 |
Date |
July/08/2022 |
Time |
Japan Standard Time (UTC+9h) |
|
11:30~13:00 |
Speaker |
Taisuke OtsuThe London School of Economics and Political Sciences
|
Title |
Likelihood Inference Under Alternative Asymptotics
|
Abstract |
The seminar will be based on the following three papers: (i) Empirical Likelihood for Network Data (with Yukitoshi Matsushita); (ii) Multiway Empirical Likelihood (with Harold Chiang and Yukitoshi Matsushita); (iii) Jackknife Empirical Likelihood: Small Bandwidths, Sparse Network and High-Dimensional Asymptotics (with Yukitoshi Matsushita).
|
Venue |
Click here to join the Zoom
Meeting Meeting ID: 313 176 9518
Passcode: 119549
|
Chair |
Qihui Chen The Chinese University of Hong Kong, Shenzhen
|
Hosts |
TEDS and The Chinese University of Hong Kong, Shenzhen
|
|
Date |
May/27/2022 |
Time |
Japan Standard Time (UTC+9h) |
|
9:30 ~ 11:00 |
Speaker |
Zhentao Shi Georgia Institute of Technology
|
Title |
Boosted Hodrick-Prescott Filter Is More General Than You Think |
Abstract |
Recent historical economic events such as the Great Recession and the Covid-19 recession have reignited the debate on econometric methodology for filtering low-frequency macroeconomic time series. While the Hodrick-Prescott (HP) filter remains, arguably, the off-the-shelf choice, the boosted HP (bHP) filter emerges as its modern machine learning upgrade for the data-rich environment. This paper sheds light on understanding the versatility of bHP in fitting time series of various nonstationary patterns. We establish a new asymptotic result that each iteration of the HP filter shrinks complex exponential functions proportionally toward zero. This founding opens a unified approach to tackle trends generated by unit root processes, higher order integrated processes, and local-to-unity processes. In an asymptotic framework where the penalty tuning parameter expands with the sample size, this theoretical property explains the smoothing effect of the original HP filter, and more importantly, it guarantees that the bHP filter consistently restores all the underlying trending processes mentioned above, in conjunction with potential deterministic trends and structural breaks. When it is applied to the universe of the FRED-QD dataset, bHP timely captures the downturns at crises and recessions and the recoveries that followed. Buttressed by its sound theoretical coverage and empirical performance, the automated bHP filtering procedure is well suited as a general-purpose machine for low-frequency macroeconomic time series.
|
Chair |
Xinyu Zhang Chinese Academy of Sciences
|
Hosts |
TEDS and Chinese Academy of Sciences
|
|
Date |
April/26/2022 |
Time |
Japan Standard Time (UTC+9h) |
|
10:50 ~ 12:20 |
Speaker |
Katsumi Shimotsu The University of Tokyo
|
Title |
Testing the Order of Multivariate Normal Mixture Models |
Abstract |
Finite mixtures of multivariate normal distributions have
been widely used in empirical applications in diverse fields such as statistical genetics and
statistical finance. Testing the number of components in multivariate normal mixture models is a
long-standing challenge even in the most important case of testing homogeneity. This paper develops
a likelihood-based test of the null hypothesis of M components against the alternative hypothesis of
M+1 components for a general M>=1. We derive the asymptotic distribution of the proposed EM test
statistic under the null hypothesis and local alternatives and show the validity of parametric
bootstrap. The simulations show that the proposed test has good finite sample size and power
properties.
|
Zoom |
Click here to join the Zoom
Meeting Meeting ID: 857 7302 7215 Passcode: 674245
|
Venue |
Koganei Campus, Hosei University. If you want to join in person, please contact qliu@hosei.ac.jp.
|
Chair |
Qingfeng Liu Hosei University
|
Hosts |
TEDS and Hosei University
|
|
Date |
April/12/2022 |
Time |
Japan Standard Time (UTC+9h) |
|
10:00 ~ 11:30 |
Speaker |
Jinchi Lv University of Southern California
|
Title |
High-Dimensional Knockoffs Inference for Time Series Data |
Abstract |
The recently introduced framework of model-X knockoffs provides a flexible tool for exact finite-sample false discovery rate (FDR) control in variable selection in arbitrary dimensions without assuming any dependence structure of the response on covariates. It also completely bypasses the use of conventional p-values, making it especially appealing in high-dimensional nonlinear models. Existing works have focused on the setting of independent and identically distributed observations. Yet time series data is prevalent in practical applications in various fields such as economics and social sciences. This motivates the study of model-X knockoffs inference for time series data. In this paper, we make some initial attempt to establish the theoretical and methodological foundation for the model-X knockoffs inference for time series data. We suggest the method of time series knockoffs inference (TSKI) by exploiting the idea of subsampling to alleviate the difficulty caused by the serial dependence. We establish sufficient conditions under which the original model-X knockoffs inference combined with subsampling still achieves the asymptotic FDR control. Our technical analysis reveals the exact effect of serial dependence on the FDR control. To alleviate the practical concern on the power loss because of reduced sample size cause by subsampling, we exploit the idea of knockoffs with copies and multiple knockoffs. Under fairly general time series model settings, we show that the FDR remains to be controlled asymptotically. To theoretically justify the power of TSKI, we further suggest the new knockoff statistic, the backward elimination ranking (BE) statistic, and show that it enjoys both the sure screening property and controlled FDR in the linear time series model setting. The theoretical results and appealing finite-sample performance of the suggested TSKI method coupled with the BE are illustrated with several simulation examples and an economic inflation forecasting application. This is a joint work with Chien-Ming Chi, Yingying Fan and Ching-Kang Ing.
|
Video |
Click here to watch the video
|
Chair |
Yoshimasa Uematsu Hitotsubashi University
|
Hosts |
TEDS and Hitotsubashi University
|