CRETA is honored to invite Professor Yuya Sasaki from Vanderbilt University as a visitor on May 18. During his visit, Prof. Sasaki will lecture on Deconvolution Methods: Theory and Practical Guide on CRETA Workshop on Advanced Econometrics 22. The workshop is due to take place on May 18 (Friday) at NTU, College of Management Building 1, Room 405 (台大管理學院一號館405教室). Please be sure to register your attendance online by noon, May 16 (Wednesday).
*Date: May 18, 2018 (Fri.), 14:00 pm – 16:50 pm
*Venue: NTU, College of Management Building 1, Room 405 (台大管理學院一號館405教室)
*Topics: Deconvolution Methods: Theory and Practical Guide
This lecture presents statistical and econometric theories of deconvolution and their applications in economics. Topics include deconvolution kernel density estimation under known error distributions (1988–), deconvolution kernel density estimation under unknown error distributions with auxiliary sample (1993–), deconvolution kernel density estimation under unknown symmetric error distributions with repeated observations (1996–), deconvolution kernel density estimation under unknown error distributions with repeated observations (1998–), deconvolution errors-in-variables regression estimation under known error distributions (1993–), deconvolution errors-in-variables regression estimation under unknown error distributions with repeated observations (2002–), point-wise inference for density functions under known error distributions (1991–), point-wise inference for errors-in-variables regressions under known error distributions (1992–), uniform inference for density functions under known error distributions (2007–), uniform inference for density functions under unknown error distributions (2017–), uniform inference for errors-in-variables regressions under unknown error distributions (2017–), bandwidth selection by normal reference, bandwidth selection by Parseval’s Identity, bandwidth selection by the SIMEX method, applications in earnings dynamics and income dynamics, applications in auction models, and applications in production analysis. The lecture starts with an overview of the history of the literature. The emphasis is placed on the practice of how to compute estimates, how to form confidence bands, and how to choose bandwidths and regularization parameters. The lecture also presents a guideline on which methods may be used under various scenarios of empirical research, including the case where data with validation sample is available and the case where data with repeated observations is available. A selected set of theories and methods is illustrated with real data analysis.
講者介紹:
[About the Speaker]
Professor Yuya Sasaki is an associate professor of economics at Vanderbilt University since 2017. The field of specialization is econometrics. Professor Sasaki received a Ph.D. at Brown University in 2012, and served as an assistant professor of economics at Johns Hopkins University from 2012 to 2017. He is currently an associate editor of Journal of Econometric Methods. He has contributed to a number of research areas including panel data, quantile regressions, regression discontinuity and kink designs, measurement errors, deconvolution, and robust inference.