사단법인 한국통계학회

본문 바로가기
사이트 내 전체검색

게시판BBS

학술행사/연구정보

HOME > 게시판 BBS > 학술행사/연구정보

[안내] 중앙대학교 데이타과학연구소 세미나

페이지 정보

작성자 관리자 작성일09-06-04 15:44 조회3,743회

본문

중앙대학교 통계학과 및 데이타과학연구소에서는 미국 University of Georgia 통계학과의 박철우 교수와 안정연 교수를 모시고 아래와 같이 세미나를 개최하오니 관심 있는 분들의 많은 참여를 부탁 드립니다.

일시 : 2009/6/19 (금요일), PM 4:00~ 5:50
장소 : 자연과학대학(수림과학관) 7층 9705 강의실
 
발표 1 : PM4:00 ~4:50


발표자: 박철우 교수 (University of Georgia)
Title: Exploratory Time Series Analysis Based on Multiscale Statistical Tools 
Abstract:
The first part of the talk introduces Internet traffic data analysis using wavelets and SiZer (SIgnificance of ZERo crossings of the derivative).  It is important to characterize burstiness of Internet traffic and find the causes for building models that can mimic real traffic. To achieve this goal, exploratory analysis tools and statistical tests are needed, along with new models for aggregated traffic. The intricate fluctuations of Internet traffic are explored in various respects and lessons on long range dependence and nonstationarities from real data analyses are summarized.

In the second part of the talk, we conduct an investigation of the null hypothesis distribution for
functional magnetic resonance imaging (fMRI) data using multiscale analysis. We remove spatial dependence using Principal Component Analysis and analyze the detrended data with a long enough time horizon to study possible long-range temporal dependence. Our multiscale approach shows that even for resting data, i.e. "null" or ambient thought, some voxel time series cannot be modeled by white noise and need long-range dependent type error structure, while for other voxels white noise is a reasonable assumption.


발표 2 : PM5:00~5:50


발표자: 안정연 교수 (University of Georgia)
Title: Maximal Data Piling for HDLSS Data Analysis
Abstract:
In clustering with high dimension, low sample size data, the traditional distance measures between clusters, such as complete
linkage and single linkage, etc., are no longer valid. The maximal data piling distance which originates from binary classification, is a distance between hyperplanes of the two classes and also a natural distance measure between HDLSS clusters. In this talk, we introduce a hierarchical clustering method based on the maximal data piling distance. Applications to microarray gene expression datasets will be presented.