Search

This study applied machine learning models to estimate stellar rotation periods from corrected light curve data obtained by the NASA Kepler mission. Traditional methods often struggle to estimate rotation periods accurately due to noise and variability in the light curve data. The workflow involved using initial period estimates from the LS-Periodogram and Transit Least Squares techniques, followed by splitting the data into training, validation, and testing sets. We employed several machine learning algorithms, including Decision Tree, Random Forest, K-Nearest Neighbors, and Gradient Boosting, and also utilized a Voting Ensemble approach to improve prediction accuracy and robustness. The analysis included data from multiple Kepler IDs, providing detailed metrics on orbital periods and planet radii. Performance evaluation showed that the Voting Ensemble model yielded the most accurate results, with an RMSE approximately 50\% lower than the Decision Tree model and 17\% better than the K-Nearest Neighbors model. The Random Forest model performed comparably to the Voting Ensemble, indicating high accuracy. In contrast, the Gradient Boosting model exhibited a worse RMSE compared to the other approaches. Comparisons of the predicted rotation periods to the photometric reference periods showed close alignment, suggesting the machine learning models achieved high prediction accuracy. The results indicate that machine learning, particularly ensemble methods, can effectively solve the problem of accurately estimating stellar rotation periods, with significant implications for advancing the study of exoplanets and stellar astrophysics., Comment: 15 pages, 8 figures. Submitted for publication in A&A

Raw light curve data from exoplanet transits is too complex to naively apply traditional outlier detection methods. We propose an architecture which estimates a latent representation of both the main transit and residual deviations with a pair of variational autoencoders. We show, using two fabricated datasets, that our latent representations of anomalous transit residuals are significantly more amenable to outlier detection than raw data or the latent representation of a traditional variational autoencoder. We then apply our method to real exoplanet transit data. Our study is the first which automatically identifies anomalous exoplanet transit light curves. We additionally release three first-of-their-kind datasets to enable further research., Comment: 12 pages, 4 figures, 4 tables, Accepted at NeurIPS 2021 (Workshop for Machine Learning and the Physical Sciences)

We present far-infrared and submillimeter maps from the Herschel Space Observatory and the James Clerk Maxwell Telescope of the debris disk host star AU Microscopii. Disk emission is detected at 70, 160, 250, 350, 450, 500 and 850 micron. The disk is resolved at 70, 160 and 450 micron. In addition to the planetesimal belt, we detect thermal emission from AU Mic's halo for the first time. In contrast to the scattered light images, no asymmetries are evident in the disk. The fractional luminosity of the disk is $3.9 \times 10^{-4}$ and its mm-grain dust mass is 0.01 MEarth (+/- 20%). We create a simple spatial model that reconciles the disk SED as a blackbody of 53 +/- 2 K (a composite of 39 and 50 K components) and the presence of small (non-blackbody) grains which populate the extended halo. The best fit model is consistent with the "birth ring" model explored in earlier works, i.e., an edge-on dust belt extending from 8.8-40 AU, but with an additional halo component with an $r^{-1.5}$ surface density profile extending to the limits of sensitivity (140 AU). We confirm that AU Mic does not exert enough radiation force to blow out grains. For stellar mass loss rates of 10-100x solar, compact (zero porosity) grains can only be removed if they are very small, consistently with previous work, if the porosity is 0.9, then grains approaching 0.1 micron can be removed via corpuscular forces (i.e., the stellar wind)., Comment: accepted to the Astrophysical Journal

We examine far-infrared and submillimeter spectral energy distributions for galaxies in the Infrared Space Observatory Atlas of Bright Spiral Galaxies. For the 71 galaxies where we had complete 60-180 micron data, we fit blackbodies with lambda^-1 emissivities and average temperatures of 31 K or lambda^-2 emissivities and average temperatures of 22 K. Except for high temperatures determined in some early-type galaxies, the temperatures show no dependence on any galaxy characteristic. For the 60-850 micron range in eight galaxies, we fit blackbodies with lambda^-1, lambda-2, and lambda^-beta (with beta variable) emissivities to the data. The best results were with the lambda^-beta emissivities, where the temperatures were ~30 K and the emissivity coefficient beta ranged from 0.9 to 1.9. These results produced gas to dust ratios that ranged from 150 to 580, which were consistent with the ratio for the Milky Way and which exhibited relatively little dispersion compared to fits with fixed emissivities., Comment: AJ, 2003, in press

We investigate star formation along the Hubble sequence using the ISO Atlas of Spiral Galaxies. Using mid-infrared and far-infrared flux densities normalized by K-band flux densities as indicators of recent star formation, we find several trends. First, star formation activity is stronger in late-type (Sc - Scd) spirals than in early-type (Sa - Sab) spirals. This trend is seen both in nuclear and disk activity. These results confirm several previous optical studies of star formation along the Hubble sequence but conflict with the conclusions of most of the previous studies using IRAS data, and we discuss why this might be so. Second, star formation is significantly more extended in later-type spirals than in early-type spirals. We suggest that these trends in star formation are a result of differences in the gas content and its distribution along the Hubble sequence, and it is these differences that promote star formation in late-type spiral galaxies. We also search for trends in nuclear star formation related to the presence of a bar or nuclear activity. The nuclear star formation activity is not significantly different between barred and unbarred galaxies. We do find that star formation activity appears to be inhibited in LINERs and transition objects compared to HII galaxies. The mean star formation rate in the sample is 1.4 Msun/yr based on global far-infrared fluxes. Combining these data with CO data gives a mean gas consumption time of 6.4 x 10^8 yr, which is ~5 times lower than the values found in other studies. Finally, we find excellent support for the Schmidt Law in the correlation between molecular gas masses and recent star formation in this sample of spiral galaxies., Comment: Accepted by The Astronomical Journal, 20 pages, 14 figures