Developments of techniques for multivariate data analysis

点击翻译按钮获取中文摘要

We all learn from experiences in everyday life and make predictions for the future. Statistics is an area of study in which we systematically investigate optimal ways of learning from experiences (data) and make rational predictions. Ian Hacking, a philosopher at the University of Toronto, once noted that statistics, as a means of empirical science, was one of the greatest inventions of the 20th century, comparable to the discovery of the theory of relativity by Einstein. I propose the following projects for further advancements of statistics. Each may represent a small step. Accumulation of these small steps, however, goes a long way toward the goal of statistics: 1. Regularized multivariate analysis 2. Information integration from multiple sources 3. Prediction of an event from another 4. Acceleration of learning processes 5. Representation of the relationships among variables