Dynamic Biostatistics : Modeling, Development of Methodology and Theory

1.提出了一种估计李雅普诺夫指数的方法来检测观测数据中的混沌，并证明了其一致性。2.开发了复杂数据的风险评估方法，并将其应用于研究环境污染物对日本母乳喂养婴儿甲状腺激素状况的联合影响。3.为小鼠淋巴瘤试验和细胞转化试验的体外毒性研究开发了一种新的数据分析程序。4.提出了一种基于径向基函数网络的模式识别机的构造方法，该方法可以方便地应用于高维连续和不连续数据的分析。5.建立了基于基因组不稳定性假说的辐射致癌数学模型。6.基于广义药代动力学理论，考虑污染物的累积测量与其流入和排出体外的速率，二恶英化合物在居民外周血中的浓度的随机效应模型的开发。7.建立了cDNA微阵列基因表达数据的数学模型，并提出了一种有效的数据归一化方法。（在模型和方法上申请了两项专利。）8.扩展了我们的分层模型的变化率的分子进化，以适应多个位点，并提出了一个程序来测试基因座之间的相关进化。9.考虑连锁相位的不确定性，开发了一种遗传算法对远交群体进行QTL分析。10.修正了边际结构模型中的权系数，对标准化死亡率进行了非参数估计，表明了标准化方法与边际结构模型之间的一般关系。

1. Developed a method of estimating Lyapunov exponent to detect chaos in observational data and proved its consistency. 2. Developed a method of risk assessment for complex data and applied it for studying joint effects of environmental contaminants on thyroid hormone status of breast fed infants in Japan. 3. Developed a new procedure of data analysis for the mouse lymphoma assay and the cell transformation assay in vitro toxicity studies. 4. Developed a technique for constructing pattern recognition machine based on radial basis function network that can be easily applied to analyze the high-dimensional continuous and discontinuous data. 5. Developed a mathematical model of radiation carcinogenesis based on the genomic instability hypothesis. 6. Based on a generalized phamacokinetic theory that accounts for accumulated measurements of contaminants with rates of their inflow into and elimination out of a body, a random effects model for concentration of Dioxin compounds in peripheral blood of residents is developed. 7. Developed a mathematical model of cDNA microarray data for gene expressions, and proposed an efficient method for normalization of the data. ( applied two patents on the model and method.) 8. Extended our hierarchical model of change in rate of molecular evolution to accommodate multiple loci and proposed a procedure to test correlated evolution between loci. 9. Developed a genetic algorithm to do QTL analysis for outbreeding populations, taking uncertainty of linkage phase into account. 10. Modified weights in marginal structural models to estimate the standardized mortality ratio nonparametrically and showed general relationship between standardization methods and marginal structural models.