OAI/Analysis/LipschitzHilbert/Estimates/Hellinger.lean
6 declarations, 0 contain sorry, 0 declared as axioms
The finite probability estimate in the phase argument
This is the Hellinger-affinity calculation in §7, proof of phase:quantile-gap,
of A uniform Hilbert transform estimate for Lipschitz directions.
The finite probability calculation is proved directly.
- def
OAI.LipschitzHilbert.affinityline 21Source documentation
Hellinger affinity of finite probability weights.
noncomputable def affinity {ι : Type*} [Fintype ι] (p q : ι → ℝ) : ℝ := … - theorem
OAI.LipschitzHilbert.l1_sq_le_affinityline 25Source documentation
Squared total variation estimate, before division of the
ℓ¹distance by two.theorem l1_sq_le_affinity {ι : Type*} [Fintype ι] (p q : ι → ℝ) (hp : ∀ i, 0 ≤ p i) (hq : ∀ i, 0 ≤ q i) (hp1 : ∑ i, p i = 1) (hq1 : ∑ i, q i = 1) : (∑ i, |p i - q i|) ^ 2 ≤ 4 * (1 - affinity p q ^ 2) := by …Mentions: OAI.LipschitzHilbert.affinity
- theorem
OAI.LipschitzHilbert.event_deficit_le_l1_halfline 64Source documentation
Each event deficit is at most the total variation distance.
theorem event_deficit_le_l1_half {ι : Type*} [Fintype ι] (p q : ι → ℝ) (hp1 : ∑ i, p i = 1) (hq1 : ∑ i, q i = 1) (s : Finset ι) : 2 * |∑ i ∈ s, (p i - q i)| ≤ ∑ i, |p i - q i| := by … - theorem
OAI.LipschitzHilbert.event_deficit_sq_le_one_sub_affinity_sqline 84Source documentation
The event estimate used to turn a quantile deficit into a fourth-moment gap.
theorem event_deficit_sq_le_one_sub_affinity_sq {ι : Type*} [Fintype ι] (p q : ι → ℝ) (hp : ∀ i, 0 ≤ p i) (hq : ∀ i, 0 ≤ q i) (hp1 : ∑ i, p i = 1) (hq1 : ∑ i, q i = 1) (s : Finset ι) : (∑ i ∈ s, (p i - q i)) ^ 2 ≤ 1 - affinity p q ^ 2 := by …Mentions: OAI.LipschitzHilbert.affinity
- theorem
OAI.LipschitzHilbert.subprobability_event_deficit_sqline 98Source documentation
The completed triple law may have mass less than one. Adjoining its missing mass at one point, where the independent law has mass zero, preserves affinity.
theorem subprobability_event_deficit_sq {ι : Type*} [Fintype ι] (p q : ι → ℝ) (hp : ∀ i, 0 ≤ p i) (hq : ∀ i, 0 ≤ q i) (hp1 : ∑ i, p i ≤ 1) (hq1 : ∑ i, q i = 1) (s : Finset ι) : (∑ i ∈ s, (q i - p i)) ^ 2 ≤ 1 - affinity p q ^ 2 := by …Mentions: OAI.LipschitzHilbert.affinity
- theorem
OAI.LipschitzHilbert.strict_gap_of_event_deficitline 132Source documentation
A definite event deficit forces the strict fourth-moment gain. This is the last algebraic step of the manuscript's quantile-gap argument.
theorem strict_gap_of_event_deficit {ι : Type*} [Fintype ι] (p q : ι → ℝ) (hp : ∀ i, 0 ≤ p i) (hq : ∀ i, 0 ≤ q i) (hp1 : ∑ i, p i ≤ 1) (hq1 : ∑ i, q i = 1) (s : Finset ι) {Q g : ℝ} (hQ : 0 ≤ Q) (hg : 0 ≤ g) (hgain : Q ^ 2 ≤ affinity p q) (hdeficit : g ≤ ∑ i ∈ s, (q i - p i)) : Q ≤ 1 - g ^ 2 / 4 := by …Mentions: OAI.LipschitzHilbert.affinity