基于互信息的适用于高维数据的因果推断算法

摘要

推断数据间存在的因果关系是很多科学领域中的一个基础问题。然而现在暂时还没有快速有效的方法对高维数据进行因果推断。为此，提出了一种基于互信息的适应于高维数据的因果推断算法，该算法采取将高维网络结构学习问题分解成每一个节点的因果网络结构学习问题的策略。在第一阶段，利用基于互信息的条件独立性测试算法寻找目标节点的父子节点；在第二阶段，利用一种混合的方向识别算法对目标节点与其父子节点之间的方向进行判别，所有节点迭代完后得到一个完整的因果网络。数据实验表明，该算法在高维数据的情况下要优于目前其他的算法。%Discovering causality from data set is one of the basic problems in many scientific fields.However,there is still no fast and effective method to discover causality from high dimensional data.This paper proposed an approach for large scale causality discovery based on mutual information.The algorithm decomposed an n-dimensional network into n parts with respect to each node.In the first phase,it explored conditional independence tests based on mutual information to discover the causal-ities of a certain node.In the second phase,it conducted a hybrid direction learning approach to distinguish the directions be-tween nodes.It constructed the whole structure of the high dimensional network at the end of the last iteration.Experimental results show that the proposed algorithm outperforms other methods in high dimensional case.