Uncovering Soft Information from Stock Market Conference Calls: Asset Management Perspectives

Investors, regulators, and the general public consume a wealth of textual information every day. Recent advancements in artificial intelligence make machine-reading of textual information plausible. We tackle text mining of financial conference call transcripts—calls of significant corporate events that are widely followed by investors and institutional investors. Our conference calls data include over 200,000 calls calls held by North American companies. We will use textual analysis and machine learning in computer science, combined with large-scale portfolio-formation and regression analysis in financial economics, to uncover systematic patterns in conference calls that will affect future stock returns. Our project aims to enhance the trading strategies of the sponsoring organization, to ultimately benefit its Canadian institutional clients. Our project will also contribute innovatively to the academic literature on textual analysis of conference calls.

Intern: 
Chufeng Hu
Faculty Supervisor: 
Alan Huang
Project Year: 
2018
Province: 
Ontario
Program: