Search NASA⌕ Search

SEARCH · Search NASA

Results for “Random forests”

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 973 records · Page 54

Improving Detection of Disease Re-emergence Using a Web-Based Tool (RED Alert): Design and Case Analysis Study

Background: Currently, the identification of infectious disease re-emergence is performed without describing specific quantitative criteria that can be used to identify re-emergence events consistently. This practice may lead to ineffective mitigation. In addition, identification of factors contributing to local disease re-emergence and assessment of global disease re-emergence require access to data about disease incidence and a large number of factors at the local level for the entire world. This paper presents Re-emerging Disease Alert (RED Alert), a web-based tool designed to help public health officials detect and understand infectious disease re-emergence. Objective: Our objective is to bring together a variety of disease-related data and analytics needed to help public health analysts answer the following 3 primary questions for detecting and understanding disease re-emergence: Is there a potential disease re-emergence at the local (country) level? What are the potential contributing factors for this re-emergence? Is there a potential for global re-emergence? Methods: We collected and cleaned disease-related data (eg, case counts, vaccination rates, and indicators related to disease transmission) from several data sources including the World Health Organization (WHO), Pan American Health Organization (PAHO), World Bank, and Gideon. We combined these data with machine learning and visual analytics into a tool called RED Alert to detect re-emergence for the following 4 diseases: measles, cholera, dengue, and yellow fever. We evaluated the performance of the machine learning models for re-emergence detection and reviewed the output of the tool through a number of case studies. Results: Our supervised learning models were able to identify 82%-90% of the local re-emergence events, although with 18%-31% (except 46% for dengue) false positives. This is consistent with our goal of identifying all possible re-emergences while allowing some false positives. The review of the web-based tool through case studies showed that local re-emergence detection was possible and that the tool provided actionable information about potential factors contributing to the local disease re-emergence and trends in global disease re-emergence. Conclusions: To the best of our knowledge, this is the first tool that focuses specifically on disease re-emergence and addresses the important challenges mentioned above.

60 APPLIED LIFE SCIENCES↗

Estimating Forest Vertical Structure from Multialtitude, Fixed-Baseline Radar Interferometric and Polarimetric Data

Parameters describing the vertical structure of forests, for example tree height, height-to-base-of-live-crown, underlying topography, and leaf area density, bear on land-surface, biogeochemical, and climate modeling efforts. Single, fixed-baseline interferometric synthetic aperture radar (INSAR) normalized cross-correlations constitute two observations from which to estimate forest vertical structure parameters: Cross-correlation amplitude and phase. Multialtitude INSAR observations increase the effective number of baselines potentially enabling the estimation of a larger set of vertical-structure parameters. Polarimetry and polarimetric interferometry can further extend the observation set. This paper describes the first acquisition of multialtitude INSAR for the purpose of estimating the parameters describing a vegetated land surface. These data were collected over ponderosa pine in central Oregon near longitude and latitude -121 37 25 and 44 29 56. The JPL interferometric TOPSAR system was flown at the standard 8-km altitude, and also at 4-km and 2-km altitudes, in a race track. A reference line including the above coordinates was maintained at 35 deg for both the north-east heading and the return southwest heading, at all altitudes. In addition to the three altitudes for interferometry, one line was flown with full zero-baseline polarimetry at the 8-km altitude. A preliminary analysis of part of the data collected suggests that they are consistent with one of two physical models describing the vegetation: 1) a single-layer, randomly oriented forest volume with a very strong ground return or 2) a multilayered randomly oriented volume; a homogeneous, single-layer model with no ground return cannot account for the multialtitude correlation amplitudes. Below the inconsistency of the data with a single-layer model is followed by analysis scenarios which include either the ground or a layered structure. The ground returns suggested by this preliminary analysis seem too strong to be plausible, but parameters describing a two-layer compare reasonably well to a field-measured probability distribution of tree heights in the area.

