Figure 11From: Extracting complements and substitutes from sales data: a network perspectiveDistributions of the relative number of shared flavour compounds of all product pairs \(\{rf_{ij}\}\) (“all”), of complementary pairs \(\{rf_{ij}: W_{ij}^{(c)} > 0\}\) (“com”) and of substitute pairs \(\{rf_{ij}: W_{ij}^{(s)} > 0\}\) (“sub”), where probabilities of \(rf_{ij}=0\) and \(rf_{ij}=1\) are of interestBack to article page