Projects
Risk Analysis Research Center has been proceeding with the following seven projects in basic and applied fields.
Data Infrastructure for Risk Analysis
Project LeaderKazuhiro Minami
We are engaged in various activities to establish a data-sharing platform for risk science research; we collect and link datasets in various domains, such as healthcare, social science, economics, and environmental studies. We focus particularly on studying privacy-preserving techniques, such as anonymization and statistical disclosure control, in order to provide researchers with a data-editing tool to mask sensitive information on individuals and organizations. We conduct our research projects in collaboration with the Joint Support-Center for Data Science Research at the Research Organization of Information and Systems (ROIS).
Mathematical Risk Analysis
Project Leader
Satoshi Kuriki
To quantify the risk factors of natural disasters, serious illnesses, or major accidents, which do not occur frequently but entail serious damage when they do, such events need to be mathematically formulated, and the tail behaviors of their distributions should be examined statistically. In this project, we conduct research on the field of mathematics and computational methods dealing with the tail region of distribution, such as extreme value theory, copula theory, and multiple comparisons. In addition, we investigate mathematical statistics related to risk management in general, including stochastic process theory for complicated data structures and applications of semiparametric theory. We also aim to promote research exchanges with researchers in Japan and overseas by holding international and domestic workshops.
Environmental Risk Analysis
Project Leader
Koji
Kanefuji
The impact of human activity on the global environment is increasing. Thus, quantitative methods to accurately take stock of the environmental situation are becoming increasingly important to implement effective measures for the next generation. In this project, we conduct research on statistical analysis methods which form the basis of environmental risk assessments for water, air, and soil, environmental monitoring, setting of environmental standards, and many other activities. In close, cross-disciplinary cooperation with environmental science fields, we also aim to provide quantitative analysis and assessment methods targeting various issues concerning the Earth’s environment.
Risk Analysis for Resource Management
Project Leader
Atsushi Yoshimoto
Through growth and production processes, renewable resources including those from forests, agriculture, and fisheries are not used only to produce direct market goods; they also produce non-market goods related to ecosystem services by controlling the intensity, timing, and location anthropogenic activities. In this project, we focus on risk analysis of natural resource management through a sustainable socioeconomic system, by integrating this work with the implementation of field surveys intended to predict and control natural resources. We develop deterministic and stochastic mathematical models to capture natural resource dynamics, considering changes in socioeconomic activities and optimization models to solve resource management problems. Through the use of these models, we seek a better policy on a more efficient and effective utilization of natural resources.
Statistical Seismology
Project Leader
Jiancang Zhuang
The statistical seismological research group develops statistical models for quantitative analysis of earthquake occurrences and the relationship between seismicity and other geophysical phenomena, geochemical observations, techniques of probabilistic earthquake forecasting, and methods for evaluating forecasts. With an interest in societal implementation, our research topics also include probability earthquake early warning, optimization of observation seismic/geodetic networks, and earthquake insurance. In interdisciplinary exchanges in risk-related sciences, we aim to promote the application of hazard-intensity based techniques, including modelling, inference, and forecasting, in the study of occurrences of more general random events. Additionally, we aim to develop new hazard intensity models by understanding the risk structures such as the underlying causal relationships.
Spatio-Temporal Statistical Modeling and Applications
Project Leader
Daisuke Murakami
We conduct research on the evaluation of spatio-temporal characteristics such as spatial correlation and temporal correlation, as well as modeling, factor analysis, and uncertainty assessment, with a focus on geographical phenomena. Based on these findings, we carry out applied researches aimed at addressing a wide range of social issues, relating the environment, diseases, and the economy.
Computational Risk Analysis
Project Leader
Mirai Tanaka
With the recent accumulation of huge, diverse, and complex data everywhere, computational inference using such big data has become an effective approach in modern risk analysis. The key to this approach is advanced computational methods based on mathematics. In this project, we study computational inference techniques using big data and their applications to risk analysis, as well as the underlying computational methods and related mathematics. Examples of the former are sparse modeling, uncertainty quantification, Bayesian inference in high or infinite dimensions, statistical methods using nonlinear models, etc. Examples of the latter are computational Bayesian statistics, mathematical optimization, etc.