Table 1.
No | Group | Reference | Graph type | Num. features |
---|---|---|---|---|
1 | Node size | Chakrabarti et al. (2005) | Binary | 1 |
2 | Graph density | Chakrabarti et al. (2005) | Binary | 1 |
3 | Degree statistics | Barabasi et al. (2004) | Binary | 4 |
4 | Edge weight statistics | Chakrabarti et al. (2005) | Weight | 4 |
5 | Density wrt. weight cutoffs | Chakrabarti et al. (2005) | Weight | 7 |
6 | Degree correlation statistics | Stelzl et al. (2005) | Binary | 3 |
7 | Clustering coefficient statistics | Barabasi et al. (2004) | Binary | 3 |
8 | Topological coefficient statistics | Stelzl et al. (2005) | Binary | 3 |
9 | First Eigenvalues | Chakrabarti et al. (2005) | Binary | 3 |
10 | Protein weight/size statistics | Cherry et al. (1997) | 4 |
Each row represents a group of similar features. We use 33 features divided into 10 groups. See supporting website for more details. The second column lists the name of the feature group and the third column provides the references. The fourth column specifies which type of graph is used to derive the property.