|タイトル||Parameter Identification in Structural Equation Models|
|開催日時||2017年3月1日 14:30-15:30 ※いつもと時間が異なります|
|講演者||Prof. Mathias Drton氏 (University of Washington)|
|内容||We treat statistical models that relate random variables of interest via a linear equation system that features stochastic noise.
These models, also known as structural equation models, are naturally represented by a mixed graph, with directed edges indicating non-zero coefficients in the linear equations and bidirected edges indicating possible correlations among noise terms.
In this talk we report on progress on combinatorial conditions for parameter identifiability.
Identifiability holds if coefficients associated with the edges of the graph can be uniquely recovered from the covariance matrix they define.