Examples of clustering coefficients for different sign and weight configurations.

<p>, and are the clustering coefficient indices proposed by <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0088669#pone.0088669-Watts1" target="_blank">[25]</a>, <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0088669#pone.0088669-Onnela1" target="_blank">[31]</a> and <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0088669#pone.0088669-Zhang1" target="_blank">[9]</a>, respectively. , and are the corresponding indices generalized to the signed case. Solid lines (–) represent edges of weight equal to 1 in absolute value, and dashed lines (–) represent edges of weight close to 0. Green lines are positive and red lines are negative. Edge weights are ignored in the computation of the unweighted clustering coefficients and . In each triangle one edge is negative. Note however that it is irrelevant for the value of the signed clustering coefficients which of the three edges is the negative one. We considered the case of a negative triangle with one negative edge, but we could have equally considered the case of a negative triangle with three negative edges without affecting the results.