Collective Classification Techniques: An Experimental Study
Collective classification is the area in machine learning, in which unknown nodes in the network are classified based on the classes assigned to the known nodes and the network structure only. Three collective classification algorithms were described and examined in the paper: Iterative Classification (ICA), Gibbs Sampling (GS) and Loopy Belief Propagation (LBP). Experiments on various networks revealed that greater number of output classes decreases classification accuracy,GS provides better results than ICA and LBP outperforms others for dense structures while it is worse for sparse networks.
KeywordsGibbs Sampling Unknown Node Gibbs Sampling Algorithm Label Assignment Loopy Belief Propagation
Unable to display preview. Download preview PDF.
- 3.Gallagher, B., Eliassi-Rad, T.: Leveraging label-independent features for classification in sparsely labeled networks: An empirical study. In: Proceedings of Second ACM SIGKDD Workshop on Social Network Mining and Analysis, SNA-KDD 2008 (2008)Google Scholar
- 7.Lu, Q., Getoor, L.: Link-based classification. In: Proceedings of 20th International Conference on Machine Learning, ICML, San Francisco, pp. 496–503 (2003)Google Scholar
- 9.Nooy, W., Mrvar, A., Batagelj, V.: Exploratory Social Network Analysis with Pajek, ch. 11. Cambridge University Press (2004)Google Scholar
- 10.Pearl, J.: Probabilistic reasoning in intelligent systems. Morgan Kaufmann (1988)Google Scholar
- 12.Sen, P., Namata, G., Bilgic, M., Getoor, L., Gallagher, B., Eliassi-Rad, T.: Collective classification in network data. Artificial Intelligence Magazine 29(3), 93–106 (2008)Google Scholar
- 13.Taskar, B., Abbeel, P., Koller, D.: Discriminative probabilistic models for relational data. In: Proceedings of 18th Conference in Uncertainty in Artificial Intelligence. Morgan Kaufmann, San Francisco (2002)Google Scholar