BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//jEvents 2.0 for Joomla//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:b83aef1224806df6a95ee3400e34badc
CATEGORIES:Seminars
CREATED:20180802T094306
SUMMARY:Svetlana Bryzgalova - London Business School
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:\n\nForest Behind the Trees (joint with Markus Pelger and Jason Zhu)\n\n\nA
 bstract: \nSorting-based strategy of building portfolios has been a default
  empirical approach in asset pricing for creating both test assets and fact
 or-mimicking returns. One of the natural limitations of this technique, how
 ever, is its inability to adequately reflect the information contained in m
 ore than 2 characteristics and their interaction. Yet recent advances in em
 pirical asset pricing have repeatedly highlighted the importance of the lat
 ter, e.g. Freyberger et al (2017), Kozak et al (2018). We propose to analyz
 e the effect of a large number of characteristics on expected stock returns
  with the machine learning technique known as random forest. As an ensemble
  learning method for classification, the new approach is particularly well-
 suited for building composite cross-sections of portfolios that reflect the
  rich conditional information contained in a large number of characteristic
 s simultaneously, and can be viewed as a natural generalization of the conv
 entional sorting-based strategies. We build decision trees for various sets
  of stock-specific characteristics, and demonstrate that the new approach i
 s able to create cross-sections that a) reflect the information in a joint 
 conditional distribution of characteristics, b) are challenging to price ba
 sed on the conventional models, even when pitted against the tradable facto
 rs based on the underlying characteristics, and c) imply investment strateg
 ies that achieve yearly out-of-sample Sharpe ratios above 2.\n
DTSTAMP:20260405T192437Z
DTSTART:20181011T163000Z
DTEND:20181011T180000Z
SEQUENCE:0
TRANSP:OPAQUE
END:VEVENT
END:VCALENDAR