Title :
Online causal discovery
Author :
Yu, Kui ; Wu, Xindong ; Wang, Hao
Author_Institution :
Dept. of Comput. Sci., Hefei Univ. of Technol., Hefei, China
Abstract :
The standard causal discovery assumes that all variables are available from the beginning. In this paper, we consider an untouched scenario in which not all variables are available in advance. We call this scenario online causal discovery which assumes that the target of interest is given in advance while the other variables are unknown. With this situation, an online algorithm is presented which consists of two phases: online growing and online shrinking phase. Experimental results validate our algorithms compared with a state-of-the-art standard algorithm of causal discovery.
Keywords :
belief networks; learning (artificial intelligence); Bayesian network; online causal discovery algorithm; online growing phase; online shrinking phase; Algorithm design and analysis; Bayesian methods; Classification algorithms; Heuristic algorithms; Markov processes; Measurement; Probability distribution; Bayesian network; causal discovery; online causal discovery;
Conference_Titel :
Cognitive Informatics (ICCI), 2010 9th IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-8041-8
DOI :
10.1109/COGINF.2010.5599825