Treuhaft, Robert N.↗

Using high spatial resolution satellite imagery to map forest burn severity across spatial scales in a Pine Barrens ecosystem

As a primary disturbance agent, fire significantly influences local processes and services of forest ecosystems. Although a variety of remote sensing based approaches have been developed and applied to Landsat mission imagery to infer burn severity at 30 m spatial resolution, forest burn severity have still been seldom assessed at fine spatial scales (≤ 5 m) from very-high-resolution (VHR) data. Here we assessed a 432 ha forest fire that occurred in April 2012 on Long Island, New York, within the Pine Barrens region, a unique but imperiled fire-dependent ecosystem in the northeastern United States. The mapping of forest burn severity was explored here at fine spatial scales, for the first time using remotely sensed spectral indices and a set of Multiple Endmember Spectral Mixture Analysis (MESMA) fraction images from bi-temporal — pre- and post-fire event — WorldView-2 (WV-2) imagery at 2 m spatial resolution. We first evaluated our approach using 1 m by 1 m validation points at the sub-crown scale per severity class (i.e. unburned, low, moderate, and high severity) from the post-fire 0.10 m color aerial ortho-photos; then, we validated the burn severity mapping of geo-referenced dominant tree crowns (crown scale) and 15 m by 15 m fixed-area plots (inter-crown scale) with the post-fire 0.10 m aerial ortho-photos and measured crown information of twenty forest inventory plots. Our approach can accurately assess forest burn severity at the sub-crown (overall accuracy is 84% with a Kappa value of 0.77), crown (overall accuracy is 82% with a Kappa value of 0.76), and inter-crown scales (89% of the variation in estimated burn severity ratings (i.e. Geo-Composite Burn Index (CBI)). Lastly, this work highlights that forest burn severity mapping from VHR data can capture heterogeneous fire patterns at fine spatial scales over the large spatial extents. This is important since most ecological processes associated with fire effects vary at the < 30 m scale and VHR approaches could significantly advance our ability to characterize fire effects on forest ecosystems.

54 ENVIRONMENTAL SCIENCES↗

Front Range Wildland Fires: Evaluating the Efficacy of Remote Sensing Imagery in Monitoring Forest Fuels Treatment Methods

Over the last several decades, wildfire frequency and severity in forested areas along Colorado’s Front Range have increased due to a buildup of fuels. This has led to an increase in forest treatments, as well as an increased need to evaluate the success of these treatments. Remote sensing products offer an efficient and cost-effective way to monitor forest treatments; however, not all remote sensing products and analysis techniques have been explored by Coloradan land managers. Specifically, project partners at the Colorado State Forest Service (CSFS) and the Colorado Forest Restoration Institute (CFRI) were interested in using an effective and streamlined method of mapping canopy cover to better monitor forest treatment success. To support their needs, the NASA DEVELOP Front Range Wildland Fires team explored National Agricultural Imagery Program (NAIP) imagery at different spatial resolutions and numbers of training points with NASA’s Shuttle Radar Topography Mission (SRTM) Data Elevation Model (DEM) as a predictor in addition to NAIP imagery spectral predictors. From this analysis, we created classified canopy cover rasters, and compared accuracy metrics across model iterations. We also determined that the best performing model, with an overall accuracy of 0.900 uses 2021 NAIP imagery at 2-meter resolution, 800 training points, 200 testing points, does not use topographic predictors, and reclassifies shadow pixels via a pre-selected NDVI threshold.

Remote Sensing↗

Field evaluation of semi‐automated moisture estimation from geophysics using machine learning

Geophysical methods can provide three-dimensional (3D), spatially continuous estimates of soil moisture. However, point-to-point comparisons of geophysical properties to measure soil moisture data are frequently unsatisfactory, resulting in geophysics being used for qualitative purposes only. This is because (1) geophysics requires models that relate geophysical signals to soil moisture, (2) geophysical methods have potential uncertainties resulting from smoothing and artifacts introduced from processing and inversion, and (3) results from multiple geophysical methods are not easily combined within a single soil moisture estimation framework. To investigate these potential limitations, an irrigation experiment was performed wherein soil moisture was monitored through time, and several surface geophysical datasets indirectly sensitive to soil moisture were collected before and after irrigation: ground penetrating radar, electrical resistivity tomography (ERT), and frequency domain electromagnetics (FDEM). Data were exported in both raw and processed form, and then snapped to a common 3D grid to facilitate moisture prediction by standard calibration techniques, multivariate regression, and machine learning. A combination of inverted ERT data, raw FDEM, and inverted FDEM data was most informative for predicting soil moisture using a random regression forest model (one-thousand 60/40 training/test cross-validation folds produced root mean squared errors ranging from 0.025–0.046 cm 3 /cm 3 ). This cross-validated model was further supported by a separate evaluation using a test set from a physically separate portion of the study area. Machine learning was conducive to a semi-automated model-selection process that could be used for other sites and datasets to locally improve accuracy.

54 ENVIRONMENTAL SCIENCES↗

Integrating Reservoirs into the Dissolved Organic Matter Versus Primary Production Paradigm: How Does Chlorophyll-$a$ Change Across Dissolved Organic Carbon Concentrations in Reservoirs?

