A Note from our hosts:
• If navigating via GPS, be sure to include the full address: 401 East City Avenue, Bala Cynwyd, PA 19004. There will be plenty of parking available in our lot. Our office is also accessible via bus (the 44 and 65).
• Check-in will take place in our first floor lobby. Please bring a photo ID to obtain a visitor badge and entrance to the talk before you head upstairs to our 9th floor game room.
• Our front and side doors will be open until 7pm. If you have any problems entering, please call Liz (215)[masked]
• We will have food! We’re ordering pizza (including vegetarian options), salads, and, SIG’s favorite, Federal Donuts, that will be there by 5:30pm. We’ll also have drinks and coffee.
For our October 2016 meetup, we will gather at SIG (http://sig.com/) and have two speakers.
5:30pm: Pizza and Networking
6:00pm: First talk by Michael Catalano-Johnson
6:45pm: Second talk by Michael Becker
7:30pm Depart for more networking
Bias, Dependence, and Skewed Populations in the Search for Successful Trading Strategies (Michael Catalano-Johnson (https://www.linkedin.com/in/mcj-at-sig))
Sampling biases, hidden dependencies between samples and the difficulty of testing for small effects are well understood statistical phenomena. But despite their simplicity, they rear their heads frequently in the search for anomalies in financial markets. We discuss these issues in general and describe how they arise specifically in examples from quantitative research and financial engineering.
Michael Catalano-Johnson joined SIG in 2001 and has been the Head of Quantitative Research since 2015. Over the past 13 years, he has worked with numerous business units at SIG. Before joining SIG, Michael worked as an Associate at the consulting firm Daniel H. Wagner Associates and as an Assistant Professor of Mathematics at Swarthmore College. He is a graduate of New College of Florida and received his Ph.D. in Mathematics from Brandeis University.
Single Batch Artisanally Crafted Algorithms (Michael Becker (http://twitter.com/beckerfuffle))
For our second talk, Michael Becker will talk about his quest for better book recommendations. He'll teach you the really simple technique he uses for gathering data from almost any web site. He'll walk you through the data munging process. Finally he'll show you how he built a simple classifier.
Michael Becker is a Senior Data Scientist at Penn Medicine where he is building machine learning systems to improve patient outcomes by providing real-time predictive applications that empower clinicians to identify at risk individuals. In his spare time Michael organizes the DataPhilly Meetup group and despite being terrified of public speaking he presents regularly at community events and conferences. On the internet he can be found at http://beckerfuffle.com/