Search NASA⌕ Search

SEARCH · Search NASA

Results for “regression to the mean”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
NASA NTRS1994

Modeled soft X-ray solar irradiances

Solar soft X-rays have historically been inaccurately modeled in both relative variations and absolute magnitudes by empirical solar extreme ultraviolet (EUV) irradiance models. This is a result of the use of a limited number of rocket data sets which were primarily associated with the calibration of the AE-E satellite EUV data set. In this work, the EUV91 solar EUV irradiance model has been upgraded to improve the accuracy of the 3.0 to 5.0 nm relative irradiance variations. The absolute magnitude estimate of the flux in this wavelength range has also been revised upwards. The upgrade was accomplished by first digitizing the SOLRAD 11 satellite 4.4 to 6.0 nm measured energy flux data set, then extracting and extrapolating a derived 3.0 to 5.0 nm photon flux from these data, and finally by performing a correlation between these derived data and the daily and 81-day mean 10.7 cm radio flux emission using a multiple linear regression technique. A correlation coefficient of greater than 0.9 was obtained between the dependent and independent data sets. The derived and modeled 3.0 to 5.0 nm flux varies by more than an order of magnitude over a solar cycle, ranging from a flux below 1 x 10(exp 8) to a flux greater than 1 x 10(exp 9) photons per sq cm per sec. Solar rotational (27-day) variations in the flux magnitude are a factor of 2. The derived and modeled irradiance absolute values are an order of magnitude greater than previous values from rocket data sets related to the calibration of the AE-E satellite.

Tobiska, W. Kent↗
NASA NTRS2006

A New Algorithm for Retrieving Aerosol Properties Over Land from MODIS Spectral Reflectance

Since first light in early 2000, operational global quantitative retrievals of aerosol properties over land have been made from MODIS observed spectral reflectance. These products have been continuously evaluated and validated, and opportunities for improvements have been noted. We have replaced the original algorithm by improving surface reflectance assumptions, the aerosol model optical properties and the radiative transfer code used to create the lookup tables. The new algorithm (known as Version 5.2 or V5.2) performs a simultaneous inversion of two visible (0.47 and 0.66 micron) and one shortwave-IR (2.12 micron) channel, making use of the coarse aerosol information content contained in the 2.12 micron channel. Inversion of the three channels yields three nearly independent parameters, the aerosol optical depth (tau) at 0.55 micron, the non-dust or fine weighting (eta) and the surface reflectance at 2.12 micron. Finally, retrievals of small magnitude negative tau values (down to -0.05) are considered valid, thus normalizing the statistics of tau in near zero tau conditions. On a 'test bed' of 6300 granules from Terra and Aqua, the products from V5.2 show marked improvement over those from the previous versions, including much improved retrievals of tau, where the MODIS/AERONET tau (at 0.55 micron) regression has an equation of: y = 1.01+0.03, R = 0.90. Mean tau for the test bed is reduced from 0.28 to 0.21.

Levy, Robert C.↗
NASA NTRS2011

CO2 Washout Testing of the REI and EM-ACES Space Suits

Requirements for using a space suit during ground testing include providing adequate carbon dioxide (CO2) washout for the suited subject. Acute CO2 exposure can lead to symptoms including headache, dyspnea, lethargy and eventually unconsciousness or even death. Symptoms depend on several factors including partial pressure of CO2 (ppCO2), duration of exposure, metabolic rate of the subject and physiological differences between subjects. The objective of this test was to characterize inspired oronasal ppCO2 in the Rear Entry I-Suit (REI) and the Enhanced Mobility Advanced Crew Escape Suit (EM-ACES) across a range of workloads and flow rates for which ground testing is nominally performed. Three subjects were tested in each suit. In all but one case, each subject performed the test twice to allow for comparison between tests. Suit pressure was maintained at 4.3 psid. Subjects wore the suit while resting, performing arm ergometry, and walking on a treadmill to generate metabolic workloads of approximately 500 to 3000 BTU/hr. Supply airflow was varied at 6, 5 and 4 actual cubic feet per minute (ACFM) at each workload. Subjects wore an oronasal mask with an open port in front of the mouth and were allowed to breathe freely. Oronasal ppCO2 was monitored real-time via gas analyzers with sampling tubes connected to the oronasal mask. Metabolic rate was calculated from the total CO2 production measured by an additional gas analyzer at the air outlet from the suit. Real-time metabolic rate was used to adjust the arm ergometer or treadmill workload to meet target metabolic rates. In both suits, inspired CO2 was primarily affected by the metabolic rate of the subject, with increased metabolic rate resulting in increased inspired ppCO2. Suit flow rate also affected inspired ppCO2, with decreased flow causing small increases in inspired ppCO2. The effect of flow was more evident at metabolic rates greater than or equal to 2000 BTU/hr. Results were consistent between suits, with the EM-ACES demonstrating slightly better CO2 washout than the REI suit, but not statistically significant. Regression equations were developed for each suit to predict the mean inspired ppCO2 as a function of metabolic rate and suit flow rate. This paper provides detailed descriptions of the test hardware, methodology and results, as well as implications for future ground testing in the REI and EM-ACES.

