Statistical Data Analysis
We conduct education and research on statistical theories and methodologies required to analyze data on various issues in the environmental and life sciences, natural and social sciences.
Statistical science is a fundamental technology of data analysis and machine learning, and provides the most effective means of presenting an objective view based on scientific evidence. We wish to contribute to solving various issues, with making use of drastically improving computer ability.
We study statistical models for analyzing complex phenomena observed in environmental and life sciences, as well as methods and criteria for selecting the optimal statistical model. For example, we investigate how to calculate efficiently estimates and selection criteria for regularized regression models, and whether the proposed methods have good statistical performance.
Recently, observed data have become increasingly large and complex, and we are working on the development of statistical methods called multivariate analysis to analyze them efficiently. For example, we are developing methods to identify data with heterogeneous behavior and to visually interpret data.