Title: Optimal detection of sparse mixtures with applications to high-dimensional classification
Authors: Tetyana Pavlenko - KTH Royal Institute of Technology (Sweden) [presenting]
Abstract: The focus is on the sparse mixture detection problem for a general non-Gaussian case. We present a class of tests procedures and provide an explicit characterization of the optimal detection boundary under mild regularity conditions. Applications of the obtained results are demonstrated for the adaptive feature selection in high-dimensional classification.