Mitchell, Kate↗
NASA NTRS2012

CO2 Washout Testing of the REI and EM-ACES Space Suits

When a space suit is used during ground testing, adequate carbon dioxide (CO2) washout must be provided for the suited subject. Symptoms of acute CO2 exposure depend on partial pressure of CO2 (ppCO2), metabolic rate of the subject, and other factors. This test was done to characterize inspired oronasal ppCO2 in the Rear Entry I-Suit (REI) and the Enhanced Mobility Advanced Crew Escape Suit (EM-ACES) for a range of workloads and flow rates for which ground testing is nominally performed. Three subjects were tested in each suit. In all but one case, each subject performed the test twice. Suit pressure was maintained at 4.3 psid. Subjects wore the suit while resting, performing arm ergometry, and walking on a treadmill to generate metabolic workloads of about 500 to 3000 BTU/hr. Supply airflow was varied between 6, 5, and 4 actual cubic feet per minute (ACFM) at each workload. Subjects wore an oronasal mask with an open port in front of the mouth and were allowed to breathe freely. Oronasal ppCO2 was monitored in real time by gas analyzers with sampling tubes connected to the mask. Metabolic rate was calculated from the total CO2 production measured by an additional gas analyzer at the suit air outlet. Real-time metabolic rate was used to adjust the arm ergometer or treadmill workload to meet target metabolic rates. In both suits, inspired CO2 was affected mainly by the metabolic rate of the subject: increased metabolic rate significantly (P < 0.05) increased inspired ppCO2. Decreased air flow caused small increases in inspired ppCO2. The effect of flow was more evident at metabolic rates . 2000 BTU/hr. CO2 washout values of the EM-ACES were slightly but not significantly better than those of the REI suit. Regression equations were developed for each suit to predict the mean inspired ppCO2 as a function of metabolic rate and suit flow rate. This paper provides detailed descriptions of the test hardware, methodology, and results as well as implications for future ground testing in the REI-suit and EM-ACES.

Mitchell, Kathryn C.↗
NASA NTRS2020

Modeling Whitecaps on Global Scale

Whitecaps play an important role in the surface-atmosphere interactions across the ocean. They are directly linked to the energy dissipation rate during wave breaking and transfer of heat, momentum, and gas/aerosol exchange from the wind to the sea surface. Although the first models of W were dependent only on wind speeds, a large number of diverse models based on wind and sea state which include wave height, wave age, friction velocity, and stability effects have been proposed since then. However, it is recognized that most of the proposed W models have strong systematic (e.g., zonal bias) and random errors when compared against observations. This is partly due to the differences in environmental conditions, measurement techniques, and geographical locations among these studies. But, some of these biases are linked to the inability of the proposed models to capture the variability in W in certain wind/wave regimes. Despite the knowledge of existing biases, W residual relationships from the models with wind and wave fields remain highly uncertain, with residual trends varying between the published studies. Here, we take advantage of the availability of relatively dense observations of W from WindSat microwave satellite retrievals in combination with the University of Miami wave model which was recently incorporated within the NASA GMAO/GEOS system (GEOS-UMWM). We use Windsat W retrievals to assess and constrain the previously published W models and understand the relationships of residuals from models in different wind/wave regimes. We link these unexplained residual variations to additional factors such as swell index, drag coefficient etc and add information to the existing whitecap models. Since Windsat retrievals cover wide range of environmental conditions, it helps to reduce the uncertainties associated with differences in measurement techniques. Regression of wind-wave fields against all Windsat data points (CTL) results in larger residuals for lower wave age and W is overestimated upto ~4% for wave age < 10 and underestimated by upto ~2% as wave age increases. We attest to this bias by considering two approaches. One is to perform regression separately for different stages of wave development such as developing sea, fully developed, and wind sea regimes thereby understanding the sensitivity of regression coefficients to sea state (EXP1). Another is to derive coefficients of W models in EXP1 as a function of additional wind/wave factors such as swell index, drag coefficient, and mean squared slope, deriving more nonlinear W models (EXP2). EXP2 provides reduction in Root Mean Squared Error (RMSE) by 0.1-0.3%. Sea surface drag has a stronger relationship with regression coefficients compared to swell index.These additional factors provide improved parameterizations in different wind and wave age regimes, with smaller unexplained/residual variations in W that has been a major concern in the W community.

