Exponential measures as Bernstein representing measures #
The exponential probability measure with rate r > 0, transported from ℝ to ℝ≥0, has
Laplace transform
t ↦ r / (r + t).
This gives a continuous, non-atomic example of Bernstein's theorem: at unit rate the function
t ↦ 1 / (1 + t) is represented by the measure with density e⁻ˣ on [0, ∞). Unlike the
Dirac examples, this exercises a genuinely continuous representing measure.
The measure is defined in TauCeti.Probability.Distributions.Exponential.Basic by pushing
Mathlib's ProbabilityTheory.expMeasure forward along Real.toNNReal. A positive-rate
exponential random variable is nonnegative almost surely, so this transport retains the law and
turns its moment-generating-function formula into the required Laplace-transform formula.
Main declarations #
TauCeti.laplaceTransform_nnrealExpMeasure: its Laplace transform isr / (r + t).TauCeti.representsLaplace_nnrealExpMeasure: the resulting Bernstein representation.TauCeti.bernsteinMeasure_one_div_one_add: the canonical Bernstein measure oft ↦ 1 / (1 + t)is the unit-rate exponential measure.
References #
- R. Schilling, R. Song, Z. Vondraček, Bernstein Functions: Theory and Applications, 2nd ed., Example 1.4 and Theorem 1.4.
The Laplace transform of the exponential measure of rate r > 0 is r / (r + t) throughout
its maximal finiteness domain -r < t.
The positive-rate exponential measure represents t ↦ r / (r + t) in Bernstein's theorem.
The canonical Bernstein representing measure of t ↦ r / (r + t), for positive r, is the
exponential measure of rate r on ℝ≥0.
At unit rate, the exponential measure on ℝ≥0 represents t ↦ 1 / (1 + t).
The canonical Bernstein representing measure of t ↦ 1 / (1 + t) is the unit-rate
exponential measure on ℝ≥0.