金免疫层析试条OD-浓度曲线的神经动力学拟合

Neural dynamics fitting of the OD-concentration curve of gold immunochromatography strip

针对金免疫层析试条定量检测系统中OD(光密度)-浓度的拟合曲线会出现检测值与实际值偏差较大,容易导致定性结果错误的情况,提出以最大绝对误差最小为评价指标的曲线拟合方案,并转化为相应的约束优化问题使用神经动力学优化算法进行求解。仿真数据实验表明神经动力学曲线拟合方法明显优于插值法和三次样条,与最小二乘法相比在等同条件下50次曲线拟合的平均最大绝对误差降低14%;通过金免疫层析试条定量检测仪的一组标定数据实验表明三次多项式基本符合OD值与浓度正相关关系,且此时神经动力学拟合曲线的最大误差与最小二乘法相比降低25%;实验结果表明该文提出的神经动力学曲线拟合方法结果收敛稳定,且有效降低最大绝对误差,为金免疫层析试条定量检测提供一种新的较简单和精确的曲线拟合方法。

Incorrect qualitative result is easy to be caused by large deviation between detection value and actual value on the OD （optical density） - concentration fitting curve, which may result in inaccurate result of gold immunochromatographic strip qualitative detection system. A curve fitting scheme of taking the maximum absolute error minimum as evaluation index is hereby proposed. Corresponding constrained optimization problem is converted to neural dynamics optimization algorithm for solution. Simulation data experiment shows that neural dynamical curve fitting method is obviously superior to the interpolation method and cubic spline. Compared with the least square method, the average maximum absolute error of 50 - times curve fitting under the same condition is reduced by 14% ; the experimental results of a set of calibration data from quantitative detection instrument of the gold immunoehromatographic strip show that cubic polynomial basically complies with normal phase relation between OD value and concentration, and the maximum error of the fitting curve is reduced by 25 % compared with that of the least square method; the experimental results show that neural dynamical curve fitting method put forward in the paper features stable convergence and the maximum absolute error is reduced effectively; a new curve fitting method for quantitative detection of gold immunochromatographic strip is provided.