Speaker: Kate Hu
Title: Use Auxiliary Data to Improve Statistical Inference
Abstract: The open data revolution has substantially increased the availability and accessibility of auxiliary data. With appropriate tools, this auxiliary information can improve statistical inference at almost no cost. This talk focuses on my research from three directions regarding the intelligent use of auxiliary information. In the first part, I describe my research results on enhancing the estimation precision in semiparametric inference with two-phase epidemiological studies by tapping into often overlooked auxiliary information embedded in phase I samples. In the second part, I summarize how I used auxiliary data from stationary monitoring sites to design an air quality monitoring program with mobile sensing platforms at scale. In the third part, I introduce my ongoing research on adjusting for unmeasured and mismeasured confounders by leveraging negative control variables from auxiliary data. These three research stories demonstrate how statisticians can provide scientists with new tools to harness the potential of auxiliary data for improving the validity and efficiency of their scientific studies. In the end, I will introduce the new data and complex scientific problems in precision agriculture, opportunities for statisticians, my future research goal to connect precision agriculture to precision health, and how my current research paves the way toward this goal.
Note: Seminars are free and open to the public. Reception to follow.
