This shows you the differences between two versions of the page.
seminars:stat:191114 [2019/10/31 07:44] qyu created |
seminars:stat:191114 [2019/11/13 12:58] (current) qyu |
||
---|---|---|---|
Line 1: | Line 1: | ||
+ | <WRAP centeralign>##Statistics Seminar##\\ Department of Mathematical Sciences</WRAP> | ||
+ | |||
+ | <WRAP 70% center> | ||
+ | ^ **DATE:**|Thursday, November 14, 2019 | | ||
+ | ^ **TIME:**|1:15pm -- 2:15pm | | ||
+ | ^ **LOCATION:**|WH 100E | | ||
+ | ^ **SPEAKER:**|Baozhen Wang, Binghamton University | | ||
+ | ^ **TITLE:**|Conformal prediction and the limit of distribution-free conditional predictive inference| | ||
+ | </WRAP> | ||
+ | \\ | ||
+ | |||
+ | <WRAP center box 80%> | ||
+ | <WRAP centeralign>**Abstract**</WRAP> | ||
+ | Conformal prediction uses past experience to determine | ||
+ | precise levels of confidence in new predictions. The proposed methodology | ||
+ | allows for the construction of a distribution-free prediction bands for | ||
+ | the response variable using any algorithm. The resulting prediction bands | ||
+ | preserves marginal coverage guarantee, where predictive coverage holds on | ||
+ | average over all possible test points, but is not sufficient for many | ||
+ | practical applications. We aim to produce predictive conditional coverage | ||
+ | rather than marginally with certain types of constrains and assumptions, | ||
+ | while still being possible to achieve in a distribution-free setting. | ||
+ | |||
+ | </WRAP> | ||
+ | |||
+ | |||
+ | |||
+ | |||