BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20250704T140159EDT-6257HniU7J@132.216.98.100 DTSTAMP:20250704T180159Z DESCRIPTION:Title: The Inner Partial Least Square A Probe into the Necessar y” Dimension Reduction.\n\nAbstract: http://users.stat.umn.edu/~liux3771/ \n\n\nThe partial least square (PLS) algorithm retains the combinations of predictors that maximize the covariance with the outcome. The Fisherian i nterpretation of PLS remained a mystery until Cook et al. (2013) showed th at it results in a predictor envelope\, which is the smallest reducing sub space of Σ X that contains the coefficient. This paper is motivated by fin dings after making a seemingly trivial change to the PLS: what if we chang e the max in PLS to min? Counterintuitively\, this does not calculate the complement of the traditional PLS space. Instead\, it results in a new spa ce: the largest reducing subspace of Σ X that is contained in the coeffici ent matrix space. We define the modified PLS as the inner PLS and the resu lting space as the inner predictor envelope space. Unlike the traditional PLS that removes irrelevant information\, the inner PLS incorporates the k nowledge that some information is purely relevant. Consequently\, the inne r PLS algorithm can lead to a more efficient regression estimator than the PLS in certain scenarios\; however\, it is not the most efficient under t he inner predictor envelope model. Therefore\, we derive the maximum likel ihood estimator and provide a non-Grassmannian optimization technique to c ompute it. We confirm the efficiency gain of our estimators both in simula tions and real-world data from the China Health and Nutrition survey.\n\n  \n\nVia Zoom - Please visit our website at: www.mcgill.ca/epi-biostat-occh /news-events/seminars/biostatistics \n DTSTART:20220921T193000Z DTEND:20220921T203000Z SUMMARY:Lan Liu\, PhD\, University of Minnesota at Twin Cities URL:/mathstat/channels/event/lan-liu-phd-university-mi nnesota-twin-cities-342052 END:VEVENT END:VCALENDAR