Statistical Methods on Challenging Issues of Biosciences

点击翻译按钮获取中文摘要

It is difficult to overemphasize the importance of statistical science to the development of the natural sciences and engineering. Advances in virtually every field, including advanced technologies, environmental science, manufacturing, agriculture, and life sciences, ultimately depend on statistical developments. Novel methods in statistical and biostatistical sciences occupy a central role in much of the modern sciences, finance, economics and life sciences that heavily rely on drawing information from massive data. As sciences and engineering advance, new features of data emerge and traditional statistical methods become inadequate to handle those challenges. Existing methods have often been developed under the "ideal" assumption that data are "perfect". In reality, however, data from which scientific inference is drawn are far from "perfect". Measurement error and missing values are ubiquitously associated with data and they typically break down usual statistical methods. Although the growing literature and emerging lexicon have raised awareness of the critical impact that "imperfect" data can have on statistical inference and applications, effective methods on handling "imperfect" data remain elusive. The primary objective of my research is to develop innovative and rigorous statistical methodology to tackle various challenging problems induced by "imperfect" data. Flexible modeling strategies and valid inference methods will be developed to address fundamental issues concerning measurement error and missing observations. The proposed research will provide new insights and valuable results that will benefit both researchers and industrial partners. The anticipated benefits include a considerably marked effect on both enhancing methodological standards and providing a broad range of applications in highly impacted fields. This will lead a significantly direct and positive economic impact on our society.