Skip to main content
Computational Social Networks
Fig. 6 | Computational Social Networks

Fig. 6

From: Utilizing the simple graph convolutional neural network as a model for simulating influence spread in networks

Fig. 6

This figures presents the results of applying the approaches to the dataset of ‘email-Eu-core’ [43, 44] part of the SNAP collection [45]. In a, b the nodes in the first ten departments belong to category ’two’ (red), the nodes in departments 11 & 12 to category ’three’ (orange), and the rest are in category ’one’ (blue). There is asymmetric influence with category ’three’ having the most which affects the label association with different values of k. For a, Eq. 4 is used, and in b, Eq. 7 is used. With c, d, the same number of nodes for the different categories is used, but instead of choosing those nodes by department, they are selected from the ordered degree centrality rank. Both pairs of plots demonstrate how Eq. 4 depicts a mechanism for influence spread upon the number of walks, and with Eq. 7 the decay over the length with the carry-on effect

Back to article page