Z~i[̈ē

ûm点

FQOOXNPROij 15:15 - 17:00

ꏊFcwH@55S قQKlc
(http://www.waseda.jp/jp/campus/okubo.html)

uځF

1. 15:15 - 16:00

@ Statistical Analysis of Self-Exciting Point Processes with
Applications to Marketing

2. 16:15 - 17:00

On Kaplan-Meier Integrals

u : Professor Winfried Stute,  Univ. Giessen, Germany

---------------------------------------------------------
Abstract of Talk (1):
We propose and study minimum distance estimation (MDE) of para-
meters in the context of point processes. These processes are self-
exciting and have a very general class of compensators. Some non-
parametric inputs drive the associated intensities. An application to
the purchasing behavior and the impact of TV-promotion is studied in
detail.
----------------------------------------------------------
Abstract of Talk (2):
The Kaplan-Meier estimator is the efficient nonparametric estimator
of a distribution function, when the data are censored from the right.
Kaplan-Meier integrals are extensions of empirical integrals in that in-
tegration is not taken w.r.t. the classical empirical distribution function
but to KM. In the first part of this talk I will review my work on KM
over the last 15 years. In the second part some recent extensions to
multivariate censored observations are discussed.
----------------------------------------------------------

ûm点

FQOOWNPORPijPRFRO|PVFOO

ꏊFcwۉc(wpZ^[jA@iVjSK
http://www.waseda.jp/jp/campus/index.html

uځF

iPj13:30 - 15:00

Inferential aspects of flexible models generated by perturbation of
symmetric distributions.

Professor Anna Clara Monti, University of Sannio, Italy.

(2) 15:30 - 17:00

FROM DISTRIBUTION-FREENESS TO SEMIPARAMETRIC EFFICIENCY
Sixty years of rank-based inference

Professor Marc Hallin,  Universite Libre de Bruxelles, Belgium.

---------------------------------------------------------
Abstract of Talk (1):
Flexible models generated by perturbation of symmetric distributions are
very appealing and suitable for a wide variety of statistical applications
in a number of different fields because of their nice distributional
properties and various stochastic representations. However, inference about
these models can be problematic for the following reasons: (i) when the
perturbed symmetric distribution is normal, the information matrix can be
singular; (ii) the maximum likelihood estimates of the model parameters that
accommodate skewness and kurtosis can be infinite, even though the true
values of the parameters are finite, and (iii) important sub-models, such as
the normal distribution, can correspond to values of the model parameter
that are on the boundary of the parameter space.
This talk begins with an overview of flexible models and the problems for
inference are discussed. Attention then focuses on one popular flexible
model that has recently emerged, the skew-t model. The standardized indexes
of skewness and kurtosis are both unbounded in this model, so it offers
considerable flexibility for fitting data in which substantial deviations
from normality occur. Consequently, the skew-t model offers an alternative
to robust procedures based on y-functions. Model reparametrizations are
available that remedy the problems associated with the singularity of the
information matrix and inference on the boundary of the parameter space.
These reparametrizations and robustness issues are addressed in the talk.
--------------------------------

Abstract of TALK (2):
The modern history of ranks in statistics started in 1945 with Frank
Wilcoxon's pathbreaking three pages on rank tests for location. Emphasis in
1945 was on distribution-freeness and ease of application. Since then, under
the impulse of such names as Chernoff, Savage, Hodges, Lehmann, Hajek, and
Le Cam, rank-based methods have followed the development of contemporary
statistics, and turned into a complete body of modern, flexible and powerful
techniques. In this talk, we show how this evolution, from
distribution-freeness to group invariance and tangent space projections,
eventually may reconcile the enemy brothers of statistics---efficiency and
robustness.
----------------------------------------

Waseda Seminar on Time Series and Statistical Finance

Ռ (A) vȊwɂ鐔I@̗_Ɖp
@@( \ҁFJ@M (cw) ) ɂ

:  2008N122 ()  14F00 - 17:00

ꏊF  cwHwTPقPWKOU

MAPF  http://www.sci.waseda.ac.jp/campus-map/

ueF

iPj 14F00|15F30 F

Dual P-Values, Evidential Tension and Balanced Tests

Donald Poskitti Monash University, Australiaj

iQj 15:30|17:00 F

On the Theory of Statistical Prediction

Denis Bosq@( Universite Pierre et Marie Curie ( Paris VI )j

Waseda Seminar on Time Series and Statistical Finance

Ռ (A) vȊwɂ鐔I@̗_Ɖp
( \ҁFJ@M (cw) ) ɂ

:  2007N124 ()  14F30

ꏊF  cwHwTPقPWKOU

MAPF  http://www.sci.waseda.ac.jp/campus-map/

ueF

iPj 14F30|15F00 F

Resampling Procedure in Estimation of Optimal Portfolios for VAR(p) Returns of Assets

