亚利桑那大学Bin Zhang:Trajectory-Based Deep Learning Model for Recurrent Event Risk Prediction-华中科技大学管理学院
·学院主页 院长信箱 书记信箱 联系我们 FAQ 本科 | 硕士 | 博士 | MBA | EMBA | MPAcc | EDP
English Version
亚利桑那大学Bin Zhang:Trajectory-Based Deep Learning Model for Recurrent Event Risk Prediction
更多
发布日期:2018-06-13 点击数:

喻园管理论坛 2018年第64期(总第408期)

演讲主题: Trajectory-Based Deep Learning Model for Recurrent Event Risk Prediction

主 讲 人: Bin Zhang(张彬)助理教授, 亚利桑那大学

主 持 人: 杨彦武教授,工商管理系

活动时间: 2018年7月5日(周四)上午9:30-11:00

活动地点: 管理学院119室

主讲人简介: Bin Zhang(张彬)is an assistant professor at Department of Management Information Systems, Eller College of Management, University of Arizona. He is also affiliated member of Artificial Intelligence Lab, University of Arizona. His research interests are Social Network Analysis, Analytical Methods for Large Social Networks, Statistical Modeling for Social Network Problems, Business Intelligence, Machine Learning and Bayesian Statistics.

活动简介: Recurrent event refers to the situation where an event happens again within a specific time interval after the initial occurrence. In the real world, recurrent events can represent failure of system, internet security breach, etc. To alleviate severe consequences of recurrent event, it is crucial to proactively predict its risk. Such prediction is challenging because the evolution of events, also called trajectory, is dynamic and complex. The state-of-the-art studies apply statistical models that assume homogeneity among all observations and use static predictors in a period, failing to consider each observation’s heterogeneous trajectory. Our approach – TADEL (Trajectory-Based Deep Learning) – is motivated to tackle the problem with the existing approaches by capturing various trajectories and accounting for observation heterogeneity.

信息资源
友情链接