报告题目：A Kernel Tweedie Compound Poisson Model
报告摘要： The Tweedie GLM is a widely used method for predicting insurance premiums and loss reserving. However, the structure of the logarithmic mean is restricted to a linear form in the Tweedie GLM, which can be too rigid for many applications. The growing applications of Tweedie models motivate us to develop a much more flexible nonparametric Tweedie models in a reproducing kernel Hilbert space. The resulting estimator is called Ktweedie, which has multiple advantages over the classical Tweedie GLM by incorporating nonlinearity, nonadditivity, and complex interactions in the final estimator. We develop an efficient algorithm for solving the entire solution path of Ktweedie. Extensive simulations are conducted to show the very competitive finite sample performance of Ktweedie. We further demonstrate the application of Ktweedie by using rate making data and loss reserving data.
Lian Yi, 加拿大麦吉尔大学生物统计学博士，于麦吉尔大学获统计学和药理学本科学位和流行病学硕士学位。主要研究方向为高维统计、稀疏统计学习、机器学习、凸优化，以及统计方法在医疗健康、药物安全、生物信息、保险精算等领域的应用。拥有生物统计分析员及保险公司数据科学家的实习经历。并在 EUROPEAN UROLOGY、Ecotoxicology And Environmental Safety 等国际知名期刊发表多篇高质量文章。
报告题目：A sparse high dimensional generalized varying coefficient model for identifying genetic variants associated with regional methylation level
报告摘要： Varying coefficient models offer the flexibility to learn the dynamic changes of regression coefficients. Despite their good interpretability and diverse applications, in high-dimensional settings, existing estimation methods for such models have important limitations. For example, we routinely encounter the need for variable selection when faced with a large collection of covariates with nonlinear/varying effects on outcomes, and no ideal solutions exist. One illustration of this situation could be identifying a subset of genetic variants with local influence on methylation levels in a regulatory region. To address this problem, we propose a composite sparse penalty that encourages both sparsity and smoothness for the varying coefficients. We present an efficient proximal gradient descent algorithm to obtain the penalized estimation of the varying regression coefficients in the model. A comprehensive simulation study has been conducted to evaluate the performance of our approach in terms of estimation, prediction and selection accuracy. We show that the inclusion of smoothness control yields much better results than having the sparsity-regularization only. Using an adaptive version of our penalty function, we can achieve notable additional performance gains. The method has been implemented in R package sparseSOMNiBUS available on GitHub
杨羿，现任加拿大麦吉尔大学 (McGill Uiniversity) 数学与统计学系副教授 (Associate Professor)，计算机系兼职教授，计量生物学项目成员，2008-2015年就读于美国明尼苏达大学，获得统计学博士学位，计算机与统计学双硕士学位。主要研究领域为统计机器学习与数据挖掘，统计计算，高维统计推断，及统计学方法在生物信息学，医学，精算学上的应用。已在Journal of the American Statistical Association、Biometrika等统计学顶级期刊发表多篇高质量文章。