΁@icwHwj

Abstract

We discuss a resampling proceduer in estimation of optimal portfolios when the return is a vector-valued non-Gaussian autoregressive process of order p. Then it is shown a consistency between the portfolio estimation error and resampled one for the expected portfolio return and portfolio risk. We construct their confidence intervals numerically. The result shows that the confidence intervals are applicable to investigation of actual portfolio estimation errors.

iQj 15F00|16F30 F

On goodness of fit tests for diffusion and point processes

Yuri Kutoyants (Universite du Maine, France)

Abstract

We present a review of several results concerning the construction of the Cramer-von Mises and Kolmogorov-Smirnov type goodness-of-fit tests for continuous time processes. As the models we take a stochastic differential equation with small noise, ergodic diffusion process, Poisson process and self-exciting point processes. For every model we propose the tests which provide the asymptotic size $\alpha$ and discuss the behaviour of the power function under local alternatives. The results of numerical simulations of the tests are presented.

cvwW

FQOOVNPOTij@PTFOO|
ꏊFcwHwTTّRc
(map) http://www.sci.waseda.ac.jp/campus-map/

ueF

iPj  PTFOO|PTFRO
Empirical likelihood approach for non-Gaussian locally stationary processes
@_icwۋ{wj

Abstract
We apply empirical likelihood method to non-Gaussian locally stationary processes. Based on the central imit theorem for locally stationary processes, we calculate the asymptotic distribution of empirical likelihood ratio statistics. Using this method, we can constract confidence inference on various important indices in time series analysis.

iQj@PTFRO|PUFRO

CONDITIONAL INFERENCE IN THE COINTEGRATED VECTOR AUTOREGRESSIVE MODEL
Kees Jan van Garderen (University of Amsterdam)

Abstract
A VAR model with normal disturbances is a Curved Exponential Model. Cointegration imposes further curvature, which means that in addition to reasons for conditioning in nonstationary timeseries given by Johansen (1995), there are further reasons due to the curvature. This paper investigates the effects of conditioning for inference on the speed-of-adjustment coefficients. We show that for some realizations the sample is far less informative than might be expected ex-ante. This should be taken into account when making inference. Conditioning is therefore crucial. We show that conditional inference can be carried out using the observed information instead of the expected information.

iRj@PUFRO|PVFOO
Garderen ͂ł̒kb

ĹAG茤iPVUTOOWPji\ҁ@J@MjɂB

cvO[vu

FQOOTNQVij@PUFOO|PVFOO
ꏊFcwHwTPقPVKOU
(map) http://www.sci.waseda.ac.jp/campus-map/

uځF On accuracy of approximations for multivariate scale mixtures in statistical applications.
uҁFProfessor Ulyanov, V.V. ( Moscow State University )

cvO[vu

FQOOTNPQPij@PTFTO|PUFTO
ꏊFcwHwTTقPKc
(map) http://www.sci.waseda.ac.jp/campus-map/
(ʏƏꏊAԂقȂ܂̂łӂB)

uځF Heavy-tailed elliptical families: optimal inference for shape
uҁFProfessor M. Hallin and Professor D. Paindaveine ( Universite Libre de Bruxelles, Belgium )

cvO[vu

FQOOSNPQQPi΁j@PUFOO|PVFOO
ꏊFcwHwTPقPVKOW
(map) http://www.sci.waseda.ac.jp/campus-map/

uځF Fractional constant elasticity of variance model
uҁFProfessor Ngai Hang Chan ( Chenese University of Hong Kong )

cvZ~i[

FQOOSNPPT@PTFOO|PUFOO
ꏊFcwHwTPقPWKOQ
(map) http://www.sci.waseda.ac.jp/campus-map/

uځFAnalysis of Longitudinal Data with I(2) Signals
uҁFProf. Bob. Shumway ( University of California, Davis )

w@Ռiajۑԍ 15340204
͓I@ɂ`W̉̈萫̌
\ҁFēc@ǍOicwj

Recent Developments in Nonlinear Time Series Analysis with Applications to Finance
@@@@@@
F2004N1 22-23
ꏊFcw(wpZ^[)ۉcAOc(22)Ac(23)
@@@@@@@@@ nScwԁAkPOA
(}bv) http://www.waseda.ac.jp/koho/guide/nisiw.html
ӔCҁFJ@Micwj

vO
January 22(Room 3)
Session I ( Chair: Rainer Dahlhaus )

