT2 or the wavelet feature sets with a BDT algorithm, accurate classification is possible purely from light curve data, without the need for any redshift information.

79 ASTRONOMY AND ASTROPHYSICS↗

Automatic discovery of optimal classes

A criterion, based on Bayes' theorem, is described that defines the optimal set of classes (a classification) for a given set of examples. This criterion is transformed into an equivalent minimum message length criterion with an intuitive information interpretation. This criterion does not require that the number of classes be specified in advance, this is determined by the data. The minimum message length criterion includes the message length required to describe the classes, so there is a built in bias against adding new classes unless they lead to a reduction in the message length required to describe the data. Unfortunately, the search space of possible classifications is too large to search exhaustively, so heuristic search methods, such as simulated annealing, are applied. Tutored learning and probabilistic prediction in particular cases are an important indirect result of optimal class discovery. Extensions to the basic class induction program include the ability to combine category and real value data, hierarchical classes, independent classifications and deciding for each class which attributes are relevant.

Cheeseman, Peter↗

Onboard Hyperspectral Image Classification via Transfer Learning for Communication-Limited Spacecraft

Employing deep-learning and artificial-intelligence (AI) techniques onboard spacecraft can dramatically improve priority data selection to ensure more effective use of the available downlink. However, deployment of effective deep-learning models requires significant training on the ground, which may not be feasible, due to limited data available in an unexplored environment. Therefore, this research explores building robust classification models for onboard data processing where training data is highly limited using transfer-learning techniques. In this paper, we focus on the use case of hyperspectral imaging for remote sensing, a domain where the high dimensionality of the data from the sensor can rapidly saturate the downlink bandwidth. With this bottleneck, there is an impending need to autonomously and robustly classify data onboard to optimize downlink of high-impact measurements, thus maximizing the scientific utility per bit transmitted to the ground. This paper examines the use of deep neural networks onboard for hyperspectral image classification in a communication-limited scenario to analyze how the models perform with limited training data. The use of transfer learning can ameliorate the issue of poor generalization by transferring features learned from training on a large source dataset for one classification task to the target classification task with limited training data. For two deep-learning models from literature, we compare the accuracy of the models trained using transfer learning to models trained from scratch using a random weight initialization with varying amounts of training data. We demonstrate the feasibility and performance of running inference of the deep-learning models on representative flight-like hardware.

Advanced Avionics, Machine Learning, Data Processi↗

The 13th Data Release of the Sloan Digital Sky Survey: First Spectroscopic Data from the SDSS-IV Survey Mapping Nearby Galaxies at Apache Point Observatory

Here, the fourth generation of the Sloan Digital Sky Survey (SDSS-IV) began observations in 2014 July. It pursues three core programs: the Apache Point Observatory Galactic Evolution Experiment 2 (APOGEE-2), Mapping Nearby Galaxies at APO (MaNGA), and the Extended Baryon Oscillation Spectroscopic Survey (eBOSS). As well as its core program, eBOSS contains two major subprograms: the Time Domain Spectroscopic Survey (TDSS) and the SPectroscopic IDentification of ERosita Sources (SPIDERS). This paper describes the first data release from SDSS-IV, Data Release 13 (DR13). DR13 makes publicly available the first 1390 spatially resolved integral field unit observations of nearby galaxies from MaNGA. It includes new observations from eBOSS, completing the Sloan Extended QUasar, Emission-line galaxy, Luminous red galaxy Survey (SEQUELS), which also targeted variability-selected objects and X-ray-selected objects. DR13 includes new reductions of the SDSS-III BOSS data, improving the spectrophotometric calibration and redshift classification, and new reductions of the SDSS-III APOGEE-1 data, improving stellar parameters for dwarf stars and cooler stars. DR13 provides more robust and precise photometric calibrations. Value-added target catalogs relevant for eBOSS, TDSS, and SPIDERS and an updated red-clump catalog for APOGEE are also available. This paper describes the location and format of the data and provides references to important technical papers. The SDSS web site provides links to the data, tutorials, examples of data access, and extensive documentation of the reduction and analysis procedures. DR13 is the first of a scheduled set that will contain new data and analyses from the planned ~6 yr operations of SDSS-IV.

79 ASTRONOMY AND ASTROPHYSICS↗

Analysis of the classification of US and Canadian intensive test sites using the Image 100 hybrid classification system

The author has identified the following significant results. Labeling of wheat rather than total grains, particularly with only one acquisition, led to significant overestimates in some segments. The Image-100 software and procedures were written to facilitate classification of the LACIE segments but were not designed to record data for later accuracy assessment. A much better evaluation would have been possible if accuracy assessment data had been collected following each satisfactory classification.

