Local Collective Classification Algorithm

A Local Collective Classification Algorithm is a Collective Classification Algorithm that makes use of Local Feature Vectors.

• AKA: Local Collective Classification.
• Example(s):
  • (Neville & Jensen, 2000)
  • (Chakrabarti et al., 1998c).
• See: Global Collective Classification Algorithm.

---

---

References

2009

• (Bilgic & Getoor, 2009) ⇒ Mustafa Bilgic, and Lise Getoor. (2009). “Reflect and Correct: A misclassification prediction approach to active inference.” In: ACM Transactions on Knowledge Discovery from Data (TKDD), 3(4). doi:10.1145/1631162.1631168
  • There are many collective classification models proposed to date that make different modeling assumptions about these dependencies. They can be grouped into two broad categories. In the first category, local collective classification models, the collective models consist of a collection of local vector-based classifiers, such as logistic regression. For the this category of collective models, each object is described as a vector of its local attributes X_i and an aggregation of attributes and labels of its neighbors. Examples include [[Chakrabarti et al. [1998]], [[Neville and Jensen [2000]], [[Lu and Getoor [2003a]], [[Macskassy and Provost [2007]], and [[McDowell et al. [2007]].

2007

• (Macskassy & Provost, 2007) ⇒ Sofus Macskassy, and Foster Provost. (2007). “Classification in Networked Data: A toolkit and a univariate case study.” In: Journal Machine Learning, 8.
• (McDowell et al., 2007) ⇒ L. McDowell, K. M. Gupta, and David W. Aha. (2007). “Cautious Inference in Collective Classification.” In: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI 2007).

2000

• (Neville & Jensen, 2000) ⇒ Jennifer Neville, and David Jensen. (2000). “Iterative Classification in Relational Data.” In: Proceedings of the Workshop on Statistical Relational Learning.

1998

• (Chakrabarti et al., 1998c) ⇒ Soumen Chakrabarti, Byron Dom, and Piotr Indyk, (1998). “Enhanced Hypertext Categorization Using Hyperlinks.” In: Proceedings of the 1998 ACM SIGMOD Conference (SIGMOD 1998). doi:10.1145/276304.276332

---