Queue Mining: From Predictive to Prescriptive Analytics in Congested Systems.

Prof. Arik Senderovich, University of Toronto, Canada

Wednesday November 25th, 2017, 12:00pm (noon).
Location: Zoom

Predictive and prescriptive analytics are the backbone of data-driven process optimization in organizations. In the talk, I will introduce queue mining, which is a set of data-driven methodologies for descriptive, predictive, and prescriptive analysis of congested systems such as hospitals, public transportation systems and contact centers. We shall start with an overview of queue mining and its interaction with process mining and machine learning. Next, we will switch to focus to predictive analytics and present congestion graph mining, a feature engineering approach that transforms transactional logs into useful predictive ML models. In the last part of the talk, we shall show how one switches from predictive analytics, that focus on predicting time-to-events of interest, to prescriptive analytics that aim at changing and improving the underlying system in a data-driven fashion. To demonstrate the usefulness of the methods we will provide results of applying queue mining methods to real-world datasets coming from the Technion’s Service Enterprise Engineering lab (SEELab).

About the Speaker. Arik Senderovich is an Assistant Professor at the Faculty of Information (University of Toronto). He holds a PhD in Data Science (2017), an MSc degree in Statistics (2012) and a BSc degree in Industrial Engineering and Management (2006): all three degrees from the Technion – Israel Institute of Technology. Before his appointment at the Faculty of Information, he received the Lyon Sachs scholarship (awarded to one PhD grad per year) and worked as a postdoctoral fellow in the Toronto Intelligent Decision Engineering Laboratory (TIDEL) at the University of Toronto. Arik’s research focuses lies on the intersection between Operations Management, Data Science, and Artificial Intelligence. Currently, he focuses on developing methodologies for automatically learning models of complex and congested environments (such as hospitals and public transportation systems) from data logs. His research has recently received an acknowledgement at the Fifteenth International Conference on Business Process Management in Barcelona (2017) where he was awarded with both the inaugural best dissertation award and the best paper award for that year.