Raman, Aishwarya↗
NASA NTRSDate not supplied

Comparison Study of Machine Learning Techniques to Predict Flight Energy Consumption for Advanced Air Mobility

This paper addresses the need to predict the flight energy consumption of aerial vehicles in the presence of wind using machine learning techniques. The presented work is critical to achieving sustainable and efficient operations for Advanced Air Mobility (AAM) and to evaluating the readiness of the ground-supporting energy infrastructure, e.g., electric grid and AAM portals. The flight energy consumption is described using the "energy per meter" (EPM) metric. We present a comparison study of influential machine learning techniques in predicting EPM using real-world flight test data. We presented new results of using the Decision Tree, Random Forest, and linear regression techniques, along with our previous results using the Recurrent Neural Network and Feed Forward Neural Network techniques. The comparison results show that the Linear Regression method outperforms other methods on the basis of the Mean Squared Error and error variance.

Machine Learning↗
DOE OSTI2023

Robust global detection of forced changes in mean and extreme precipitation despite observational disagreement on the magnitude of change

Detection and attribution (D&A) of forced precipitation change are challenging due to internal variability, limited spatial, and temporal coverage of observational records and model uncertainty. These factors result in a low signal-to-noise ratio of potential regional and even global trends. Here, we use a statistical method – ridge regression – to create physically interpretable fingerprints for the detection of forced changes in mean and extreme precipitation with a high signal-to-noise ratio. The fingerprints are constructed using Coupled Model Intercomparison Project phase 6 (CMIP6) multi-model output masked to match coverage of three gridded precipitation observational datasets – GHCNDEX, HadEX3, and GPCC – and are then applied to these observational datasets to assess the degree of forced change detectable in the real-world climate in the period 1951–2020. We show that the signature of forced change is detected in all three observational datasets for global metrics of mean and extreme precipitation. Forced changes are still detectable from changes in the spatial patterns of precipitation even if the global mean trend is removed from the data. This shows the detection of forced change in mean and extreme precipitation beyond a global mean trend is robust and increases confidence in the detection method's power as well as in climate models' ability to capture the relevant processes that contribute to large-scale patterns of change. We also find, however, that detectability depends on the observational dataset used. Not only coverage differences but also observational uncertainty contribute to dataset disagreement, exemplified by the times of emergence of forced change from internal variability ranging from 1998 to 2004 among datasets. Furthermore, different choices for the period over which the forced trend is computed result in different levels of agreement between observations and model projections. These sensitivities may explain apparent contradictions in recent studies on whether models under- or overestimate the observed forced increase in mean and extreme precipitation. Lastly, the detection fingerprints are found to rely primarily on the signal in the extratropical Northern Hemisphere, which is at least partly due to observational coverage but potentially also due to the presence of a more robust signal in the Northern Hemisphere in general.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2012

Can Selforganizing Maps Accurately Predict Photometric Redshifts?

