Feature Selection In High Dimensional Data