9:30 - 10:10  : LAN theorem for non-Gaussian locally stationary processes and its applications.
Junichi HIRUKAWA ( Waseda University ) and Masanobu TANIGUCHI ( Waseda University )

10:10 - 10:50 : Application of Bernstein polynomials for density function estimation.
Yoshihide KAKIZAWA ( Hokkaido University )

11:00 - 11:50 : Optimal signed rank tests for elliptical VARMA models.
Marc HALLIN ( Universite Libre de Bruxelles )

Session II ( Chair: Kokyo Naga )

13:00 - 13:40 : Discrimination and clustering for fundamental frequency patterns of
infant and parent data based on time series regression models.
Hiroko KATO ( NTT Communication Science Laboratories )
and Masanobu TANIGUCHI ( Waseda University )

13:40 - 14:30 : Nearest neighbor ARX modeling of spatial time series with application
to localization and connectivity study of functional MRI data.
Tohru OZAKI ( Institute of Statistical Mathematics )

14:30 - 15:20 : Solving high-dimensional inverse problems by spatiotemporal Kalman
filtering.
Andreas GALKA ( Institute of Statistical Mathematics )

Session III ( Chair: Benoit Laine )

15:30 - 16:20 : Signal extraction by state space modeling
Genshiro KITAGAWA ( Institute of Statistical Mathematics )

16:20 - 17:10 : Statistical inference for time-varying ARCH processes.
Rainer DAHLHAUS ( University of Heidelberg )

---------------------------------------------------------------------------------------

January 23(Room 1)
Session IV ( Chair: Oliver Linton )

9:30 - 10:20  : Finite Sample Distributions of the Empirical
Likelihood Estimator and GMM Estimator.
Naoto KUNITOMO ( University of Tokyo )

10:20 - 11:00 : Change-point detection in time series regression models.
Takayuki SHIOHAMA ( Hitotsubashi University )

11:00 - 11:50 : Partial mixing and Edgeworth expansion for stochastic process.
Nakahiro YOSHIDA ( University of Tokyo )

Session V ( Chair: Yoshihide Kakizawa )

13:00 - 13:40 : Outlier detection in time series.
@@@@@@@@ Kokyo NAGA ( Komazawa University )

13:40 - 14:30 : On nonparametric and semiparametric testing of multivariate time
series. Yoshihiro YAJIMA ( University of Tokyo )

Session VI ( Chair: Marc Hallin )

14:40 - 15:20 : Autoregression depth.
Benoit LAINE ( Universite Libre de Bruxelles )

15:20 - 16:10 : An asymptotic test theory of the fractional cointegration rank.
Yuzo HOSOYA ( Tohoku University )

16:10 - 17:00 : Estimation of semiparametric ARCH() models by kernel smoothing.
Oliver LINTON ( London School of Economics )

---------------------------------------------------------------------------------------

Satellite Seminar ̂m点
@(Chair: Masanobu Taniguchi)

January 24, 2004 : 16:00 - 17:00 :  School of Science & Engineering, Waseda University
Room 51- 18 - 02( cwHwTP18
K02 ,@iqcnwˎRԓk10
@@@@@@@@@@@@@@@@@@(}bv) http://www.sci.waseda.ac.jp/campus-map/

Nonparametric MLEs and empirical spectral processes for locally stationary processes.
By Rainer DAHLHAUS ( University of Heidelberg )

---------------------------------------------------------------------------------------

January 27, 2004 : 16:00 - 17:00 :  School of Science & Engineering, Waseda University
Room 51- 18 - 02( cwHwTP18K02 )

Asymptotic expansions for some semiparametric program evaluation estimators.
By Oliver LINTON ( London School of Economics )

---------------------------------------------------------------------------------------

January 29, 2004 : 16:00 - 17:00 :  School of Science & Engineering, Waseda University
Room 51- 18 - 02( cwHwTP18K02 )

From multivariate location to multivariate autoregression quantities : conditional depth revisited
By Marc HALLIN and Benoit LAINE ( Universite Libre de Bruxelles )

FPQPPi؁j@PRFRO|PSFROAPTFOO|PUF OO
ꏊFcwHwTPقPWKOQ

uځFX^C̃phbNXƏk̐E
uҁFwoϊw@vې@B玁

gbvy[W