### Abstract

Original language | English |
---|---|

Title of host publication | PMLR: Proceedings of Machine Learning Research |

Pages | 16-18 |

Number of pages | 3 |

Volume | 15 |

Publication status | Published - 11 Apr 2011 |

### Publication series

Name | PMLR: Proceedings of Machine Learning Research |
---|---|

Volume | 15 |

ISSN (Print) | 1938-7228 |

### Fingerprint

### Cite this

*PMLR: Proceedings of Machine Learning Research*(Vol. 15, pp. 16-18). (PMLR: Proceedings of Machine Learning Research; Vol. 15).

}

*PMLR: Proceedings of Machine Learning Research.*vol. 15, PMLR: Proceedings of Machine Learning Research, vol. 15, pp. 16-18.

**Discussion of "learning equivalence classes of acyclic models with latent and selection variables from multiple datasets with overlapping variables".** / ZHANG, Jiji; SILVA, Ricardo.

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed)

TY - GEN

T1 - Discussion of "learning equivalence classes of acyclic models with latent and selection variables from multiple datasets with overlapping variables"

AU - ZHANG, Jiji

AU - SILVA, Ricardo

PY - 2011/4/11

Y1 - 2011/4/11

N2 - Learning equivalence classes of acyclic models with latent and selection variables from multiple datasets with overlapping variables is discussed. The problem of inferring the presence of latent variables, their relation to the observables, and the relation among themselves, is considered. A different approach for identifying causal structures, one that results in much simpler equivalence classes, is provided. It is found that the computational cost is much higher than the procedure implemented, but if datasets are individually of modest dimensionality, it might be doable in practice. From the point of view of search algorithms for optimizing structure, much of the machinery of combinatorial optimization could optimize the penalized composite likelihood score by enforcing constraints such that the independence models over different subsets of variables agree on the overlapping sets.

AB - Learning equivalence classes of acyclic models with latent and selection variables from multiple datasets with overlapping variables is discussed. The problem of inferring the presence of latent variables, their relation to the observables, and the relation among themselves, is considered. A different approach for identifying causal structures, one that results in much simpler equivalence classes, is provided. It is found that the computational cost is much higher than the procedure implemented, but if datasets are individually of modest dimensionality, it might be doable in practice. From the point of view of search algorithms for optimizing structure, much of the machinery of combinatorial optimization could optimize the penalized composite likelihood score by enforcing constraints such that the independence models over different subsets of variables agree on the overlapping sets.

UR - http://proceedings.mlr.press/v15/zhang11a.html

UR - http://commons.ln.edu.hk/sw_master/6568

M3 - Conference paper (refereed)

VL - 15

T3 - PMLR: Proceedings of Machine Learning Research

SP - 16

EP - 18

BT - PMLR: Proceedings of Machine Learning Research

ER -