A Framework for Inference Using Goodness of Fit Tests Based on Ensemble of Phi-Divergences

摘要: In this paper we study the inferential use of goodness fit tests in a non-parametric setting. The utility such will be demonstrated for test case spectrum sensing applications cognitive radios. We provide first comprehensive framework decision fusion an ensemble goodness-of-fit testing procedures through Ensemble Goodness-of-Fit test. Also, introduce generalized family functionals and kernels called Φ-divergences which allow us to formulate that are parameterized by single parameter. performance these is simulated under Gaussian non-Gaussian noise MIMO show uncertainty statistics or non-Gaussianity noise, general, phi-divergence based particular, significantly superior energy detector with reduced implementation complexity. false alarm rates our proposed maintained at fixed level over wide variation channel distributions. Additionally, describe collaborative spatially separated version robust combining distributed setting quantify significant collaboration gains achieved.