We present an unsupervised machine-learning approach that can be employed for estimating photometric redshifts. The proposed method is based on a vector quantization called the self-organizing-map (SOM) approach. A variety of photometrically derived input values were utilized from the Sloan Digital Sky Survey's main galaxy sample, luminous red galaxy, and quasar samples, along with the PHAT0 data set from the Photo-z Accuracy Testing project. Regression results obtained with this new approach were evaluated in terms of root-mean-square error (RMSE) to estimate the accuracy of the photometric redshift estimates. The results demonstrate competitive RMSE and outlier percentages when compared with several other popular approaches, such as artificial neural networks and Gaussian process regression. SOM RMSE results (using delta(z) = z(sub phot) - z(sub spec)) are 0.023 for the main galaxy sample, 0.027 for the luminous red galaxy sample, 0.418 for quasars, and 0.022 for PHAT0 synthetic data. The results demonstrate that there are nonunique solutions for estimating SOM RMSEs. Further research is needed in order to find more robust estimation techniques using SOMs, but the results herein are a positive indication of their capabilities when compared with other well-known methods

Way, Michael J.↗
NASA NTRS2017

Seasonal Clear-Sky Flux and Cloud Radiative Effect Anomalies in the Arctic Atmospheric Column Associated with the Arctic Oscillation and Arctic Dipole

The impact of the Arctic Oscillation (AO) and Arctic Dipole (AD) on the radiative flux into the Arctic mean atmospheric column is quantified. 3-month-averaged AO and AD indices are regressed with corresponding surface and top-of-atmosphere (TOA) fluxes from the CERES-SFC and CERES-TOA EBAF datasets over the period 2000-2014. An increase in clear-sky fluxes into the Arctic mean atmospheric column during fall is the largest net flux anomaly associated with AO, primarily driven by a positive net longwave flux anomaly (i.e. increase of net flux into the atmospheric column) at the surface. A decrease in the Arctic mean atmospheric column cloud radiative effect during winter and spring is the largest flux anomaly associated with AD, primarily driven by a change in the longwave cloud radiative effect at the surface. These prominent responses to AO and AD are widely distributed across the ice-covered Arctic, suggesting that the physical process or processes that bring about the flux change associated with AO and AD are distributed throughout the Arctic.

Hegyi, Bradley M.↗
NASA NTRS1983

An allpass filter design for luminance and chrominance separation of NTSC signals

Abstract- Digital techniques for separation of luminance and chrominance components from composite NTSC broadcast television signals are explored. First, comb filters are introduced. Secondly, filters based on linear regression are developed. Finally, a third filtering form, the allpass filter is considered. Techniques are compared based on mean squared error criteria and subjective testing. It is demonstrated that the allpass filtering technique is superior to both the comb filtering as well as the regression filtering.

Border, D. A.↗
DOE OSTI2025

Computer-aided design of stability enhanced nicotinamide cofactor biomimetics for cell-free biocatalysis

Cell-free biocatalysis (CFB) is an efficient and environmentally friendly method to synthesize molecules such as pharmaceuticals, biochemicals, and biofuels through the in vitro use of enzyme cascades. These enzymes often require redox cofactors to drive chemical reactions. Natural redox cofactors (NAD(P)H) are expensive to isolate, motivating synthetic nicotinamide cofactor biomimetics (NCBs) as a cost-effective solution. A select handful of NCBs have been identified as potential NAD(P)H alternatives with comparable or improved redox capabilities, however, they display a tendency to degrade in common buffers. In this study, a library of 132 NCB candidates is systematically generated, over 85% of which have not been characterized in the literature, to expand the diversity of currently explored NCBs. The decomposition mechanism of NCBs in phosphate is evaluated using density functional theory (DFT), revealing protonation at the nicotinamide C5 position as a reporter of cofactor stability. Based on this result, we trained a linear regression model on DFT calculated descriptors to predict NCB stability in phosphate buffer, achieving mean absolute error (MAE) and root mean squared error (RMSE) values within computational accuracy. Analysis of key atomic descriptors and qualitative trends in our dataset informed the design of novel NCB candidates we propose with optimized stability. This work enables researchers to predict the relative stability of NCBs before synthesis, thereby streamlining the process to make CFB more affordable and viable at industry scales.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRS2012

Lower Limb Venous Compliance is Different Between Men and Women Following 60 Days of Head-Down Bedrest but Is Not Associated with Venoconstriction Dysfunction

