Hadamard derivative

Summary

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 edit

A map   between Banach spaces   and   is Hadamard-directionally differentiable[2] at   in the direction   if there exists a map   such that

 
for all sequences   and  .

Note that this definition does not require continuity or linearity of the derivative with respect to the direction  . Although continuity follows automatically from the definition, linearity does not.

Relation to other derivatives edit

Applications edit

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

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

See also edit

References edit

  1. ^ Shapiro, Alexander (1990). "On concepts of directional differentiability". Journal of Optimization Theory and Applications. 66 (3): 477–487. CiteSeerX 10.1.1.298.9112. doi:10.1007/bf00940933. S2CID 120253580.
  2. ^ a b Shapiro, Alexander (1991). "Asymptotic analysis of stochastic programs". Annals of Operations Research. 30 (1): 169–186. doi:10.1007/bf02204815. S2CID 16157084.
  3. ^ Fang, Zheng; Santos, Andres (2014). "Inference on directionally differentiable functions". arXiv:1404.3763 [math.ST].