On Identifying Significant Edges in Graphical Models

Details On Identifying Significant Edges in Graphical Models On Identifying Significant Edges in Graphical Models

2011-04-05

arxiv.org/abs/1104.0896v3

Statistics

Machine Learning

18 pages, 6 figures. Presented at the Conference for Artificial Intelligence in Medicine (AIME '11), Workshop on Probabilistic

Scientific paper

Graphical models such as Bayesian networks have been used successfully to capture dependencies between entities of interest from observational data sets. Such dependencies may also reflect possible causal relationships between these entities. Graphical abstractions can also provide system-level insights, hence of great interest across a spectrum of disciplines. Identifying significant edges in the resulting graph have traditionally relied on the choice of an ad-hoc threshold, which can have a pronounced impact on the network topology and conclusions. In the present study, a statistically-motivated approach that obviates the need for an ad-hoc threshold is proposed for identifying significant edges. Several aspects of the proposed approach are investigated across synthetic as well as experimental data sets.

Affiliated with

Also associated with

No associations

Rating

On Identifying Significant Edges in Graphical Models does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.

If you have personal experience with On Identifying Significant Edges in Graphical Models, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and On Identifying Significant Edges in Graphical Models will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFWR-SCP-O-321717