Orthogonal array construction and confidentiality in both survey and data-mining contexts

点击翻译按钮获取中文摘要

My research interests for the next few years will be primarily focused on developing adaptive statistical methods and investigating their theoretical properties for the analysis of large and complex datasets.In the context of complex surveys, protecting sensitive information from disclosure has become more and more challenging as the magnitude of such surveys increases. Specific methodologies may involve constructing nearly-orthogonal mixed-level arrays, which by itself is an interesting research problem in the area of modern experimental design, especially when designing input settings for complex computer simulators, as the realistic surrogate to the much more complex physical phenomena, is desired.It is also worth noticing that the issues of maintaining confidentiality in large data warehouses, closely related to the above-mentioned research problem, has drawn significant interests after data-mining techniques have been widely used by government agencies and businesses in processing public access data that also includes sensitive information. In particular, I intend to explore issues relating to algorithmic strategies in the area of privacy-preserving data-mining research and expect the results to be quite influential to industries involved in data-mining.