Fig. 2
From: The recipe similarity network: a new algorithm to extract relevant information from cookbooks
Structural properties of the three recipe networks as a function of the intersection threshold (\(t\)). The x-axis values indicate that the intersection graph is built with increasing \(\:t\). By increasing \(\:t\), we remove weaker links from the network, performing a link removal analysis of the network’s structural properties. We remove links with weight \(\:w<t\). \(\:t\)=0 indicates that we hold links with \(\:w>0\), and it corresponds to the original network; \(\:t=0.1\) indicates that we hold links with \(\:w>0.1\) and remove links of \(\:w\le\:0.1\); \(\:t=0.2\) indicates that we hold links with \(\:w>0.2\) and remove links of \(\:w\le\:0.2\), and so on. Network structural indicators keys: number of links (\(\:L\)), average link weight (\(\:<w>\)), node degree (\(\:<k>\)), node strength (\(\:<s>\)), network diameter (\(\:D\)) and largest connected component (\(\:LCC\)). Each structural property is normalised by the maximum value for that indicator. Cookbook keys are: Corpus del patrimoni culinari català (CTrad), El Celler de Can Roca (Roca), 100 Plats indispensables de la cuina catalana (101).
