RIT on Statistics Archives for Fall 2016 to Spring 2017
Organizational Meeting
When: Tue, September 15, 2015 - 3:30pm
Where: MATH 1313
Speaker: Dr. Paul Smith (UMCP) -
Clustering: Algorithms, Applications, and potential research extensions
When: Tue, September 22, 2015 - 3:30pm
Where: MATH 1313
Speaker: Cheng Jie (UMCP) -
Clustering
When: Tue, September 29, 2015 - 3:30pm
Where: Math 1313
Speaker: Jie Cheng & Alex Estes (UMCP) -
Performance of Classifiers
When: Tue, October 6, 2015 - 3:30pm
Where: MTH 1313
Speaker: Lijuan Cao (UMCP) -
Principal Components Analysis
When: Tue, October 13, 2015 - 3:30pm
Where: MTH 1313
Speaker: Franck Ndjakou Njeunje (UMCP) -
Spectral Methods
When: Tue, October 20, 2015 - 3:30pm
Where: MTH 1313
Speaker: David Bekkerman (UMCP) -
Text Clustering and Classification
When: Tue, November 3, 2015 - 3:30am
Where: MTH 1313
Speaker: Lijuan Cao (UMCP) -
Clustering and the Bootstrap
When: Tue, November 10, 2015 - 3:30am
Where: MTH 1313
Speaker: Paul Smith (UMCP) -
Representative Traffic Management Initiatives
When: Tue, November 10, 2015 - 3:30pm
Where: MTH 1313
Speaker: Alex Estes (UMCP) -
Organizational meeting
When: Mon, February 15, 2016 - 3:00pm
Where: MTH 0201
Speaker: Takumi Saegusa (UMCP) -
Abstract: This semester's RIT will deal with Empirical Process Theory with applications to Semiparametric Models.
Organizational Meeting
When: Mon, February 22, 2016 - 3:00pm
Where: MTH 0201
Speaker: Takumi Saegusa (UMCP) -
TBD
When: Mon, February 29, 2016 - 3:00pm
Where: MTH0201
Speaker: Takumi Saegusa (UMCP) -
Z-theorems
When: Mon, March 7, 2016 - 3:00pm
Where: MTH0201
Speaker: Takumi Saegusa (UMCP) -
Cox model with right censoring
When: Mon, March 21, 2016 - 3:00pm
Where: MTH0201
Speaker: Takumi Saegusa (UMCP) -
Martingales versus Empirical Process Theory in Proving the Functional Central Limit Theorem for Repeated Logrank Statistics
When: Mon, March 28, 2016 - 3:00pm
Where: MTH0201
Speaker: Eric Slud (UMCP) -
Abstract: In this talk, I will begin by defining the logrank statistic for testing equality of survival outcomes in a two-group survival experiment and briefly explain its relation to martingales. The Martingale Central Limit Theorem turns out to be a very good way to prove a Central Limit Theorem for this kind of statistic, when the data-sample is large. However, suppose one wishes to prove a Functional Central Limit Theorem covering the large-sample behavior of the logrank statistic calculated repeatedly over time, considered as a stochastic process. Then we will see that empirical process techniques and NOT martingales are the best way to prove a large-sample limit theorem.
The material of the lecture will intersect material from survival analysis (logrank statistic, which is discussed for example in the Survival Analysis text of Klein and Moeschberger), from a couple of well-known papers one by Tsiatis (1982), one by Sellke (1984) and one by me (1984), and from empirical-process material that can be found in later papers of Ming Gao Gu.
Cox Model (continued)
When: Mon, April 11, 2016 - 3:00pm
Where: MTH0201
Speaker: Takumi Saegusa (UMCP) -