学术报告

当前位置:学术交流 >> 学术报告 >> 浏览文章

2018年厦门大学数据科学研讨会

发布时间:2018年11月22日 浏览次数: 文章作者:2018-11-25 发布者:mathky

会议名称:2018年厦门大学数据科学研讨会

会议日期:2018年11月25日

会议地点:厦门大学行政楼A306

日期

时间

事项

地点

8:45-9:00

开幕式、合影

厦门大学

海韵校区

行政楼A306

9:00-9:30

郭建华(东北师范大学):从统计学到数据科学

主持人

梁薇

9:30-10:00

陈敏中国科学院): 统计学的魅力

10:00-10:30

茶歇

10:30-11:00

周勇华东师范大学/中国科学院): 高低数据下市场波动率和杠杆效应研究

主持人

胡杰

11:00-11:30

孙六全(中国科学院/广州大学):生存分析简介及研究现状

11:30-12:00

梁薇(厦门大学): Mean Empirical Likelihood

14:00-14:30

胡杰(厦门大学): Bayesian Detection of Event Spreading Pattern from Multivariate Binary Time Series

厦门大学

海韵校区

行政楼A306

14:30-16:00

统计学科发展咨询会

 

 


学术报告题目与摘要

 

从统计学到数据科学

郭建华东北师范大学

 

摘要: 大数据的涌现,使得世界进入到数据科学的时代。而统计学,作为收集和分析数据的科学,究竟在数据科学中占有什么样的地位?本次演讲,我将从世界观、价值观、方法论等几方面,根据自己的一些认识,探讨统计学和数据科学的一些内在逻辑。

 

 

统计学的魅力

陈敏(中国科学院)

Abstract:

 

 

高低数据下市场波动率和杠杆效应研究

周勇(华东师范大学/中国科学院)

AbstractThe leverage effect is an extensively important phenomenon that describes the (usually) negative relation between stock returnsand their volatility.However, many previous works are conducted over full observed asset prices or with additive error, and few of them have carefully studied its estimation via taking into account the market microstructure, especially with high-frequency data. Few scholars have carefully studied the estimation of leverage effect under the setting when market microstructure noise is a parametric function of the trading information, includes but not limit to trade type, trade volume and bid-ask spread. We propose a new estimator of the leverage effect (NLE) in the general caseto take account into auxiliary information of microstructure noiseto improve the existed estimator. The proposed estimator provides a faster converge rate n^{1/4} than the estimator in Wang and Mykland (2014) (WLE) with additive noise, which has the same rate as WLE without noise. The results of simulation and proofs demonstrate that the proposed estimator with market noise have little difference from WLE without market noise.

 

 

生存分析简介及研究现状

孙六全(中国科学院/广州大学)

摘要:简要介绍生存分析和生物统计中出现的一些基本概念和一些删失数据,包括:右删失数据,截断数据,区间删失数据,纵向数据和复发事件数据等。最后介绍其研究现状。

 

 

Mean Empirical Likelihood

梁薇(厦门大学)

Abstract: Empirical likelihood methods are widely used in different settings to construct the confidence regions for parameters which satisfy the moment constraints. However, the empirical likelihood ratio confidence regions may have poor accuracy, especially for small sample sizes and multi-dimensional situations. In this paper, we propose a novel Mean Empirical Likelihood (MEL) method. This new method constructs a new pseudo dataset using the means of observation values to define the empirical likelihood ratio and we prove that this MEL ratio satisfies the Wilks' theorem. Simulations with different examples are given to assess its finite sample performance, which shows that the confidence regions constructed by Mean Empirical Likelihood is much more accurate than that of the other Empirical Likelihood methods.

 

 

 

Bayesian Detection of Event Spreading Pattern from Multivariate Binary Time Series

胡杰(厦门大学)

Abstract: This report concerns the inference of event spreading path from noisy event occurrence time series. We propose a probabilistic tree model for the event spreading path, and provide a full Bayesian approach with the Metropolis-Hastings algorithm to detect the tree path from multivariate binary time series. Our approach allows a node spreading the event in a probabilistic way to an arbitrary number of downstream nodes, and allows the input data to contain nodes which are not inside the event spreading path. Simulations on large scale synthesized data sets show that our method can accurately detect the event spreading topologies and have good one-step-ahead prediction accuracy. Our method outperforms existing approaches such as greedy algorithms and the genetic algorithm. We also apply our method to the PM2.5 data in provincial capitals of North and East China. The inferred path is largely consistent with the spreading trend recognized by the public.