Time Series Analysis and Computing, and Robust Statistical Methods for Modeling Serially Correlated Data

点击翻译按钮获取中文摘要

My research program is to develop robust methods in testing for temporal trends or shifts where the data are noisy and serially correlated. It will also engage in incorporating nonparametric rank-based distance with multiple endpoints and applying it to population comparison studies. The topics are motivated by my collaborative activities with scientists in environmental science, biology, and health outcomes research. The methods proposed for time series trend analysis will provide new ways for scientists to discover changes in our water resources, wildlife populations, and food protein processes. The methods proposed for time series intervention analysis will lead to improved designs for more effective socio-economic intervention programs, especially for public drug policy intervention programs. The robust methods for multiple endpoints will allow researchers to identify patients' true behaviour patterns through patient-reported outcomes. Among other topics my proposed research will also be dedicated to powerful computer algebra algorithms for resolving difficult inference questions. Overall the proposed methods as general statistical methodologies will play a significant role for researchers in answering important questions in science and the social sciences.