An Overview of Joint Modeling of Longitudinal and Time to Event Data

11/03/2022 at 1:37 pm

At the thirteenth annual American Conference on Pharmacometrics (ACoP13), I gave a presentation that was an introduction/overview of joint models of longitudinal and time to event data. The slides from the talk can be downloaded here: acop4b-overview

Since joint modeling is the basis of many models used for pharmacometric applications, the intent of these slides was to give some of the general concepts together with multiple references for someone to get started if they are interested in learning more about this class of joint models. Some of the content is taken from other presentations and I have tried to make sure the original source is cited in all cases.

The general structure of the slides is:

  1. What is Joint Modeling of longitudinal and time to event data
  2. How do we develop joint models.
  3. Stepwise versus simultaneous estimation
  4. Bayesian Joint Models
  5. Applications of Joint Models with directions to stan code
  6. Additional

JSM 2010: An Example of Performance Analysis for Network Community Detection

08/02/2010 at 9:08 pm

Abstract: Community detection algorithms have many applications ranging from search engines on the world wide web to the detection terrorist networks. While the computer scientists are trying to detect “clumpiness” in networks, as statisticians we are analyzing the performance of the algorithm. We pick a node and determine its expected number of neighbors with a degree greater than or equal to this chosen node. This is colloquially referred to the number of nodes in the chosen node’s “bucket”. We show that, for a multigraph with an arbitrary node degree distribution, both the number of nodes in the bucket and the number of pairs of nodes in the bucket are asymptotically finite. This is in contrast to an Erdos Renyi random graph where this quantity increases with O(n).

Poster: JSM 2010 Poster

Post Election Audits in Iowa

09/12/2009 at 10:17 pm

Abstract: During the summer of 2008, several statisticians in Iowa learned of developing post-election audit legislation for the state, a collective effort of state legislators, county auditors and active citizens. Since then, faculty and students at three Iowa universities have worked with audit advocates around the state and election audit experts from around the country to promote risk-limiting audit legislation in Iowa. The input from a number of statisticians with experience and expertise in post-election audits has helped shape our strategy, which is focused on developing a dialogue between statisticians and election officials. Our initial contacts with local officials have underscored the critical importance of communication in promoting risk-limiting audits. Our efforts in 2009 will include additional student involvement and exploring pilot audits for Iowa’s next general election.

Slides: Post Election Audits Presentation

The Pros of Pro Bono Statistical Consulting

08/01/2009 at 10:00 pm

Abstract: Students involved in pro bono statistical work, while acquiring hands-on skills that contribute to their own professional development, also embrace a spirit of volunteerism that will likely stay with them throughout their careers. In recent years, an increasing number of student statisticians have become actively engaged in pro bono work and have helped contribute to an expansion of volunteerism in the profession. The panel includes statisticians who have volunteered their time during graduate school and beyond, and have been invited to reflect on their experiences.

Slides: JSM 2009 StatCom Slides