Physlib

PhyslibAlpha.ProbabilisticTheory.JordanOrderUnit.CStarAlgebra.Statistics

Spectral formulas for Jordan moments

i. Overview

The statistics of an observable `a` in a state `ω` of a C⋆-algebra are given by its spectral measure `μ_{ω,a}`: `ω(f(a)) = ∫ f dμ_{ω,a}`. The Jordan moments of `a` are defined through Jordan powers `a^{[n]}`. For a single observable these agree with the ordinary powers `aⁿ` of the C⋆-algebra, so every moment, and in particular the variance, is an integral against the spectral measure.

ii. Key definitions and results

  • `JB.jpow_eq_pow` : Jordan powers of a single observable are ordinary powers.
  • `JB.moment_eq_integral` : `moment n (ω.onObservables) a = ∫ y, y^n ∂(realSpectralMeasure ω a)`
  • `JB.variance_eq_integral_sq_sub` : the variance as `∫y² dμ - (∫y dμ)²`

iii. Table of contents

  • A. Jordan powers of a single element are ordinary powers
  • B. Moments as integrals against the outcome distribution

A. Jordan powers of a single element are ordinary powers

B. Moments as integrals against the outcome distribution

4 declarations

theorem

Jordan powers of self-adjoint elements equal ordinary powers

Let AA be a star-algebra and a∈selfAdjoint(A)a \in \text{selfAdjoint}(A) be a self-adjoint element. For any natural number n∈Nn \in \mathbb{N}, the nn-th Jordan power of aa, denoted a[n]a^{[n]} and defined recursively by a[0]=1a^{[0]} = 1 and a[n+1]=a∘a[n]a^{[n+1]} = a \circ a^{[n]} (where ∘\circ is the Jordan product), is equal to the ordinary power ana^n in AA.

theorem

nn-th moment equals ∫yn dμω,a\int y^n \, d\mu_{\omega, a}

Let AA be a unital C∗C^*-algebra and ω\omega be a state on AA (a unital positive linear map from AA to C\mathbb{C}). For any self-adjoint element a∈Aa \in A (an observable) and any natural number n∈Nn \in \mathbb{N}, the nn-th moment of aa in the state ω\omega is given by the integral of the nn-th power function with respect to the real spectral measure μω,a\mu_{\omega, a} associated with ω\omega and aa: momentn(ω,a)=∫Ryn dμω,a(y)\text{moment}_n(\omega, a) = \int_{\mathbb{R}} y^n \, d\mu_{\omega, a}(y) Here, the moment is defined by applying the state to the nn-th Jordan power of the observable, which for a single observable coincides with the ordinary power ana^n.

theorem

ω(a)=∫y dμω,a\omega(a) = \int y \, d\mu_{\omega, a}

Let AA be a C∗C^*-algebra, ω\omega be a state on AA, and aa be a self-adjoint element (an observable). The value of the state ω\omega on the observable aa is equal to the integral of the identity function with respect to the real spectral measure μω,a\mu_{\omega, a} associated with ω\omega and aa: ω(a)=∫y dμω,a(y)\omega(a) = \int y \, d\mu_{\omega, a}(y) This identifies the mean (the expected value) of the observable as its first moment.

theorem

Varω(a)=∫y2 dμ−(∫y dμ)2\text{Var}_{\omega}(a) = \int y^2 \, d\mu - (\int y \, d\mu)^2 for spectral measures of observables

Let AA be a unital C∗C^*-algebra, ω\omega be a state on AA, and a∈Aa \in A be a self-adjoint element (an observable). Let μω,a\mu_{\omega, a} denote the real spectral measure (the probability distribution) of aa in the state ω\omega. The variance of aa with respect to ω\omega is equal to the difference between the second moment and the square of the first moment of its spectral measure: Varω(a)=∫Ry2 dμω,a(y)−(∫Ry dμω,a(y))2\text{Var}_{\omega}(a) = \int_{\mathbb{R}} y^2 \, d\mu_{\omega, a}(y) - \left( \int_{\mathbb{R}} y \, d\mu_{\omega, a}(y) \right)^2