Primary production in freshwater ecosystems is largely a function of light and nutrient availability, both of which have been changing in many lakes and reservoirs in response to anthropogenic pressures. Recent studies focusing on natural lakes have found a hump-shaped response of primary production (sometimes measured as chlorophyll-$a$) to dissolved organic matter (DOM, measured as dissolved organic carbon, DOC), which has both light-absorbing chromophoric properties and DOM-bound nutrients. We used the United States National Lakes Assessment dataset to integrate reservoirs into this paradigm in comparison with natural lakes and assessed the relative differences in the predicted response’s model structure, regression parameter values, and drivers of the chlorophyll-$a$ residuals. We found that chlorophyll-$a$ in reservoirs exhibited a hump-shaped response to DOC, while natural lakes from this dataset were better fit with a linear response, differing from previous studies focused on boreal lakes. Despite this, reservoirs had a greater maximum chlorophyll-a response compared to natural lakes in this study (45.5 versus 33.8 μg L -1 ), which occurred at a lower DOC concentration threshold (18.3 versus 26.4 mg L -1 ) when compared using quadratic models. Reservoirs had lower median light:nutrient values compared to natural lakes, and greater median surface area and total phosphorus (TP), that can all influence the light environment and the peak chlorophyll-a responses. In both reservoirs and natural lakes, chlorophyll-$a$ residuals were most strongly influenced by TP, where TP < 25-30 µg L -1 suppressed chlorophyll-a residuals and higher TP amplified them. Light:nutrient values were somewhat important predictors, and patterns with chlorophyll-$a$ residuals supported previous work showing low light:nutrient values amplified chlorophyll-$a$ responses and higher values suppressed them. In conclusion, quantifying the shape of the response of primary production to DOM quantity and quality as well as the drivers of the residuals, namely TP for lakes and reservoirs in this dataset, will be important for understanding the effects that changes in water quality may have on primary production and freshwater ecosystem processes.

59 BASIC BIOLOGICAL SCIENCES↗

Integrating Natural Language Processing and Machine Learning Algorithms to Categorize Oncologic Response in Radiology Reports

A significant volume of medical data remains unstructured. Natural language processing (NLP) and machine learning (ML) techniques have shown to successfully extract insights from radiology reports. However, the codependent effects of NLP and ML in this context have not been well-studied. Between April 1, 2015 and November 1, 2016, 9418 cross-sectional abdomen/pelvis CT and MR examinations containing our internal structured reporting element for cancer were separated into four categories: Progression, Stable Disease, Improvement, or No Cancer. We combined each of three NLP techniques with five ML algorithms to predict the assigned label using the unstructured report text and compared the performance of each combination. The three NLP algorithms included term frequency-inverse document frequency (TF-IDF), term frequency weighting (TF), and 16-bit feature hashing. The ML algorithms included logistic regression (LR), random decision forest (RDF), one-vs-all support vector machine (SVM), one-vs-all Bayes point machine (BPM), and fully connected neural network (NN). The best-performing NLP model consisted of tokenized unigrams and bigrams with TF-IDF. Increasing N-gram length yielded little to no added benefit for most ML algorithms. With all parameters optimized, SVM had the best performance on the test dataset, with 90.6 average accuracy and F score of 0.813. The interplay between ML and NLP algorithms and their effect on interpretation accuracy is complex. The best accuracy is achieved when both algorithms are optimized concurrently.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

What matters the most? Understanding individual tornado preparedness using machine learning

Scholars from various disciplines have long attempted to identify the variables most closely associated with individual preparedness. Therefore, we now have much more knowledge regarding these factors and their association with individual preparedness behaviors. However, it has not been sufficiently discussed how decisive many of these factors are in encouraging preparedness. In this article, we seek to examine what factors, among the many examined in previous studies, are most central to engendering emergency preparedness in individuals particularly for tornadoes by utilizing a relatively uncommon machine learning technique in disaster management literature. Using unique survey data, we find that in the case of tornado preparedness the most decisive variables are related to personal experiences and economic circumstances rather than basic demographics. Our findings contribute to scholarly endeavors to understand and promote individual tornado preparedness behaviors by highlighting the variables most likely to shape tornado preparedness at an individual level.