Space flight-induced orthostatic intolerance (OI) is more prevalent in female (F) than male (M) astronauts. The mechanisms explaining the higher incidence of OI in F are unclear. We tested the hypothesis that venous compliance would be higher in F more than M following 6 deg head-down bed rest (BR) and would be associated with constrictor dysfunction. Using 2-D ultrasound, dorsal hand (DHV) and dorsal foot (DFV) vein compliances were determined in 24 subjects (10 F, 14 M; 35 +/- 1 yr) by measuring mean diameter response to increasing congestion pressure (0, 20, 30, and 40 mmHg) before and after 60 d of BR. Constrictor function was assessed by intravenous infusions of Ketorolac (KE; 1.5 ig/min) Phenylephrine (PE; 3160 ng/min), and L-NMMA (50 ig/min). The effects of BR between F vs. M and hand vs. foot were determined using mixed-effects linear regression. DFV but not DHV compliance changed in response to BR (p=0.012). Mean DFV increased significantly (0.903 mm to 1.191mm) in F but decreased in M (1.353 mm to 1.154 mm). DFV constrictor response was not different between sexes in response to BR (KE; p=0.647, PE; p=0.717, and L-NMMA; p=0.825). These BR data suggest that the higher incidence of OI in F astronauts may be related to increased lower limb venous compliance, contributing to blood pooling upon standing. Notably, changes to DFV compliance was not accompanied by impaired constrictor function.

Westby, Christian M.↗
DOE OSTI2024

Evaluation of the hyperspectral radiometer (HSR1) at the Atmospheric Radiation Measurement (ARM) Southern Great Plains (SGP) site

Abstract. The Peak Design Ltd hyperspectral radiometer (HSR1) was tested at the Atmospheric Radiation Measurement (ARM) user facility Southern Great Plains (SGP) site in Lamont, Oklahoma, for 2 months from May to July 2022. The HSR1 is a prototype instrument that measures total (Ftotal) and diffuse (Fdiffuse) spectral irradiance from 360 to 1100 nm with a spectral resolution of 3 nm. The HSR1 spectral irradiance measurements are compared to nearby collocated spectral radiometers, including two multifilter rotating shadowband radiometers (MFRSRs) and the Shortwave Array Spectroradiometer–Hemispheric (SASHe) radiometer. The Ftotal at 500 nm for the HSR1 compared to the MFRSRs has a mean (relative) difference of 0.01 W m−2 nm−1 (1 %–2 %). The HSR1 mean Fdiffuse at 500 nm is smaller than the MFRSRs' by 0.03–0.04 (10 %) W m−2 nm−1. The HSR1 clear-sky aerosol optical depth (AOD) is also retrieved by considering Langley regressions and compared to collocated instruments such as the Cimel sunphotometer (CSPHOT), MFRSRs, and SASHe. The mean HSR1 AOD at 500 nm is larger than the CSPHOT's by 0.010 (8 %) and larger than the MFRSRs' by 0.007–0.017 (6 %–18 %). In general, good agreement between the HSR1 and other instruments is found in terms of the Ftotal, Fdiffuse, and AODs at 500 nm. The HSR1 quantities are also compared at other wavelengths to the collocated instruments. The comparisons are within ∼ 10 % for the Ftotal and Fdiffuse, except for 940 nm, where there is relatively larger disagreement. The AOD comparisons are within ∼ 10 % at 415 and 440 nm; however, a relatively larger disagreement in the AOD comparison is found for higher wavelengths.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Applying machine learning and quantum chemistry to predict the glass transition temperatures of polymers

Glass transition temperature (T g ) is important for understanding the physical and mechanical properties of a polymer material because it relates to the thermal energy required to transition between a hard glassy state and a soft rubbery one. Over the years, various models have been developed for predicting this thermal property from molecular structure to aid in designing novel polymers in selected classes. This work builds on those efforts by utilizing both machine learning (ML) and quantum chemistry (QC) techniques to develop models that can predict T g values from the molecular structure under different data availability scenarios and for a wide variety of polymer types. For the ML model, a graph convolutional network (GCN) was used to map topological polymer features; this model was trained against a dataset of more than 7500 T g values and resulted in a root mean square error (RMSE) of 38.1 °C. The QC-based regression model was trained on 83 T g values and produced an RMSE of 34.5 °C. In conclusion, this work demonstrated that while both model techniques produce accurate predictions and are suitable for different data availability scenarios, the QC-based regression model offered a more interpretable model framework with significantly less training data.

