🏁 Annual Kick-Off 🗓️

NO COST & PARKING FREE
PLUS Lunch will be provided with a vegetarian option
Stick around for the business meeting to learn about next year's initiatives! Please share this event with a friend: meetup.com/sc_asa

Keynote by Dr. Shujie Ma, Department of Statistics, UCR:
"A Robust and Efficient Approach to Causal Inference Based on Sparse Sufficient Dimension Reduction"

Students Presenters:
Ms. Mingming Liu, graduate student, UCR (Title: Subgroup analysis in longitudinal data)

Mr. Gordon David Mosher, undergraduate student, UCR (Title: Prior Weights for Iterative Hard Thresholding: Using Julia to analyze the human genome)

Schedule:
9am-10:15am : Keynote Presentation (Dr. Shujie Ma)
10:15am-10:30am : Coffee Break
10:30am-11am : Presentation by Mingming Liu
11am-11:30am : Presentation by Gordon David Mosher
11:30am-12:30pm : Lunch
12:30pm-2pm : SCASA Business meeting

About:

KEYNOTE: A Robust and Efficient Approach to Causal Inference Based on Sparse Sufficient Dimension Reduction (by Dr. Shujie Ma)

A fundamental assumption used in causal inference with observational data is that treatment assignment is ignorable given measured confounding variables. This assumption of no missing onfounders is plausi-ble if a large number of baseline covariates are included in the analysis, as we often have no prior knowledge of which variables can be important confounders. Thus, estimation of treatment effects with a large number of covariates has received considerable attention in recent years. Most of the existing meth-ods require specifying certain parametric models involving the outcome, treatment and confounding vari-ables, and employ a variable selection procedure to identify confounders. However, selection of the right set of confounders depends on correct specification of the working models. The bias due to model mis-specification and incorrect selection of confounders can yield misleading results. Weproposes a new ro-bust and efficient approach for inference about the average treatment effect via a flexible modeling strate-gy incorporating penalized variable selection. Specifically, weconsider an estimator constructed based on an efficient influence function which involves a propensity score function and an outcome regression func-tion. We then propose a new sparse sufficient dimension reduction approach to estimating these two functions, without making restrictive parametric modeling assumptions. We show that the proposed esti-mator of the average treatment effect is asymptotically normal and semiparametric efficient.

STUDENT TOPIC: Subgroup analysis in longitudinal data (by Ms. Mingming Liu)
Understanding treatment heterogeneity in longitudinal studies is very important to the development of precision medicine, which seeks to tailor medical treatments to subgroups of patients with similar charac-teristics. One of the challenges is that we usually do not have a priori knowledge of the grouping infor-mation of patients with respect to treatment effect. To solve this problem, we propose a penalized ap-proach for subgroup analysis based on either a GEE model or a nonparametric mixed-effects model.

STUDENT TOPIC: Prior Weights for Iterative Hard Thresholding: Using Julia to analyze the human genome (by Mr. Gordon David Mosher)
This summer I was privileged to participate the “Bruins-In-Genomics” Summer Undergraduate Re-search Program at UCLA where I was selected to work in the Human Genetics Department in the Geffen School of Medicine. I learned to program in Julia, which is designed to be extremely fast and to handle the large data sets common in modern genomics analyses. I used my new-found Julia skills on a project in support of the multivariate genome-wide association studies component in OpenMendel, an open source statistical ge-netics package for analysis of qualitative and quantitative traits.

Read more @