Linear Regression and Calibration

When a linear relationship is observed between two variables, the correlation is quantified by a method such as linear least-squares regression. This method determines the equation for the best straight line that fits the experimental data. 

The line has the form of the Eq. 16.12, were x and y are variables, a is the slope and b is the intercept on the v axis. Linear regression determines the best values of a 

A linear relationship between a measurable parameter (like absorbance) and concentration is observable in many analytical methods, and the least-squares method is used to fit the best calibration curve. 

Normally, a minimum of 6-8 points are necessary to ensure the linearity of the calibration curve and to calculate the regression parameters. In this case, the following equations are used to calculate a and b values and their confidence limits. We assume that a s and that a is independent of x. 

The correlation coefficient r, or more frequently r2, is used to calculate the quality of the calibration line obtained (Eq. 16.21). A value of 1 indicates a perfect 

---