4 0 0 0 統計・機械学習による異分野相関を俯瞰する方法論の確立
COVID-19、SDGsの複雑に絡み合った課題に対しては、多くの異分野に跨がる学際的研究が 不可欠であり、異分野相関を俯瞰できる方法論が必要となる。従来の統合評価モデルによるアプローチには、1) 不確かさの所在が不明瞭、2) 多様なデータの十分な活用が困難、3) COVID-19のような突発事象への対応不可等の問題がある。本研究では、統計・機械学習により、この従来モデルをデータ駆動型の確率モデルに転換させて、上記の問題1)、2)を解決する方法論Iを確立する。さらにこの方法論Iを、突発事象を含め、様々な事象を包括できるモデル統合の方法論IIへと変革させ、問題3)を解決する。
4 0 0 0 OA 大橋靖雄先生の追悼記事
2 0 0 0 OA 母集団薬物動態解析の基礎:線形混合効果モデル・非線形混合効果モデルの数理
- 著者
- 船渡川 伊久子 船渡川 隆
- 出版者
- 日本計量生物学会
- 雑誌
- 計量生物学 (ISSN:09184430)
- 巻号頁・発行日
- vol.36, no.Special_Issue, pp.S33-S48, 2015-06-30 (Released:2015-09-08)
- 参考文献数
- 42
Longitudinal data are data collected repeatedly from each subject for a particular response variable over a certain time period. Specifically, in longitudinal data analysis, the researchers are interested in changes in the response levels over time and the differences in these changes among factor levels or covariates. Because of within-subject correlations, analysis methods considering the correlations or variance-covariance structures have been developed. One of the approaches is the use of mixed effects models that take into account between-subject heterogeneity by random effects. In population pharmacokinetics, the response variable corresponds to drug concentration and is analysed typically using nonlinear mixed effects models. In this article, longitudinal data analysis with a continuous response variable is introduced focusing on population pharmacokinetics. Longitudinal data analysis, linear mixed effects models, nonlinear mixed effects models, and population pharmacokinetics are discussed from a biostatistical point of view. This article is expected to be of interest to biostatisticians, pharmacologists, pharmacokineticists, and those in related fields.
1 1 1 0 OA 疫学分野での計量生物学の発展
- 著者
- 田栗 正隆 高橋 邦彦 小向 翔 伊藤 ゆり 服部 聡 船渡川 伊久子 篠崎 智大 山本 倫生 林 賢一
- 出版者
- 日本計量生物学会
- 雑誌
- 計量生物学 (ISSN:09184430)
- 巻号頁・発行日
- vol.44, no.2, pp.129-200, 2024 (Released:2024-04-25)
- 参考文献数
- 167
Epidemiology is the study of health-related states or events in specific populations and their determinants, with the aim of controlling health problems. It encompasses various research fields, such as cancer epidemiology, infectious disease epidemiology, and social epidemiology, molecular epidemiology, environmental epidemiology, genetic epidemiology, clinical epidemiology, pharmacoepidemiology, spatial epidemiology, and theoretical epidemiology, among others, and is closely related to statistics and biometrics. In analytical epidemiological studies, data is collected from study populations using appropriate study designs, and statistical methods are applied to understand disease occurrence and its causes, particularly establishing causal relationships between interventions or exposures and disease outcomes. This paper focuses on five topics in epidemiology, including infectious disease control through spatial epidemiology, cancer epidemiology using cancer registry data, research about long-term health effects, targeted learning in observational studies, and that in randomized controlled trials. This paper provides the latest insights from experts in each field and offers a prospect for the future development of quantitative methods in epidemiology.