Last Updated: 2007sep10


Topics in Astrostatistics

Statistics 310, Fall/Winter 2006-2007

Harvard University

Instructor Prof. Meng Xiao Li
Schedule Tuesdays 11:30 AM
Location Science Center Rm 705

Fall/Winter 2004-2005
Siemiginowska, A. / Connors, A. / Kashyap, V. / Zezas, A. / Devor, J. / Drake, J. / Kolaczyk, E. / Izem, R. / Kang, H. / Yu, Y. / van Dyk, D.
Fall/Winter 2005-2006
van Dyk, D. / Ratner, M. / Jin, J. / Park, T. / CCW / Zezas, A. / Hong, J. / Siemiginowska, A. & Kashyap, V. / Meng, X.-L.
Hyunsook Lee (Penn State)
7 Sep 2006
A Convex Hull Peeling Depth Approach to Nonparametric Massive Multivariate Data Analysis with Applications
Abstract: We explore the convex hull peeling process to develop empirical tools for statistical inferences on multivariate massive data. Convex hull and its peeling process has intuitive appeals for robust location estimation. We define the convex hull peeling depth, which enables to order multivariate data. This ordering process provides ways to obtain multivariate quantiles including median. Based on the generalized quantile process, we define a convex hull peeling central region, a convex hull level set, and a volume functional, which lead us to invent one dimensional mappings, describing shapes of multivariate distributions along data depth. We define empirical skewness and kurtosis measures based on the convex hull peeling process. In addition to these empirical descriptive statistics, we find a few methodologies to separate multivariate outliers in massive data sets. Those outlier detection algorithms are (1) estimating multivariate quantiles up to the level $\alpha$, (2) detecting changes in a measure sequence of convex hull level sets, and (3) constructing a balloon to exclude outliers. The convex hull peeling depth is a robust estimator so that the existence of outliers do not affect properties of inner convex hull level sets. Overall, we illustrate all these characteristics and algorithms of the convex hull peeling process through bivariate synthetic data sets. We show that these empirical procedures are applicable to real massive data set by employing Quasars and galaxies from the Sloan Digital Sky Survey.
Presentation [.pdf]
Alanna Connors, et al. (CHASC)
19 Sep 2006
A Sense of Motion:
California-Harvard Astronomy and Statistics Collaboration in 2006-2007
Join us for an overview of interdisciplinary problem solving as hosted by Harvard Statistics Dept. and Harvard-Smithsonian Center for Astrophysics, 1997-2006.
In the first Ph.D. in Astronomy ever awarded by Harvard, Cecelia Payne applied new theory and calculational techniqes from one field (Physics) to the new large survey astronomy data of its time (visible energy spectra of many stars). (
Eight decades later, we have new data, from satellites and ground-based observatories, now at many wavelengths beyond the visible. It is becoming increasing high-resolution at even previously inaccessible wavebands. The amount of data is large and slated to become much larger as new missions are launched. In astrophysics, we use the knowledge gained in the Sun/Earth environment as a strong physics-based prior to infer the behavior of all objects in the Universe.
But there is a lack: even with all these astonishing leaps in measurement, there are still basic statistical problems in data analysis techniques which require a deep statistical background to solve. These problems need the perspective of statisticians rather than astronomers.
Hence there are many problems, both simple and complex, for which even a beginning student in statistics can make a serious contribution to modern astronomy.
In our opening talk, we will give a general historical overview, including a summary of some of our work so far. Our group has tended to concentrate on astronomy data that cannot be well described by a Gauss-normal distribution: especially UV, X-ray, and gamma-ray data. However all kinds of problems are welcome. We will introduce progress and challenges from last Spring's widely attended SAMSI Special Workshop in Astrostatistics. We will also introduce a contest in the related field of high energy physics! We invite all to come and comment.
The High Energy Groove [.mov]
Pavlos Protopapas (CfA)
14 Nov 2006
New Challenge: Time Series Center
The Time Series Center is a new center hosted at the Initiative of Innovative Computing (IIC) at Harvard. The center will collect among others a large set of Astronomical time series (light curves) from various surveys.
I will describe few projects underway with emphasis on unsolved statistical question.
Presentation: [.pdf] [.ppt]
Jonathan McDowell (CXC/CfA)
12 Dec 2006
Astrophysics for Mathematicians
Astronomical image, spectroscopic, and variability data challenges for Statisticians and Mathematicians.
Presentation: [.ppt] [.pdf]
Rima Izem (Harvard)
30 Jan 2007
Reducing the dimensionality of RMFs
Stephane Blondin (CfA/OIR)
13 Mar 2007
How to Classify Spectra of Exploding Stars(?)
Professional and amateur astronomers discover several hundreds of exploding stars (supernovae) each year. These are classified into four main types based on characteristic features in their optical spectra. Nevertheless, the supernova classification scheme poses several conceptual problems, all of which point towards the need for an unbiased and automated classification system. I will present a simple cross-correlation algorithm that was adapted to this issue, although the aim of this seminar is to seek advice from the statistics community as to which tools are best suited to classify supernova spectra.
Presentation: [.pdf]
Hyunsook Lee (CfA)
10 Apr 2007
Introducing BLoCXS and using it to estimate calibration uncertainties
Abstract: The flowchart of BLoCXS is presented for a discussion among participants to refine the proposed methods in resolving calibration uncertainties from high energy astrophysics.
Presentation: [.pdf]
Flowcharts: [.ps] [.ps] [.ps]
Updated flowchart: [.pdf]
Andreas Zezas & Hyunsook Lee (CfA)
24 Apr 2007
Developing an alternative method to measure the history of stellar populations"
Optical photometry and in particular color-magnitude diagrams are a standard tool for studies of stellar populations. In this presentation we will discuss the basic astrophysical background, describe the most commonly used methods and contrast them with a new more robust method which is currently under development.
Andreas' Presentation: [.pdf] [.ppt]
Hyunsook's Presentation: [.pdf]
Jing Chen Liu (Harvard)
w. Xiao-Li Meng, Michael Ratner, Irwin Shapiro
08 May 2007
An Exploratory Statistical Study of the Measured Motion of the Guide Star Used in a New Test of Einstein's Universe
General relativity predicts that the phenomenon of "frame-dragging" slowly alters the direction of spin of Earth-orbiting gyroscopes. NASA's Gravity Probe B satellite measured the orientation of such freely falling gyroscopes with respect to a guide star. This star's motion with respect to "fixed points" in the distant universe thus needed to be measured independently. VLBI measurements spread over 14 years yielded 39 positions of this radio-emitting guide star. We use Bayesian methods and a linear regression model of the star's motion to study the effects of various assumptions involving the parameters of both the physical model and the measurement noise model. We find that a wide range of different assumptions lead to a narrow range of point estimates and error distributions.
David van Dyk (UC Irvine)
14 May 2007
CHASC meeting
John Rice (UC Berkeley)
08 June 2007
Rm 403 @ 60 Oxford
Event Weighted Tests for Periodicity in a Sequence of Photon Arrival Times: Detecting Gamma-ray Pulsars