We establish exponential inequalities for a class of V-statistics under strong mixing conditions. Our theory is developed via a novel kernel expansion based on random Fourier features and the use of a probabilistic method. This type of expansion is new and useful for handling many notorious classes of kernels.
"Exponential inequalities for dependent V-statistics via random Fourier features." Electron. J. Probab. 25 1 - 18, 2020. https://doi.org/10.1214/20-EJP411