Showing posts with label data mining. Show all posts
Showing posts with label data mining. Show all posts

Wednesday, January 7, 2015

Sunday, June 22, 2014

A Method for Analyzing Weak Orbital Signals From Gaia

Analysing weak orbital signals in Gaia data

Author:

Lucy

Abstract:

Anomalous orbits are found when minimum-chi^{2} estimation is applied to synthetic Gaia data for weak orbital signals - i.e., orbits whose astrometric signatures are comparable to the single-scan measurement error (Pourbaix 2002). These orbits are nearly parabolic, edge-on, and their major axes align with the line-of-sight to the observer. Such orbits violate the Copernican principle (CPr) and as such could be rejected. However, the preferred alternative is to develop a statistical technique that incorporates the CPr as a fundamental postulate. This can be achieved in the context of Bayesian estimation by defining a Copernican prior. With this development, Pourbaix's anomalous orbits no longer arise. Instead, orbits with a somewhat higher chi^{2} but which do not violate the CPr are selected.

Other areas of astronomy where the investigator must analyse data from 'imperfect experiments' might similarly benefit from appropriately- defined Copernican priors.

Thursday, May 29, 2014

SuperWASP Provides Evidence of Circumbinary Exoplanets

Period and amplitude variations in post-common-envelope eclipsing binaries observed with SuperWASP

Authors:

Lohr et al

Abstract:

Period or amplitude variations in eclipsing binaries may reveal the presence of additional massive bodies in the system, such as circumbinary planets. Here, we have studied twelve previously-known eclipsing post-common-envelope binaries for evidence of such light curve variations, on the basis of multi-year observations in the SuperWASP archive. The results for HW Vir provided strong evidence for period changes consistent with those measured by previous studies, and help support a two-planet model for the system. ASAS J102322-3737.0 exhibited plausible evidence for a period increase not previously suggested; while NY Vir, QS Vir and NSVS 14256825 afforded less significant support for period change, providing some confirmation to earlier claims. In other cases, period change was not convincingly observed; for AA Dor and NSVS 07826147, previous findings of constant period were confirmed. This study allows us to present hundreds of new primary eclipse timings for these systems, and further demonstrates the value of wide-field high-cadence surveys like SuperWASP for the investigation of variable stars.

Friday, February 7, 2014

Data Mining WISE Images Produces Over 22,000 Brown Dwarf Candidates

The AllWISE Motion Survey and The Quest for Cold Subdwarfs

Authors:

Kirkpatrick et al

Abstract:

The AllWISE processing pipeline has measured motions for all objects detected on WISE images taken between 2010 January and 2011 February. In this paper, we discuss new capabilities made to the software pipeline in order to make motion measurements possible, and we characterize the resulting data products for use by future researchers. Using a stringent set of selection criteria, we find 22,445 objects that have significant AllWISE motions, of which 3,525 have motions that can be independently confirmed from earlier 2MASS images yet lack any published motions in SIMBAD. Another 58 sources lack 2MASS counterparts and are presented as motion candidates only. Limited spectroscopic follow-up of this list has already revealed eight new L subdwarfs. These may provide the first hints of a "subdwarf gap" at mid-L types that would indicate the break between the stellar and substellar populations at low metallicities (i.e., old ages). Another object in the motion list -- WISEA J154045.67-510139.3 -- is a bright (J ~ 9 mag) object of type M6; both the spectrophotometric distance and a crude preliminary parallax place it ~6 pc from the Sun. We also compare our list of motion objects to the recently published list of 762 WISE motion objects from Luhman (2014). While these first large motion studies with WISE data have been very successful in revealing previously overlooked nearby dwarfs, both studies missed objects that the other found, demonstrating that many other nearby objects likely await discovery in the AllWISE data products.