387
个编辑
更改
→因果结构
用顶点表示事件，边表示[[因果关系]]的图通常是无环的。<ref>{{citation|title=Causal Learning|first1=Alison|last1=Gopnik|first2=Laura|last2=Schulz|publisher=Oxford University Press|year=2007|isbn=978-0-19-803928-0|page=4|url=http://books.google.com/books?id=35MKXlKoXIUC&pg=PA4}}.</ref>事件由时间上的先后顺序来排列，所有箭头遵循从先发生事件指向后发生事件的原则，因此也不存在环。
用顶点表示事件，边表示[[因果关系]]的图通常是无环的。<ref>{{citation|title=Causal Learning|first1=Alison|last1=Gopnik|first2=Laura|last2=Schulz|publisher=Oxford University Press|year=2007|isbn=978-0-19-803928-0|page=4|url=http://books.google.com/books?id=35MKXlKoXIUC&pg=PA4}}.</ref>事件由时间上的先后顺序来排列，所有箭头遵循从先发生事件指向后发生事件的原则，因此也不存在环。
举例来说，<font color="#ff8000"> 贝叶斯网络 Bayesian network </font>表示多个概率事件的关联网络。顶点表示事件，后续事件的发生可能性则可以通过其在有向无环图的前驱节点的发生概率计算出来。<ref>{{citation|title=Probabilistic Boolean Networks: The Modeling and Control of Gene Regulatory Networks|publisher=Society for Industrial and Applied Mathematics|first1=Ilya|last1=Shmulevich|first2=Edward R.|last2=Dougherty|year=2010|isbn=978-0-89871-692-4|page=58|url=http://books.google.com/books?id=RfshqEgO7KgC&pg=PA58}}.</ref>在此基础上，一个有向无环图的<font color="#ff8000"> 端正图 Moral graph </font>通过以下方法而得到：将单个顶点的所有父节点之间添加一条无向边，再将所有的有向边换成无向边。<ref>{{citation |last1= Cowell |first1= Robert G. |author2-link=Philip Dawid|last2=Dawid|first2=A. Philip|author3-link=Steffen Lauritzen|last3=Lauritzen|first3=Steffen L.|author4-link=David Spiegelhalter|last4=Spiegelhalter|first4=David J.|title= Probabilistic Networks and Expert Systems |publisher= Springer |year= 1999 |isbn= 0-387-98767-3 |chapter= 3.2.1 Moralization|pages= 31–33 }}.</ref>
举例来说，<font color="#ff8000">贝叶斯网络 Bayesian network </font>表示多个概率事件的关联网络。顶点表示事件，后续事件的发生可能性则可以通过其在有向无环图的前驱节点的发生概率计算出来。<ref>{{citation|title=Probabilistic Boolean Networks: The Modeling and Control of Gene Regulatory Networks|publisher=Society for Industrial and Applied Mathematics|first1=Ilya|last1=Shmulevich|first2=Edward R.|last2=Dougherty|year=2010|isbn=978-0-89871-692-4|page=58|url=http://books.google.com/books?id=RfshqEgO7KgC&pg=PA58}}.</ref>在此基础上，一个有向无环图的<font color="#ff8000"> 端正图 Moral graph </font>通过以下方法而得到：将单个顶点的所有父节点之间添加一条无向边，再将所有的有向边换成无向边。<ref>{{citation |last1= Cowell |first1= Robert G. |author2-link=Philip Dawid|last2=Dawid|first2=A. Philip|author3-link=Steffen Lauritzen|last3=Lauritzen|first3=Steffen L.|author4-link=David Spiegelhalter|last4=Spiegelhalter|first4=David J.|title= Probabilistic Networks and Expert Systems |publisher= Springer |year= 1999 |isbn= 0-387-98767-3 |chapter= 3.2.1 Moralization|pages= 31–33 }}.</ref>
另外一种具有相似因果结构的图是<font color="#ff8000"> 影响图 Influence diagram </font>。其顶点表示决策或不确定的事件，边表示两个顶点之间的因果关系。<ref>{{citation|title=The Technology Management Handbook|first=Richard C.|last=Dorf|publisher=CRC Press|year=1998|isbn=978-0-8493-8577-3|page=9-7<!-- Do not conver this hyphen into a dash! It is a section-page number, not a range of page numbers. -->|url=http://books.google.com/books?id=C2u8I0DFo4IC&pg=SA9-PA7}}.</ref>在[[流行病学]]中，这些表示因果关系的图表常常用来评估不同干预手段的效果。<ref>{{citation|title=Encyclopedia of Epidemiology, Volume 1|first=Sarah|last=Boslaugh|publisher=SAGE|year=2008|isbn=978-1-4129-2816-8|page=255|url=http://books.google.com/books?id=wObgnN3x14kC&pg=PA255}}.</ref><ref name=pearl:95>{{cite journal|last1=Pearl|first1=Judea|authorlink= Judea Pearl|title=Causal diagrams for empirical research|journal=Biometrika|date=1995|volume=82|issue=4|pages=669–709|doi=10.1093/biomet/82.4.669}}</ref>
另外一种具有相似因果结构的图是<font color="#ff8000"> 影响图 Influence diagram </font>。其顶点表示决策或不确定的事件，边表示两个顶点之间的因果关系。<ref>{{citation|title=The Technology Management Handbook|first=Richard C.|last=Dorf|publisher=CRC Press|year=1998|isbn=978-0-8493-8577-3|page=9-7<!-- Do not conver this hyphen into a dash! It is a section-page number, not a range of page numbers. -->|url=http://books.google.com/books?id=C2u8I0DFo4IC&pg=SA9-PA7}}.</ref>在[[流行病学]]中，这些表示因果关系的图表常常用来评估不同干预手段的效果。<ref>{{citation|title=Encyclopedia of Epidemiology, Volume 1|first=Sarah|last=Boslaugh|publisher=SAGE|year=2008|isbn=978-1-4129-2816-8|page=255|url=http://books.google.com/books?id=wObgnN3x14kC&pg=PA255}}.</ref><ref name=pearl:95>{{cite journal|last1=Pearl|first1=Judea|authorlink= Judea Pearl|title=Causal diagrams for empirical research|journal=Biometrika|date=1995|volume=82|issue=4|pages=669–709|doi=10.1093/biomet/82.4.669}}</ref>