标题： 模型比较的微分代数 显示英文标题

标题： Differential Algebra for Model Comparison

摘要： 我们提出了一种从噪声时间序列数据中拒绝竞争模型的方法，该方法不依赖于参数推理。 首先，我们使用微分代数消去法，仅用可测量的变量来刻画常微分方程模型。 接下来，我们利用高斯过程回归（GPR）从给定数据中提取额外信息，然后变换微分不变量。 我们利用线性代数和统计学开发了一种测试方法，以无参的方式用给定数据拒绝变换后的模型。 此算法利用了嵌入模型结构中的关于瞬态的信息。 我们通过区分数学生物学中的不同模型，展示了这种方法的强大功能。 显示英文摘要

摘要： We present a method for rejecting competing models from noisy time-course data that does not rely on parameter inference. First we characterize ordinary differential equation models in only measurable variables using differential algebra elimination. Next we extract additional information from the given data using Gaussian Process Regression (GPR) and then transform the differential invariants. We develop a test using linear algebra and statistics to reject transformed models with the given data in a parameter-free manner. This algorithm exploits the information about transients that is encoded in the model's structure. We demonstrate the power of this approach by discriminating between different models from mathematical biology.