Generalized Score Functions for Causal Discovery.

Generalized Score Functions for Causal Discovery.
复制标题

DOI:
10.1145/3219819.3220104
复制
发表时间:
2018-08
期刊:
KDD : proceedings. International Conference on Knowledge Discovery & Data Mining
影响因子:
--
通讯作者:
Glymour C
Glymour C
中科院分区:
其他
文献类型:
--
作者:
Huang B;Zhang K;Lin Y;Schölkopf B;Glymour C

文献摘要

参考文献

被引文献

相似文献

从观测数据中发现因果关系是一个基本问题。粗略地说,因果关系发现有两种方法,基于约束的方法和基于分数的方法。与基于约束的方法相比,基于分数的方法避免了多重测试问题,具有一定的优势。然而,他们中的大多数都需要对因果机制的功能形式以及数据分布进行强有力的假设,这限制了他们的适用性。实际上,底层模型类的精确信息通常是未知的。如果违反上述假设,则可能导致伪边缘和丢失边缘。在本文中,我们引入广义分数函数的因果关系发现的基础上,一般(条件)的随机变量之间的独立关系的表征,而不假设特定的模型类。特别是,我们利用回归RKHS捕获的依赖性在一个非参数的方式。由此产生的因果发现方法产生渐近正确的结果,而一般情况下,这可能有非线性因果机制,广泛的一类数据分布,混合连续和离散数据,和多维变量。合成和真实世界的数据上的实验结果表明,我们所提出的方法的有效性。
Discovery of causal relationships from observational data is a fundamental problem. Roughly speaking, there are two types of methods for causal discovery, constraint-based ones and score-based ones. Score-based methods avoid the multiple testing problem and enjoy certain advantages compared to constraint-based ones. However, most of them need strong assumptions on the functional forms of causal mechanisms, as well as on data distributions, which limit their applicability. In practice the precise information of the underlying model class is usually unknown. If the above assumptions are violated, both spurious and missing edges may result. In this paper, we introduce generalized score functions for causal discovery based on the characterization of general (conditional) independence relationships between random variables, without assuming particular model classes. In particular, we exploit regression in RKHS to capture the dependence in a non-parametric way. The resulting causal discovery approach produces asymptotically correct results in rather general cases, which may have nonlinear causal mechanisms, a wide class of data distributions, mixed continuous and discrete data, and multidimensional variables. Experimental results on both synthetic and real-world data demonstrate the efficacy of our proposed approach.
DOI: 10.1159/000330168
发表时间: 2011-01-01
期刊: HUMAN HEREDITY
影响因子: 1.8
作者:
Bakken, Trygve E.;Dale, Anders M.;Schork, Nicholas J.
通讯作者: Schork, Nicholas J.
DOI: 10.1214/14-aos1260
发表时间: 2014-12-01
影响因子: 4.5
作者:
Buehlmann, Peter;Peters, Jonas;Ernest, Jan
通讯作者: Ernest, Jan
DOI: 10.1016/s0140-6736(15)01156-3
发表时间: 2016-05-07
期刊: LANCET
影响因子: 168.9
作者:
Springmann, Marco;Mason-D'Croz, Daniel;Scarborough, Peter
通讯作者: Scarborough, Peter
DOI: 10.1016/j.neubiorev.2013.12.004
发表时间: 2014-02-01
影响因子: 8.2
作者:
Ruigrok, Amber N. V.;Salimi-Khorshidi, Gholaihreza;Suckling, John
通讯作者: Suckling, John
DOI: 10.1007/s10994-006-6889-7
发表时间: 2006-10-01
期刊: MACHINE LEARNING
影响因子: 7.5
作者:
Tsamardinos, Ioannis;Brown, Laura E.;Aliferis, Constantin F.
通讯作者: Aliferis, Constantin F.