数学模型法在RIA数据处理中的应用与分析  

THE ANALYSIS OF MATHEMATIC MODEL METHOD AND APPLICATION OF THE METHOD TO RIA DATA HANDLING

作　　者：赵玉荣[1] 陈耀华[2] 

机构地区：[1]锦州医学院数学教研室 [2]锦州医学院核医学教研室

出　　处：《辽宁医学院学报》1992年第2期1-5,共5页Journal of Liaoning Medical University (LNMU) Bimonthly

摘　　要：本文采用数学模型法取代手工作图法,对RIA标准曲线函数拟合的几种数学模型(多项式、样条函数、Logit—log、四参数或五参数Logistic及四参数单位点质量作用模型等)逐一进行分析、比较,论证其在RIA应用中的优缺点。分析得出,多项式有较好的适应性,样条函数的可塑性极好,Logit—log计算简便,易在较低层次上推广应用。多项式、样条函数模型纯属数学方法,曲线形态、规律走向完全取决于实验数据,稳定性差,参数不稳定,故逐步被否定。实践表明Logistic模型和四参数单位点质量作用模型是适应性强、稳健性好、合理而有效的模型,乃是放免工作者、统计学家和计算机专家共同合作产生的公认且广为推广的方法。前当用计算器按Logit—log模型法函数式作线性回归计算和反查求解,应是RIA标准曲线处理最低质量要求。This essay describes how the method of drawing a chart by hand was replaced by the method of mathdmatic model, several mathemtic models (multinomial, spline function logit-log, 4. 5—paramater logistic and 4—parameter single binding site mass action model)whith fit RIA calibr ation curve Were analysed and compared, their virtues and defects were discussed when they were applied to RIA. Through analysis we came to the conclusion that rnultinomal had good adcptabitity, spline function had better plasticity and logit-log was simple in Calculation,which was easy to poplarize in lower Lervel Multinomial and spline function models thoroughly belong to mathemaitc method. The forms and rules of the curve were thoroughly determined by experimental data Their stability was too bad and parameters were not stable, so they were denied gradually. The fact showed that logistic model and 4-parameter single binding size mass action model were rational and effective models which had strong adaptability and good stability. Therefore they were generally aknowledged and popularized everywhere This method was created by the joint efforts of radioimmune workers, in experts statiseics and computers At present the use ol calcnlator, acconding to logit-log model method friction to make linear retrace calc ulation and probed into the result again should e the lowest quality required by RIA calibration curve handling

关 键 词：数学模型法 放射免疫分析 标准曲线 函数拟合 LOGISTIC模型 单位点质量作用模型 质量控制 

分 类 号：R[医药卫生]