Marginalization and Collapsibility in Graphical Interaction Models

نویسندگان

چکیده

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

On Marginalization, Collapsibility and Precollapsibility

It is shown that for every undirected graph G over a nite set N and for every nonempty T N there exists an undirected graph G T over T, called the marginal graph of G for T, such that the class of marginal distributions for T of (discrete) G-Markovian distributions coincides with the class of G T-Markovian distributions. An example shows that this is not true within the framework of strictly po...

متن کامل

Adjustments and their Consequences— Collapsibility Analysis using Graphical Models

We review probabilistic and graphical rules for detecting situations in which a dependence of one variable on another is altered by adjusting for a third variable (i.e., non-collapsibility or noninvariance under adjustment), whether that dependence is causal or purely predictive. We focus on distinguishing situations in which adjustment will reduce, increase, or leave unchanged the degree of bi...

متن کامل

Beating the Likelihood: Marginalization-Based Parameter Learning in Graphical Models

Likelihood based-learning of graphical models faces challenges of computational-complexity and robustness to model error. This paper studies methods that fit parameters directly to maximize a measure of the accuracy of predicted marginals, taking into account both model and inference approximations at training time. Experiments on imaging problems suggest marginalization-based learning performs...

متن کامل

Marginalization in Composed Probabilistic Models

Composition of low-dimensional distribu­ tions, whose foundations were laid in the pa­ per published in the Proceedings of UAI'97 (Jirousek 1997), appeared to be an alterna­ tive apparatus to describe multidimensional probabilistic models. In contrast to Graphi­ cal Markov Models, which define multidimen­ sional distributions in a declarative way, this approach is rather procedural. Ordering of...

متن کامل

On the Collapsibility of Lifetime Regression Models∗

In this paper we derive conditions on the internal wear process under which the resulting time-to-failure model will be of the simple collapsible form (Oakes, 1995, Duchesne and Lawless, 2000) when the usage accumulation history is available. We suppose that failure occurs when internal wear crosses a certain threshold and/or a traumatic event causes the item to fail (Cox, 1999 and Bagdonavičiu...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: The Annals of Statistics

سال: 1990

ISSN: 0090-5364

DOI: 10.1214/aos/1176347626