Learning Polytrees with Constant Number of Roots from Data
Chow and Liu  has shown that learning trees that maximize likelihood score given data can be done in polynomial time. A generalization of directed trees are polytrees. However, Dasgupta  has proved that learning maximum likelihood polytrees from data (and even approximation of the optimal result with a constant ratio) is NPHard. Therefore, researchers have focused on learning maximum likelihood polytrees with a constant number of roots. Gaspers et al.  have presented such an algorithm with complexity O(mn3k + 4) using matroid theory. We present a direct combinatorial algorithm with complexity O(mn3k + 1).
Unable to display preview. Download preview PDF.
- 1.Chickering, D.M.: Learning Bayesian networks is NP-complete. In: Learning from data, pp. 121–130. Springer (1996)Google Scholar
- 3.Dasgupta, S.: Learning polytrees. In: Uncertainty in Artificial Intelligence, pp. 134–141 (1999)Google Scholar
- 5.Gaspers, S., Koivisto, M., Liedloff, M., Ordyniak, S., Szeider, S.: On finding optimal polytrees. In: Twenty-Sixth AAAI Conference on Artificial Intelligence (2012)Google Scholar
- 7.Verma, T.S., Pearl, J.: Equivalence and synthesis of causal models. In: Uncertainty in Artificial Intelligence (UAI). pp. 220–227 (1990)Google Scholar
- 8.Wilf, H.S.: Generatingfunctionology. Academic Press (1990)Google Scholar