论文部分内容阅读
至今大多数分析方法都要通过对分析信号校正来计算被测组分含量,因此,分析信号校正是分析中的重要问题.一元线性回归是信号校正的常用方法.1 绝对偏差和相对偏差最小二乘回归设有普通变量Y_i,随机变量X_i,回归方程形式为Y=a+bX,用绝对偏差最小二乘法拟合校准线,即求Σ((?)_i-Y_i)~2为最小值时a,b值,由方程组解得 计算简便,绘图直观是绝对偏差最小二乘法拟合校准线的一大优点,故在实际中得到广泛应用.但
So far, most analysis methods need to calculate the content of the tested components by analyzing the signal calibration, therefore, analyzing the signal calibration is an important issue in the analysis.Unit linear regression is a common method of signal calibration.1 The absolute deviation and the relative deviation minimum The common variable Y_i and the random variable X_i are obtained by regression and the regression equation is Y = a + bX. The calibration curve is fitted by the least square deviation of absolute deviation, that is, when Σ ((?) _i-Y_i) a, b value, which is easy to calculate by the system of equations. The intuition of drawing is one of the advantages of Least squares method of absolute deviation fitting calibration line, so it is widely used in practice.