Coincidence complex networks

Abstract Complex networks, which constitute the main subject of network science, have been wide and extensively adopted for representing, characterizing, modeling an ample range structures phenomena from both theoretical applied perspectives. The present work describes application real-valued Jaccard coincidence similarity indices translating generic datasets into networks. More specifically, two data elements are linked whenever between their respective features, gauged by some index, is greater than a given threshold. Weighted networks can also be obtained taking these as weights. It shown that proposed approaches lead to enhanced performance when compared cosine Pearson correlation approaches, yielding detailed description specific patterns connectivity nodes, with modularity. In addition, parameter ? introduced used control contribution positive negative joint variations considered catering flexibility while obtaining ability methodology capture interconnections emphasize modular structure illustrated quantified respectively real-world including handwritten letters raisin datasets, well Caenorhabditis elegans neuronal network. reported results pave way significant number developments.