报告题目：Generalized Additive Model and its applications
报告时间：2017年4月26日 (星期三) 14:30-15:15
报告地点：尚贤楼1楼108报告厅
报告主持人: 曹春正副教授
报告内容简介：
Generalized Additive Models (GAMs) extend linear and generalized linear models (GLMs) to include smooth functions of predictors in an additive framework. GAM has been demonstrated as a modelling tool based on its three characteristics: (1) Easy to interpret; (2) Flexible predictor functions to uncover hidden patterns in the data; (3) Regularization of predictor functions to avoid overfitting. In general, GAM has the interpretability of GLMs where the contribution of each predictor to the prediction is clearly modelled. It provides a flexibility because the smooth functions of predictors give the ability to model the relationship between the response variables and predictors, whether it is linearly or nonlinear. The data-driven estimated smooth functions are therefore most useful to describe unknown the relationship between the response variable and predictors when there is no prior knowledge. From an estimation point of view, the use of regularized nonparametric functions avoids the pitfalls of dealing with higher order polynomial terms in the linear models. From the accuracy standpoint, GAMs are competitive with or superior to other popular techniques.
In this talk, we will focus on a brief introduction about the GAM modelling technique and discuss its applications in modelling climate problems, and forecasting electricity consumption incorporating nonlinear weather influences.
专家简介：
李云博士是澳大利亚西澳电力商业情报与数据分析高级咨询顾问，高级预报和模型分析研究员。主要负责西澳电力网络的可靠性和供电量的统计建模和预报研究（包括天气极端事件对电力网络的可靠性和供电量的影响），为西澳电力高层决策和规划管理提供内部高级数据分析咨询和建议。
李云博士加盟西澳电力之前（2000-2015）曾是澳大利亚联邦科学与工业研究组织（CSIRO）的首席研究员，主要从事气候统计学的研究。研究成果包括:天气和气候极端事件的统计建模、统计降尺度模型、气候变化诊断、降水和气温的统计预报等。
李云博士是2007-2009年澳大利亚-中国气候变化双边合作的首席科学家。并与中国科学院大气物理研究所的李建平教授成功地完成了由中国发展和改革委员会（The Chinese National Development and Reform Commission）和澳大利亚气候变化部(The Australian Department of Climate Change)共同资助的项目“Research on Rainfall and Climate Change in both China and Australia”。李云博士自2007年以来，与中国科学院、中国气象局和有关大学保持长期合作，接待访问学者和联合培养8名博士研究生。
欢迎广大师生踊跃参加！
数学与统计学院
2017-04-25
