[CoWy ASA] Reminder: CO/WY 2010 Spring Meeting (April 23, 2010)

[Christopher Nelson]

Hello everyone - Just a quick reminder that the CO/WY 2010 Spring meeting
will be held at the end of this month.  Details are included below.  A job
announcement is also included below.

 

##### REMINDER: SPRING CO/WY CHAPTER MEETING (April 23, 2010) #####

 

The spring ASA CO/WY meeting will be held on Friday, April 23, 2010 at the
NCAR facility in Boulder, CO. There is no need to RSVP for this event. We
will meet in the ML-239 Damon Room located at NCAR - 1850 Table Mesa Dr.,
Boulder, CO 80305.  Directions - Take U.S. Highway 36 to Boulder Colorado .
Exit at Table Mesa Drive . Head West on Table Mesa Drive . The Mesa Lab and
Fleischmann Building are located at 1850 Table Mesa Drive where the road
ends at the base of the Rocky Mountains

 

Event: CO/WY Spring Meeting 2010

Date: April 23, 2010

Location: NCAR, ML-239 Damon Room

Time:  9:00am - 4:00pm

 

AGENDA

8am-9am: Refreshments

9am-11:30am Welcome and Presentations

11:30am-1:15pm: Lunch on your own or $5 in the cafeteria

1:15pm-4pm: Presentations

 

Speakers

 

Myung-Hee Lee [mhlee.csu at gmail.com]

Title: Clustering High Dimensional, Low Sample Size Data using Maximal Data
Piling:  We present new hierarchical clustering method for high dimension,
low sample size (HDLSS) data. The method utilizes the fact that each
individual data vector accounts for exactly one dimension in the subspace
generated by HDLSS data. The linkage that is used for measuring the distance
between clusters is the orthogonal distance between a ne subspaces generated
by each cluster. The ideal implementation would be to consider all possible
binary splits of data and choose the one that maximizes the distance
in-between. Since this is not computationally feasible in general, however,
we use singular value decomposition for its approximation. We provide
theoretical justification of the method by studying high dimensional
asymptotics. Also we obtain the probability distribution of the distance
measure under the null hypothesis of no split, which we use to propose a
criterion for determining the number of clusters. Simulation and real data
analyses with microarray data show competitive clustering performance of the
proposed method.  (This is a joint work with Jeongyoun Ahn and Youngju Yoon)


 

A.M. Santos [pintyesantos at gmail.com]

Robust estimation of mixtures of heavy-tailed distributions A. M. Santos,
IBM and UC-Denver Karen Kafadar, Indiana University. We examine the
robustness and efficiency of a variant of the EM algorithm for estimating
parameters of mixtures of h-distributions. Specifically, we use the biweight
in an EM-like algorithm to estimate location, scale, and proportion of each
component of a mixture of heavy-tailed distributions. This family of
long-tailed distributions, indexed by a tail-length parameter (h=0
corresponds to the Gaussian), provides a measure by which robustness and
efficiency can be assessed.  We compare our results to those using the
conventional EM-algorithm (that assumes Gaussian distributions) and with
Scott's L_2E (2001) approach.  Finally, we describe the application that
motivated this research and our plans for future work.

 

Jun Zhu [jun.zhu.e at gmail.com]: Variable Selection in Spatial Linear
Regression

 

Bahr, Timothy [TIMOTHY.BAHR at UCDENVER.EDU]

SLAM: Gaussian Dynamic Linear Anaysis of MeDIP-chip Data

This paper presents an efficient and powerful algorithm, GausSian Dynamic
Linear Anaysis of Methylated- chip data (SLAM), for proling DNA methylation
patterns from samples enriched for methylated DNA through
immunoprecipitation (MeDIP) and hybridized on genome tiling microarrays
(MeDIP-chip). SLAM is distinguished from other existing algorithms by its
application of MeDIP-chip specific normalization, dynamic linear smoothing
and a Probit transformation in order to accurately profile the percentage of
methylated DNA bases across the genome. SLAM is the first MeDIP-chip
analysis method that can directly compare the DNA methylation levels of
multiple tissue types, as well as being the rst method appropriate for the
analysis of tiling array data for other epigenetic marks, such as histone
methylation or acetlyation. We apply SLAM to several epigenetic pro ling
datasets. Results show that SLAM can provide more efficient, accurate and
consistent methylation estimates than other existing algorithms, implying
that SLAM has the greater potential for biological discovery. SLAM is
available as a command-line application or in a graphical user interface,
and is freely available for download at:
http://statistics.byu.edu/johnson/SLAM/.

 

Amber Hackstadt [ahacksta at lamar.colostate.edu]

A Bayesian Approach to Fitting Mixed Models Using Shape Restricted
Regression Splines, Amber Hackstadt, Mary Meyer, and Jennifer Hoeting:  We
propose a Bayesian approach to fit shape-restricted regression splines for
mixed models. Linear mixed-effects models are often used to analyze repeated
measure and longitudinal data but often the only thing that is known about
the data is the shape (monotone increasing/decreasing or concave/convex) and
smoothness of the regression functions. Shape-restricted splines are used to
model regression functions in mixed-effects models. 

 

Yuan Wang [ywang.csu at gmail.com]

Carbon Flux in the Mid-Continent Region Project: The net exchange of CO2
between the terrestrial biosphere and atmosphere remains a key uncertainty
in the carbon cycle.  An important campaign has been launched in the
Mid-Continent region of USA, and our main goal is to compare, diagnose and
reconcile two different estimates of CO2 exchange over the study region.

 

Bruce Bugbee [brucebugbee at gmail.com]

A popular topic in e commerce research is the observation and quantifying of
bidder behavior in online auction systems. A functional methodology and
cluster analysis was used to quantify bidder behaviors across two distinct
item types--collectibles (1968 Camaros) and commodities (generic digital
camera). Similar behaviors were observed with varying proportions of
occurrence  and levels of auction success."

 

 

##### JOB ANNOUCEMENT #####

 

Job Posting: Instructor in the Business School at the University of Colorado
Denver.

We're seeking a full-time instructor to teach undergraduate and graduate
(master's level) statistics courses. More information about the position is
available at
https://www.jobsatcu.com/applicants/jsp/shared/position/JobDetails_css.jsp?p
ostingId=221403. The job posting number is 809565.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/cowystats/attachments/20100407/be77492e/attachment.html