BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Empirical Asset Pricing via Ensemble Gaussian Process Regression
DTSTART:20230418T123000
DTEND:20230418T133000
DTSTAMP:20260405T120521Z
UID:3d2f31c1916285886a2dcd188670739b76e48b525b3fd3fe656b34ad
CATEGORIES:Conferences - Seminars
DESCRIPTION:Puneet Pasricha\, Postdoc EPFL\nWe introduce an ensemble learn
 ing method based on Gaussian Process Regression (GPR) for predicting condi
 tional expected stock returns given stock-level and macro-economic informa
 tion. Our ensemble learning approach significantly reduces the computation
 al complexity inherent in GPR inference and takes into account the non-sta
 tionarity of the financial data. We conduct an empirical analysis on a lar
 ge cross-section of US stocks from 1962 to 2016. Our method dominates exis
 ting machine learning models statistically and economically. Exploiting th
 e Bayesian nature of GPR\, we introduce the mean-variance optimal portfoli
 o with respect to the predictive uncertainty distribution. It significantl
 y dominates standard prediction-sorted portfolios and the S&P 500.\n\nAuth
 ors: Damir Filipovic (EPFL) and Puneet Pasricha (EPFL)\n 
LOCATION:UniL-Extranef-125
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR