Astronomy & Celestial Mechanics
the sky, computed
Concept

Gaussian process — where it appears

A statistical model of a signal as a random function whose values at any set of times are jointly Gaussian, with correlations set by a kernel. In radial-velocity work it describes the star's own activity, so that a planet's reflex is fitted against correlated noise rather than white noise.

Named by 2 essays across 2 fields — each of them below, with the objects they name alongside it.

The evidence for a period, one trial period at a time. The natural logarithm of the Bayes factor for a sinusoid at each trial period against no signal, for 120 radial velocities spread over 359 days with 2.0 m/s white noise per point and a planet of semi-amplitude 1.0 m/s at 11.3 days. The amplitude prior is Gaussian with scale 3.0 m/s in each component. At the planet's period the evidence is 14.4 — odds of about 1.8·10⁶ to one — and almost everywhere else it is near zero or negative, because a sinusoid that fits nothing is penalised for the amplitude range it wasted. The quantity that answers the question "is there a planet at some period" is not the peak but the average of the Bayes factor over the prior on the period, here uniform in the logarithm of the period from 1.5 to 1,000 days, and that average is ln B = 7.7: 6.7 lower than the peak, the Bayesian form of the look-elsewhere penalty, paid because the data had to find the period as well as the amplitude.

Odds for a planet that depend on a guess about its size

A false-alarm probability asks how often noise alone would produce a peak this tall. A Bayes factor asks something better — how much more probable the data are with a planet than without one — and pays for it with two numbers nobody measured. The width allowed for the period barely matters. The width allowed for the amplitude can turn strong evidence into none, from the same velocities.

starlight · Periodograms Two models through the same points, one with a planet. 60 simulated velocities over 200 days of a star rotating every 25 days, with spot activity drawn from a quasi-periodic process of 3 m/s amplitude and a 40-day decay, white noise of 1 m/s, and a planet of 2.5 m/s semi-amplitude on a 27.0-day orbit — 1.08 times the rotation period. Two models are fitted, each with its activity amplitude and decay time chosen by maximum likelihood: activity plus the planet (solid), whose planet comes out at 3.3 m/s, and activity alone (dashed), which has no planet in it. Both pass through the points; the likelihoods differ by 3.6 in the logarithm, a ratio of 34.9, which is not decisive. Near the rotation period a flexible model of the star can carry the planet's signal by itself.

A model of the star that can carry the planet

Starspots are modelled as correlated noise with the star's rotation period built in, and the model is good enough to remove signals that would otherwise be announced as planets. It is also good enough to remove planets. Near the rotation period no amount of data in one season separates the two, and away from it a flexible noise model quietly takes most of the evidence a planet would have had.

exoplanets · Reflex velocity

Named alongside it

The objects these essays reach for when they reach for this one.

Radial-velocityBayes factorCorrelated noiseThe Doppler effectFalse-alarm probabilityFalse-positiveLikelihood ratioLindley's paradoxLook elsewhere effectPeriodogramPriorRed noise

All concepts