Skip to main content
Figure 13 | EPJ Data Science

Figure 13

From: Extracting complements and substitutes from sales data: a network perspective

Figure 13

Pairwise NMI between the partitions of \(\mathbf{W}^{(c)}\) varying \(\alpha _{m}\) (leftmost) and varying the threshold quantile \(q_{c}\) with \(\alpha _{m} = 0.01\) (middle-left), and of \(\mathbf{W}^{(s)}\) varying \(\alpha _{l}\) with \(\alpha _{m} = 0.01\), \(q_{c} = 0.35\) (middle-right) and varying the threshold quantile \(q_{s}\) with \(\alpha _{m} = 0.01\), \(q_{c} = 0.35\), \(\alpha _{l} = 0.2\) (right-most), where the original measure is used and axis labels are shown in their titles

Back to article page