学 术 报 告（一）
题 目：Monitoring the data quality of data streams using a two-step control scheme
时 间：2018年10月17日(周三)下午14:30-15:30
地 点：闵行校区法商南楼135室
报告人：吴纯杰教授 上海财经大学统计与管理学院副院长
摘 要：In the Data-rich environments provide unprecedented opportunities for monitoring data quality. This paper focuses on the quality of data streams. We use indicator variables to measure the six dimensions of data quality and a glitch index to indicate how poor the quality is. A two-step control scheme is proposed considering two relationships: the inter- and intra-correlation. In the first step, the Mahalanobis distance (MD) is applied to the 
-type control chart to monitor the quality of a data stream. In the second step, a Shewhart control chart is built based on a weighted-sum statistic, which measures the quality of the whole process. The feasibility and effectiveness of the control scheme are illustrated through detailed simulation studies and one landslide example. The simulated results, considering the three cases of no correlation, low correlation and high correlation, show that the proposed approach can detect the mean shift in multiattribute data sensitively and robustly. The example, in which sensors are used to collect data on accelerations in Taiwan, demonstrates the superiority of our design over four traditional control charts, producing the closest type-I error to the given level and the highest power under the same type-I error. This talk is based on a joint work with Drs Miaomiao Yu and Fugee Tsung.
报告人简介：吴纯杰，南开大学数理统计理学博士，上海财经大学统计与管理学院副院长（本科教育和实验室建设），教授，博士生导师，全国工业统计教学研究会常务理事，“双法”研究会工业工程分会常务理事。主要从事应用统计和金融统计的科学研究工作，在统计过程控制、市场满意度测评、客户流失分析和金融建模等方面开展了许多有价值的研究工作。科研上发表论文20多篇，其中SCI和EI索引论文7篇，国内权威期刊7篇，主持国家自然科学基金项目2项（含在研1项），参与完成国家级课题4项，主持完成《数理统计》建设获得上海市精品课程和市重点课程优秀结项，完成上海市教改课题1项，获得上海市教学成果一等奖1项和二等奖1项和上海市哲社科学优秀成果二等奖1项，获得2017年“知行杯”上海市大学生社会实践项目大赛优秀指导教师和上海财经大学第十届“我心目中的好老师”。现任SAS公司顾问、SAS认证行业应用专家和飞凤“百脑慧”专家。
学 术 报 告（二）
题 目：Surprise sampling: an optimal subsampling design
时 间：2018年10月17日(周三)下午15:30-16:30
地 点：闵行校区法商南楼135室
报告人：郁 文副教授 复旦大学管理学院
摘 要：Sampling for surprise is a working principle of efficient sampling for the saving of computational workload among other purposes. A sample is deemed surprising if it has large error of pilot prediction or large absolute score, and will be sampled with larger sampling probability, as it in general contains more information than non-surprising samples. Such sampling schemes are particularly useful when dealing with imbalanced data. Following the working principle, we propose a sample design called surprise sampling. It caters to the specific forms of a variety of objectives. The estimation procedure is valid even if the model is misspecified and/or the pilot estimator is inconsistent. The proposed surprise sampling includes as a special case the local case-control sampling (Fithian and Hastie, 2014), which high efficiency by utilizing a clever adjustment pertained only to the logistic model. The proposed estimator also performs no worse than that of (Fithian and Hastie, 2014) under same model specification. We present theoretical justifications of the claimed advantages and optimality of the estimation and the sampling design. Numerical studies are carried out and the evidence in support of the theory is shown.
报告人简介：郁文，复旦大学管理学院副教授，复旦大学概率论与数理统计本科和博士；研究领域包括生存分析，亚组分析，经验似然，半参数统计推断等；迄今已在包括JRSSB在内的国内外统计学权威期刊发表科研论文20多篇；主持国家自然科学基金青年基金项目和面上项目、全国统计科研计划项目、高等学校博士学科点专项科研基金等各一项；连续多年获得复旦大学管理学院MBA教学评估一等奖或二等奖。