从粗糙模型中提取可证伪预测
定量方法
2007-11-24 v1
摘要
成功的预测是对任何模型最有力的验证之一。从非线性多参数模型中提取可证伪预测是复杂的,因为这类模型通常具有粗糙性(sloppiness),即对不同参数组合的敏感程度跨越多个数量级。本文讨论粗糙性如何影响最能约束模型预测的数据类型、如何使线性不确定性近似变得危险,以及如何在蒙特卡洛不确定性分析中引入计算困难。我们还提出了一个有用的测试问题,并建议改进模型交流的标准。
引用
@article{arxiv.0704.3049,
title = {Extracting falsifiable predictions from sloppy models},
author = {Ryan N. Gutenkunst and Fergal P. Casey and Joshua J. Waterfall and Christopher R. Myers and James P. Sethna},
journal= {arXiv preprint arXiv:0704.3049},
year = {2007}
}
评论
4 pages, 2 figures. Submitted to the Annals of the New York Academy of Sciences for publication in "Reverse Engineering Biological Networks: Opportunities and Challenges in Computational Methods for Pathway Inference"