{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "\n", "\n", "Following Section 4.2.2.4 of [Müller, FMP, Springer 2015], we discuss in this notebook thresholding strategies that are applicable for a wide range of matrix representations. We show the behavior of these strategies by applying them to self-similarity matrices. The functionalities are also provided by the MATLAB toolbox described in the following article:\n", "\n", "
threshold_matrix, which comprises all the thresholding variants discussed above.\n",
"\n",
"threshold_matrix: \n",
" \n",
"* Some parameter settings may not make sense (for exampling, using scaling with penalty as well as binarization). \n",
"* Depending on the strategy, one may need one threshold (in case of 'absolute' and 'relative') or two threshold parameters (in case of 'local').\n",
"* Because of rounding issues and the fact that different matrix entries may have identical values, the split as specified by $\\rho$ may not be exact. \n",
"![]() | \n",
" ![]() | \n",
" ![]() | \n",
" ![]() | \n",
" ![]() | \n",
" ![]() | \n",
" ![]() | \n",
" ![]() | \n",
" ![]() | \n",
"