讲座标题： Big Data in Finance
主讲人： Dr Stephen George PHELPS
讲座语言 英文
讲座时间： 2015-3-18（周三），10:30-12:00
讲座地点： 浦东校区北1117
讲座内容：
In this talk it will give an overview of agent-based modelling. Agent-based models are simulation models with an emphasis on modelling the behaviour of individuals, and are used to in many different disciplines ranging from biology, sociology and economics. Such models can be used to understand real-world complex-adaptive systems which are composed of interacting autonomous agents.
A key research in this domain is how, and if, these systems maintain macroscopic homoeostatic behaviour despite the fact that their constituent agents often face an incentive to disrupt the rest of the system for their own gain. The financial markets present a unique opportunity for empirically studying such systems with the recent availability of high-frequency tick-data, which records every transaction in the market, and can run to many billions of events per exchange per annum. It will give an overview of some of the insights that can be gained from studying these data-sets, as well as some of the technological and statistical challenges involved.
主讲人简介：
The core focus of Dr Stephen’s research is using agent-based modelling to understand real-world complex- adaptive systems. He works with a diverse range of collaborators in different disciplines (Biology, Finance, Computer-Science). He is particularly interested in whether models of cooperation can be validated against empirical studies, and he has had the opportunity to apply many different modelling techniques to a diverse range of data.
For example, He has applied agent-based simulation and evolutionary game-theory to the study of social networks and cooperative behaviour (Phelps, 2013). This work was invited for presentation at the Alternberg Symposia in Theoretical Biology at the Konrad Lorenz Institute, and he is in the process of validating these models against field-data from primate grooming networks in collaboration with a multi-disciplinary group consisting of an evolutionary psychologist (Y. Russell at Middlesex), computer-scientists (Katarzyna Musial-Gabrys at KCL and Mirco Musolesi at Birmingham), and an economist (W.L. Ng at Essex).
He has also applied agent-based simulation to understanding collaboration and competition in markets, which are now almost exclusively technologically mediated. A strength of this work is the development of techniques for modelling high-frequency data that can run to billions of events. He is currently developing methods for using these big data sets to systematically calibrate agent-based simulation models (Nguyen, Phelps and Ng 2014), in order to try and better understand the role of adaptation in explaining some of the phenomena that are observed in empirical financial time-series data, which cannot be accounted-for by the classical theoretical models in this field.