Hocutt, W. T.↗

The Transition from Complex Crater to Peak-Ring Basin on the Moon: New Observations from the Lunar Orbiter Laser Altimeter (LOLA) Instrument

Impact craters on planetary bodies transition with increasing size from simple, to complex, to peak-ring basins and finally to multi-ring basins. Important to understanding the relationship between complex craters with central peaks and multi-ring basins is the analysis of protobasins (exhibiting a rim crest and interior ring plus a central peak) and peak-ring basins (exhibiting a rim crest and an interior ring). New data have permitted improved portrayal and classification of these transitional features on the Moon. We used new 128 pixel/degree gridded topographic data from the Lunar Orbiter Laser Altimeter (LOLA) instrument onboard the Lunar Reconnaissance Orbiter, combined with image mosaics, to conduct a survey of craters >50 km in diameter on the Moon and to update the existing catalogs of lunar peak-ring basins and protobasins. Our updated catalog includes 17 peak-ring basins (rim-crest diameters range from 207 km to 582 km, geometric mean = 343 km) and 3 protobasins (137-170 km, geometric mean = 157 km). Several basins inferred to be multi-ring basins in prior studies (Apollo, Moscoviense, Grimaldi, Freundlich-Sharonov, Coulomb-Sarton, and Korolev) are now classified as peak-ring basins due to their similarities with lunar peak-ring basin morphologies and absence of definitive topographic ring structures greater than two in number. We also include in our catalog 23 craters exhibiting small ring-like clusters of peaks (50-205 km, geometric mean = 81 km); one (Humboldt) exhibits a rim-crest diameter and an interior morphology that may be uniquely transitional to the process of forming peak rings. Comparisons of the predictions of models for the formation of peak-ring basins with the characteristics of the new basin catalog for the Moon suggest that formation and modification of an interior melt cavity and nonlinear scaling of impact melt volume with crater diameter provide important controls on the development of peak rings. In particular, a power-law model of growth of an interior melt cavity with increasing crater diameter is consistent with power-law fits to the peak-ring basin data for the Moon and Mercury. We suggest that the relationship between the depth of melting and depth of the transient cavity offers a plausible control on the onset diameter and subsequent development of peak-ring basins and also multi-ring basins, which is consistent with both planetary gravitational acceleration and mean impact velocity being important in determining the onset of basin morphological forms on the terrestrial planets.

Baker, David M. H.↗

The Caltech-NRAO Stripe 82 Survey (CNSS) Paper. I. The Pilot Radio Transient Survey in 50 Deg.(exp. 2)

We have commenced a multiyear program, the Caltech-NRAO Stripe 82 Survey (CNSS), to search for radio transients with the Jansky VLA in the Sloan Digital Sky Survey Stripe 82 region. The CNSS will deliver five epochs over the entire approx. 270 deg.(exp. 2) of Stripe 82, an eventual deep combined map with an rms noise of approx. 40 proper motion epoch y and catalogs at a frequency of 3 GHz, and having a spatial resolution of 3 inches. This first paper presents the results from an initial pilot survey of a 50 deg.(exp. 2) region of Stripe 82, involving four epochs spanning logarithmic timescales between 1 week and 1.5 yr, with the combined map having a median rms noise of 35 proper motion epoch y. This pilot survey enabled the development of the hardware and software for rapid data processing, as well as transient detection and follow-up, necessary for the full 270 deg.(exp. 2) survey. Data editing, calibration, imaging, source extraction, cataloging, and transient identification were completed in a semi-automated fashion within 6 hr of completion of each epoch of observations, using dedicated computational hardware at the NRAO in Socorro and custom-developed data reduction and transient detection pipelines. Classification of variable and transient sources relied heavily on the wealth of multiwavelength legacy survey data in the Stripe 82 region, supplemented by repeated mapping of the region by the Palomar Transient Factory. A total of 3.9(+0.5%/-0.9%) of the few thousand detected point sources werefound to vary by greater than 30%, consistent with similar studies at 1.4 and 5 GHz. Multiwavelength photometric data and light curves suggest that the variability is mostly due to shock-induced flaring in the jets of active galactic nuclei (AGNs). Although this was only a pilot survey, we detected two bona fide transients, associated with an RS CVn binary and a dKe star. Comparison with existing legacy survey data (FIRST, VLA-Stripe 82) revealed additional highly variable and transient sources on timescales between 5 and 20 yr, largely associated with renewed AGN activity. The rates of such AGNs possibly imply episodes of enhanced accretion and jet activity occurring once every approx. 40,000 yr in these galaxies. We compile the revised radio transient rates and make recommendations for future transient surveys and joint radio-optical experiments.

galaxies: active – radio continuum: galaxies –↗