化学建模与模型集群分析

摘要

本文简单介绍了化学建模与模型集群分析的思想，并列举了基于模型集群分析的思路与框架。近年来，应用于化学建模各个方面的许多新算法包括奇异样本诊断、变量选择、模型参数与评价、稳健与模型应用域。本文通过应用于不同的数据类型，包括近红外光谱、定量构效关系及代谢组学数据，举例阐述模型集群分析方法的可行性与应用性，为未来开发化学建模新算法提供一个好的思路和框架。%In this review, the concept and idea of chemical modeling and model population analysis ( MPA) were introduced, and the recent applications of MPA-based methods to different aspects of chemical modeling were listed, including outlier detection, variable selection, model evaluation and applicability domain. In addition, the feasibility and applicability of MPA to different kinds of dataset was illustrated, such as near infrared spectroscopy, quantitative structure-activity relationship and metabolomics, which provided a better idea and framework to develop a new algorithm in chemical modeling.