54 ENVIRONMENTAL SCIENCES↗

A Cost-Benefit Analysis of Automated Physiological Data Acquisition Systems Using Data-Driven Modeling

Precision medicine and the continuous analysis of “Big data” promises to improve patient outcomes dramatically in the near future. Very recently, healthcare facilities have started to explore automatic collection of patient-specific physiological data with the aim of reducing nursing workload and decreasing manual data entry errors. In addition to those purposes, continuous physiological data can be used for the early detection and prevention of common, and possibly fatal, diseases. For instance, poor patient outcomes from sepsis, a leading cause of mortality in healthcare facilities and a major driver of hospital costs in the USA, can be mitigated when detected early using screening tools that monitor the changing dynamics of physiological data. However, the potential cost of collecting continuous physiological data remains a barrier to the widespread adoption of automated high-frequency data collection systems. In this paper, we perform cost-benefit analysis (CBA) of machine learning applied to various types of acquisition systems (with different collection intervals) to determine if the benefits of such systems will outweigh their implementation costs. Although such systems can be used in the detection of various complications, in order to showcase the immediate benefits, we focus on the early detection of sepsis, one of the major challenges of hospital systems. We present a general approach to conduct such analysis for a wide range of hospitals and highlight its applicability using a case study for a small hospital with 150 beds and 3000 annual patients where the acquisition system would collect data at 1-min intervals. Finally, we discuss how the analysis may help guide incentives/policies with regard to adopting automated data acquisition systems.

99 GENERAL AND MISCELLANEOUS↗

The Solar Influencer Next Door: Predicting Low-Income Solar Referrals and Leads

Increasing the adoption of solar among low-to-moderate income (LMI) households remains an important policy goal because of its promise to simultaneously reduce energy burden and support the just distribution of benefits of renewable energy. However, scaling LMI solar remains challenging due to affordability and access issues. Most existing LMI adoption has occurred under public-funded programs, highlighting the importance of increasing the cost-effectiveness of these programs at scale. We develop a new household-level data set on LMI solar lead acquisition, referrals, and adoption to understand the processes through which LMI solar uptake has occurred in California. Then, we develop models to predict two sub-mechanisms in the solar adoption process: whether an otherwise qualified lead becomes "lost" i.e. non-responsive to outreach and, for existing clients, whether they refer solar to others. For the program analyzed, participants received their solar system at no cost, which deemphasizes economic drivers of solar adoption and could differ from other program experiences. Both models substantially improved the accuracy of prediction relative to a baseline. Overall, we find that peer effects and solar economics are important to predicting referrals, and household demographic factors in lead loss prediction. Finally, we find that referrals are both the highest quality and largest source of LMI solar leads, providing a promising mechanism to expand LMI programs further.

customer acquisition costs↗

Comparison of machine learning and electrical resistivity arrays to inverse modeling for locating and characterizing subsurface targets

Here, this study evaluates the performance of multiple machine learning (ML) algorithms and electrical resistivity (ER) arrays for inversion with comparison to a conventional Gauss-Newton numerical inversion method. Four different ML models and four arrays were used for the estimation of only six variables for locating and characterizing hypothetical subsurface targets. The combination of dipole-dipole with Multilayer Perceptron Neural Network (MLP-NN) had the highest accuracy. Evaluation showed that both MLP-NN and Gauss-Newton methods performed well for estimating the matrix resistivity while target resistivity accuracy was lower, and MLP-NN produced sharper contrast at target boundaries for the field and hypothetical data. Both methods exhibited comparable target characterization performance, whereas MLP-NN had increased accuracy compared to Gauss-Newton in prediction of target width and height, which was attributed to numerical smoothing present in the Gauss-Newton approach. MLP-NN was also applied to a field dataset acquired at U.S. DOE Hanford site.

54 ENVIRONMENTAL SCIENCES↗

Intensity of sample processing methods impacts wastewater SARS-CoV-2 whole genome amplicon sequencing outcomes

Wastewater SARS-CoV-2 surveillance has been deployed since the beginning of the COVID-19 pandemic to monitor the dynamics in virus burden in local communities. Genomic surveillance of SARS-CoV-2 in wastewater, particularly efforts aimed at whole genome sequencing for variant tracking and identification, are still challenging due to low target concentration, complex microbial and chemical background, and lack of robust nucleic acid recovery experimental procedures. The intrinsic sample limitations are inherent to wastewater and are thus unavoidable. Here, we use a statistical approach that couples correlation analyses to a random forest-based machine learning algorithm to evaluate potentially important factors associated with wastewater SARS-CoV-2 whole genome amplicon sequencing outcomes, with a specific focus on the breadth of genome coverage. We collected 182 composite and grab wastewater samples from the Chicago area between November 2020 to October 2021. Samples were processed using a mixture of processing methods reflecting different homogenization intensities (HA + Zymo beads, HA + glass beads, and Nanotrap), and were sequenced using one of the two library preparation kits (the Illumina COVIDseq kit and the QIAseq DIRECT kit). Technical factors evaluated using statistical and machine learning approaches include sample types, certain sample intrinsic features, and processing and sequencing methods. The results suggested that sample processing methods could be a predominant factor affecting sequencing outcomes, and library preparation kits was considered a minor factor. Finally, a synthetic SARS-CoV-2 RNA spike-in experiment was performed to validate the impact from processing methods and suggested that the intensity of the processing methods could lead to different RNA fragmentation

