BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250731T211300EDT-0364NeuUU5@132.216.98.100
DTSTAMP:20250801T011300Z
DESCRIPTION:Mr. Chengyu Zhang\, a doctoral student at ºÚÁϲ»´òìÈ University in 
 the Finance area will be presenting his thesis defense entitled:\n\nThree 
 Essays on the Applications of Machine Learning in Finance\n\n \n\nWednesda
 y\, July 31\, 2024 at 10:00 a.m.\n	(The defense will be conducted on Zoom)
 \n\nStudent Committee Chair: Professor Ruslan Goyenko\n\nPlease note that 
 the presentation will be conducted on Zoom. If you wish to attend the pres
 entation\, kindly contact the PhD Office.\n\n\nABSTRACT\n\nThis thesis con
 sists of three essays that covers various topics on the applications of ma
 chine learning methods in the field of finance.\n\nThe first essay introdu
 ces a machine learning framework that non-parametrically estimates the opt
 imal dynamic portfolio strategy subject to realistic and predictable tradi
 ng costs. Conditioning on a comprehensive set of stock characteristics and
  macroeconomic indicators\, the trading-cost-aware portfolio strategy subs
 tantially outperforms market benchmarks in out-of-sample tests\, and is ro
 bust to various limits-to-arbitrage constraints. I demonstrate that incorp
 orating explicit trading-cost penalty is critical to avoid extracting perf
 ormance from small and illiquid stocks\, better capture market stress peri
 ods\, and allocate assets based on more fundamental signals.\n\nIn the sec
 ond essay\, we study whether the stock market or the options market has th
 e leading informational advantage. By conducting a horse-race using a larg
 e set of stock and option characteristics with machine learning\, we find 
 that option characteristics dominates the return predictability for both s
 tocks and options. Among option characteristics\, option illiquidity is th
 e most important predictor for both stock and option returns\, and we unco
 ver positive option illiquidity premium in the stock returns\, as an incre
 ase in derivatives trading decreases information asymmetry and stock price
  uncertainty.\n\nIn the third essay\, we train classic machine learning al
 gorithms and large language models\, LLMs\, to predict future earnings sur
 prises using textual data extracted from quarterly and annual filings of U
 .S. corporations. We observe a negative correlation between the length of 
 the MD&A section and both future earnings surprises and firm returns. In a
 ddition\, conventional machine learning methods that rely on sentiment ana
 lysis or bag-of-words techniques fail to effectively leverage past manager
 ial discussions for accurate predictions of future earnings. We find that 
 only finance-objective trained LLMs have the capacity to comprehend the co
 ntextual information embedded in 10-Q and 10-K filings to predict both pos
 itive and negative earnings surprises\, and future firm returns.\n
DTSTART:20240731T140000Z
DTEND:20240731T160000Z
SUMMARY:PhD Thesis Defense Presentation: Chengyu Zhang
URL:/desautels/channels/event/phd-thesis-defense-prese
 ntation-chengyu-zhang-358025
END:VEVENT
END:VCALENDAR