Research And Implementation Of The Benchmark Dose Assessment System

The benchmark dose (BMD) method is a scientific and effective method, which is used to derive the starting point of the health risk assessment through fitting the experimental dose-response data and mathematical models.First of all, the experimental design is studied. Generally, the experimental dose is uniformly distributed in the dose-response experiments, and then the curve obtained is quite different from the real dose-response trend sometimes, which leads to poor BMD accuracy in dose-response analysis. In this study, a design criterion approach method is used to look for improving the BMD accuracy’s dose distribution and sample layout. The conclusion shows that:the distribution of different doses and sample layout will affect BMD accuracy, when total number of samples in the experiment is under fixed conditions, choosing different model two who played minor roles in different.Secondly, in order to develop the data and model fitting system, the dose-response model and its form are analyzed, including solving dichotomous model unconstrained optimization algorithms Double dog-leg, the improvement of the numerical method of sequential quadratic programming of the problem solving constraint optimization, and likelihood ratio method for confidence intervals calculated, on this basis, the main algorithm gives detailed iterative strategy and numerical implementation steps, and the main algorithm is verified to be correct by an example.Finally, a benchmark dose assessment system is developed. The platform has the experimental design analysis and model fitting analysis capabilities, which can be used for dichotomous data and continuous data analysis to calculate.