We derive Hanson-Wright inequalities for the quadratic form of a random vector with sparse independent components. Specifically, we consider cases where the components of the random vector are sparse $\alpha$-subexponential random variables with $\alpha>0$. Our proof relies on a novel combinatorial approach to estimate the moments of the random quadratic form. In addition, we obtain a new Bernstein-type inequality for the sum of independent sparse $\alpha$-subexponential random variables. We present two applications with the sparse Hanson-Wright inequality: (1) Local law and complete eigenvector delocalization for sparse $\alpha$-subexponential Hermitian random matrices; (2) Concentration of the Euclidean norm for the linear transformation of a sparse $\alpha$-subexponential random vector.
翻译:暂无翻译