60 APPLIED LIFE SCIENCES↗

Machine Learning Prediction of the Experimental Transition Temperature of Fe(II) Spin-Crossover Complexes

Spin-crossover (SCO) complexes are materials that exhibit changes in the spin state in response to external stimuli, with potential applications in molecular electronics. It is challenging to know a priori how to design ligands to achieve the delicate balance of entropic and enthalpic contributions needed to tailor a transition temperature close to room temperature. Here, we leverage the SCO complexes from the previously curated SCO-95 data set [Vennelakanti et al. J. Chem. Phys. 159, 024120 (2023)] to train three machine learning (ML) models for transition temperature (T 1/2 ) prediction using graph-based revised autocorrelations as features. We perform feature selection using random forest-ranked recursive feature addition (RF-RFA) to identify the features essential to model transferability. Of the ML models considered, the full feature set RF and recursive feature addition RF models perform best, achieving moderate correlation to experimental T 1/2 values. We then compare ML T 1/2 predictions to those from three previously identified best-performing density functional approximations (DFAs) which accurately predict SCO behavior across SCO-95, finding that the ML models predict T 1/2 more accurately than the best-performing DFAs. In addition, we study ML model predictions for a set of 18 SCO complexes for which only estimated T 1/2 values are available. Upon excluding outliers from this set, the RF-RFA RF model shows a strong correlation to estimated T 1/2 values with a Pearson’s r of 0.82. In contrast, DFA-predicted T 1/2 values have large errors and show no correlation to estimated T 1/2 values over the same set of complexes. Overall, our study demonstrates slightly superior performance of ML models in comparison with some of the best-performing DFAs, and we expect ML models to improve further as larger data sets of SCO complexes are curated and become available for model training.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Predicting the Activity and Selectivity of Bimetallic Metal Catalysts for Ethanol Reforming using Machine Learning

Machine learning is ideally suited for the pattern detection in large uniform datasets, but consistent experimental datasets on catalyst studies are often small. Here we demonstrate how a combination of machine learning and first-principles calculations can be used to extract knowledge from a relatively small set of experimental data. The approach is based on combining a complex machine-learning model trained on an extensive computational library of transition-state energies with simple linear regression models of experimental catalytic activities and selectivities from the literature. Using the combined model, we identify the key C–C bond scission reactions involved in ethanol reforming and perform a computational screening for ethanol reforming on monolayer bimetallic catalysts with architectures TM-Pt-Pt(111) and Pt-TM-Pt(111) (TM = 3d transition metals). The model also predicts four promising catalyst compositions for future experimental studies. In conclusion, the approach is not limited to ethanol reforming but is of general use for the interpretation of experimental observations as well as for the computational discovery of novel catalytic materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Uncertainty-aware mixed-variable machine learning for materials design

Abstract Data-driven design shows the promise of accelerating materials discovery but is challenging due to the prohibitive cost of searching the vast design space of chemistry, structure, and synthesis methods. Bayesian optimization (BO) employs uncertainty-aware machine learning models to select promising designs to evaluate, hence reducing the cost. However, BO with mixed numerical and categorical variables, which is of particular interest in materials design, has not been well studied. In this work, we survey frequentist and Bayesian approaches to uncertainty quantification of machine learning with mixed variables. We then conduct a systematic comparative study of their performances in BO using a popular representative model from each group, the random forest-based Lolo model (frequentist) and the latent variable Gaussian process model (Bayesian). We examine the efficacy of the two models in the optimization of mathematical functions, as well as properties of structural and functional materials, where we observe performance differences as related to problem dimensionality and complexity. By investigating the machine learning models’ predictive and uncertainty estimation capabilities, we provide interpretations of the observed performance differences. Our results provide practical guidance on choosing between frequentist and Bayesian uncertainty-aware machine learning models for mixed-variable BO in materials design.

36 MATERIALS SCIENCE↗