36 MATERIALS SCIENCE↗
DOE OSTI2020

CMIP5 Climate Models Overestimate Cooling by Volcanic Aerosols

We compare projections of the observed hemispherical mean surface temperature (HadCRUT4.6.0.0) and the ensemble mean of CMIP5 climate models' simulations on a set of standard regression model forcing variables. We find that the volcanic aerosol regression coefficients of the CMIP5 simulations are consistently significantly larger (by 40–49%) than the volcanic aerosol coefficients of the observed temperature. The probability that the observed differences are caused just by chance is much less than 0.01. The overestimate is due to the climate models' response to volcanic aerosol radiative forcing. The largest overestimate occurs in the winter season of each hemisphere. We hypothesize that the models' parameterization of aerosol-cloud interactions within ice and mixed phase clouds is a likely source of this discrepancy. Furthermore, the models significantly underestimate the effect of solar variability on temperature for both hemispheres.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2026

A physically interpretable precursor framework for sub-seasonal prediction of Northern Hemisphere flash flourishing

Flash flourishing describes rapid vegetation increases that can quickly reshape land–atmosphere exchanges and impacts on ecosystem, yet its large-scale precursors, circulation context, and sub-seasonal predictability remain poorly understood. Here, we identified onset-stage circulation regimes across northern extratropical latitudes (NEL; >30°N) using 200 and 1000 hPa geopotential height, and examined their regional expressions over eastern Asia, western North America, and Europe. Flash flourishing onset in East Asian was associated with a baroclinic circulation regime and was preceded by a North Atlantic sea surface temperature (SST) precursor at a four-pentad lead. In contrast, onset in western North American and European preferentially occurred under barotropic regimes, preconditioned by Great Plains soil moisture at three-pentad lead and North Atlantic SST at a four-pentad lead, respectively. Ridge regression forecasts revealed regime-dependent sub-seasonal predictability, with mean out-of-sample R 2 exceeding 0.3 up to lead times of two pentads in East Asia, three pentads in western North America, and four pentads in Europe. Together, these findings established a mechanistic and regionally specific framework for anticipating rapid vegetation greening at sub-seasonal timescales.

Kong, Xiangxu [Nanjing Univ. of Information Scienc↗
DOE OSTI2023

A review on recent machine learning applications for imaging mass spectrometry studies

Imaging mass spectrometry (IMS) is a powerful analytical technique widely used in biology, chemistry, and materials science fields that continue to expand. IMS provides a qualitative compositional analysis and spatial mapping with high chemical specificity. The spatial mapping information can be 2D or 3D depending on the analysis technique employed. Due to the combination of complex mass spectra coupled with spatial information, large high-dimensional datasets (hyperspectral) are often produced. Therefore, the use of automated computational methods for an exploratory analysis is highly beneficial. The fast-paced development of artificial intelligence (AI) and machine learning (ML) tools has received significant attention in recent years. These tools, in principle, can enable the unification of data collection and analysis into a single pipeline to make sampling and analysis decisions on the go. There are various ML approaches that have been applied to IMS data over the last decade. Here, in this review, we discuss recent examples of the common unsupervised (principal component analysis, non-negative matrix factorization, k-means clustering, uniform manifold approximation and projection), supervised (random forest, logistic regression, XGboost, support vector machine), and other methods applied to various IMS datasets in the past five years. The information from this review will be useful for specialists from both IMS and ML fields since it summarizes current and representative studies of computational ML-based exploratory methods for IMS.

47 OTHER INSTRUMENTATION↗
DOE OSTI2024

Neural networks for estimation of divertor conditions in DIII-D using C III imaging

Deep learning approaches have been applied to images of C III emission in the lower divertor of DIII-D to develop models for estimating the level of detachment and magnetic configuration (X-point location and strike point radial location). The poloidal distance from the target to the C III emission front is used to represent the level of detachment. The models perform well on a test dataset not used in training, achieving $F_1$ scores as high as 0.99 for detachment state classification and root mean squared error (RMSE) as low as 2cm for front location regression. Predictions for shots with intermittent reattachment are studied, with class activation mapping used to aid in interpretation of the model predictions. Based on the success of these models, a third model was trained to predict the X-point location and strike point radial position from C III images. Though the dataset covers only a small range of possible magnetic configurations, the model shows promising results, achieving RMSE around 1cm for the test data.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