Big Data and Machine Learning in Quantitative Investment

CHAPTER 3State of Machine Learning Applications in Investment Management

Ekaterina Sirotyuk

3.1 INTRODUCTION

Excited by applications of artificial intelligence (AI) used daily via smartphone apps, home products like Alexa and Google Home, as well as matching algorithms used in services of Uber and Facebook,¹ industry professionals outside of financial services and academia wonder why not more, if not the overwhelming majority, of the investment management industry is run on algorithmic principles used by the above‐mentioned tech companies. Quite often I have had conversations with professionals and clients who speculated that if AlphaGo can learn to beat the human so fast, then in a matter of years, it is predominantly the AlphaGos of the world that will be managing institutional and retail investor money. However, aside of questions of trading costs, data collection and processing and execution infrastructure, financial markets represent a much more complex eco system of players with continuous feedback loops that continually rewrite the rule book.

3.2 DATA, DATA, DATA EVERYWHERE

In this context, a common assumption has been that access to proprietary data or big data would a priori create a long‐lasting competitive advantage for an investment strategy. For example, at conference presentations it has been discussed that corporate treasury and finance departments of global businesses with access to customer data (the likes of Ikea) hired quants to make sense out of company ...