演化自适应建模在QSAR研究中的应用  

APPLICATION OF THE GENETIC SELFADAPTING MODELING IN THE QSAR ANALYSIS

作　　者：张海月[1] 郑绍辉[1] 张明涛[1] 林少凡[1] 

机构地区：[1]南开大学中心实验室,天津300071

出　　处：《南开大学学报（自然科学版）》2003年第4期72-78,共7页Acta Scientiarum Naturalium Universitatis Nankaiensis

摘　　要：演化自适应建模利用自然选择的方法处理建模和模型优化问题,通过在模型空间中搜索问题的解来揭示输入与输出数据间存在的相关关系并以此来估计和外推系统的行为,该方法用于QSAR研究,具有稳定、适应性强、无需问题的背景知识等优点,可以获得多个较为准确描述药物构效关系的模型,既可以是线性模型,也可以是非线性模型,但对于自变量较多的问题,该方法的编码方式可能使效率下降。Genetic self-adapting modeling algorithm process the couple of input and output data by means of natural selection to build up an approximate description of the correlativity between them. Genetic self-adapting modeling could be applied in many practical problems without modification. In QSAR analysis, models are essential tools to discover the relationship between chemical structure and biological activity. Genetic self-adapting modeling provide a feasible method to search and optimize models, which can be used in QSAR research and get satisfactory result. Models are searched and selected in the runtime, without the manual intervention. Commonly, several available models will be offered for the same problem. Each of these models can be used for estimating and extrapolating the result of the system.

关 键 词：自适应建模 定量构效关系分析 卡巴醌 

分 类 号：O6-39[理学—化学]