OCHI Yoshimichi

FacultyNational University Corporation OITA UNIVERSITY
PositionDirector
 

Profile

Faculty

    National University Corporation OITA UNIVERSITY,  Director

Name

    OCHI Yoshimichi

Degree(1)

  • Ph.D., ワシントン大学, Dec. 1983
  • Master of Science, 広島大学, Mar. 1980

Address

    700 Dannoharu Oita-city, Oita 870-1192, Japan, +8197(554)7869, ochioita-u.ac.jp

Research Achievements

Field of Study

    Statistical science, Statistical Science

Research content keyword

    Exact Inference, Over Dispersion, Prallel Computing

Research Title

  • Evaluation of Associations in Categorical Data, This research aims to establish methods to analyze data with categorical responses and to evaluate their relations to associated covariates. There are cases where we can assume some ordering for them and cases where we cannot and analyses depend on such characteristics of the responses. In this research, we consider construction of association measures, modeling them with covariates, their inference procedure, diagnostics and remedies, and so on.
  • Distributed Parallel Computational Environment for Statistical Calculation, In this research, we aim to establish an environment which enables us to use computer resources efficiently in order to carry out statistical data analyses and to develop statistical methods. In current statistical data analyses, we need to deal with more flexible and sophisticated mathematical models to analyze details of complex phenomena than ever, while robust techniques which do not depend on the underlying assumption of those models are urgently on demand. At the same time, we often have a huge data set to analyze as we became acquainted with “data mining.” The power of the computer resource affects heavily on efficiencies of the analyses. In this research, we consider a distributed parallel computation environment to utilize various machines at our hand and so-called “PC-cluster” to face with such computational demand. The key issues involved here are accommodation of statistical computation to the heterogeneous distributed parallel computation environment, load balancing in such situations and re-organization of the computational algorithms themselves.

Papers

  • 2004, Exploring Effects of Dispersion Structure for Over-dispersed Categorical Data on Regression Analysis, Proceedings of IBC 2004 XXIInd International Biometric Conference, 2004
  • 2004, Evaluation of Covariate Effects for Over/Under-dispersed Categorical Data, 計算機統計学, 2004
  • 2006, On development of a courseware of statistics and synchronized web-based statistical analysis system, E-Learn 2006 Proceedings, (Eds. Thomas C. Reeves and Shirley F. Yamashita), 2006
  • 2011, Properties and Effects of Inferences on Hierarchical Generalized Linear Models in Clinical Trial Designs, Abstract of Joint Meeting of the 2011 Taipei International Statistical Symposium and 7th Conference of the Asian Regional Section of the IASC, 2011
  • 2011, Bayesian approach for Hierarchical Generalized Linear Models, Proceedings of Joint Meeting of the Korea-Japan Conference of Computational Statistics and the 25th Symposium of Japanese Society of Computational Statistics, 2011
  • 2012, Evaluating Impact on Application of Bayesian Inference to the Hierarchical Generalized Linear Models in Clinical Trial Design, XXVIth International Biometoric Conference(IBC Kobe 2012), 2012
  • 2012, Oita University, 情報処理, 2012
  • 2015, TREES GARROTE for regression analysis, 36th Annual Conference of the International Society for Clinical Biostatics Abstract book, 2015
  • 2015, Semi-Parametric Approach for Estimating Smoothing Receiver Operating Characteristic Curves with Covariates, East Asia Regional Biometric Conference 2015, Program & Abstract, 2015

Books

  • 2007, Dictionary of Data Science, 朝倉書店, 2007
  • 2007, Methods for Computational Statistics, 朝倉出版, 2008
  • 2009, Dictionary for Medical Statistics, 朝倉書店, 2010

Learned Societies to Belong

  • American Statistical Association

Grant-in-aid for Scientific Research

  • Parallel Distributed Computing for Computer Intensive Statistical Analysis under Heterogeneous Computing Environment, 2002-2004
  • Influence of correlations among categories on statistical inference for multi-categorical discrete distribution, 2007-2008
  • Development of Efficient Statistical Calculation Algorithms under Heterogeneous Parallel Computing Environment, 2009-2011
  • Efficient Statistical Computation via Shared-Distributed Memory Hybrid Parallel Processing, 2012-2014

Collaborative Research and Sponsored Research

  • 2004, 計数データに対する実験計画と解析過程の研究

Social Contribution

Social Activities

  • 2015
  • 2015
  • 2015

International Research Project

  • 2008


Copyright (c) MEDIA FUSION Co.,Ltd. All rights reserved.