Hadamard derivative

From HandWiki

In mathematics, the Hadamard derivative is a concept of directional derivative for maps between Banach spaces. It is particularly suited for applications in stochastic programming and asymptotic statistics.[1]

Definition

A map φ:𝔻𝔼 between Banach spaces 𝔻 and 𝔼 is Hadamard-directionally differentiable[2] at θ𝔻 in the direction h𝔻 if there exists a map φθ:𝔻𝔼 such thatφ(θ+tnhn)φ(θ)tnφθ(h) for all sequences hnh and tn0. Note that this definition does not require continuity or linearity of the derivative with respect to the direction h. Although continuity follows automatically from the definition, linearity does not.

Relation to other derivatives

Applications

A version of functional delta method holds for Hadamard directionally differentiable maps. Namely, let Xn be a sequence of random elements in a Banach space 𝔻 (equipped with Borel sigma-field) such that weak convergence τn(Xnμ)Z holds for some μ𝔻, some sequence of real numbers τn and some random element Z𝔻 with values concentrated on a separable subset of 𝔻. Then for a measurable map φ:𝔻𝔼 that is Hadamard directionally differentiable at μ we have τn(φ(Xn)φ(μ))φμ(Z) (where the weak convergence is with respect to Borel sigma-field on the Banach space 𝔼).

This result has applications in optimal inference for wide range of econometric models, including models with partial identification and weak instruments.[3]

See also

References

  1. Shapiro, Alexander (1990). "On concepts of directional differentiability". Journal of Optimization Theory and Applications 66 (3): 477–487. doi:10.1007/bf00940933. 
  2. 2.0 2.1 Shapiro, Alexander (1991). "Asymptotic analysis of stochastic programs". Annals of Operations Research 30 (1): 169–186. doi:10.1007/bf02204815. 
  3. Fang, Zheng; Santos, Andres (2014). "Inference on directionally differentiable functions". arXiv:1404.3763 [math.ST].