时间：2019年07月17日（星期三）9:00-18:00
地点：沙河校区，主教楼505
第一部分：Data visualization and statistical graphics
主讲人：Anastasios Panagiotelis，Monash University
Theunit covers data visualization and statistical graphics. The theory,principles and guidelines of good data visualisation as elucidated byEdward Tufte in The Visual Display of Quantitative Information and LelandWilkinson in The Grammar of Graphics, will be covered as well an Rsoftware package, ggplot2 which implements this framework. Some plotsthat will be covered include histograms, scatterplots, bubble plots, violinplots, parallel coordinate plots and mosaic plots. In addition to thesetopics we briefly cover kernel density estimation, LOESS smoothing, Zipf's Law,the Viridis color scale, database normalization and interactive graphicsthrough plotly. The lecture notes can be found athttp://users.monash.edu.au/~anastasp/DataViz/
第二部分：Adaptive and Robust Forecast Combinations
主讲人：Yuhong Yang，School of Statistics，University of Minnesota
Forecastcombination has been studied for over four decades since seminal papers byNobel laureate Granger and his coauthors. Despite many successes in methods andempirical findings, little advancement has been made on deriving forecastcombinations that are flexible to handle multiple possible scenarios of thedata generating process. Indeed, the current forecast combination methodstypically rely on strong assumptions on behaviors of the forecasts in relationto the true values (e.g., stationarity of the forecast errors), which are oftenrestrictive or hard to satisfy. Also, in economic and financial applications,outliers of forecast errors may frequently occur due to various factors such asstructure changes. These observations call for new tools to combine forecastsin adaptive ways that can automatically react to the hidden messages in thedata and can also handle outlier forecast errors robustly. In this talk, wewill present our recent results in this area. Practical examples will be givenfor illustrations.
本次活动受威尼斯432888cam2019专题学术讲座项目资助。
由威尼斯432888cam引智项目支持
