The smaller the SSq, the closer the observed values are to the predicted, the better the model predicts your data. To get around this you should square each of the residuals, which render all the values positive, then sum them. Summing these is not very useful, as even a random set of data points may generate residuals that sum close to zero. However, since some observed values will likely be above the fitted curve and some below you will get positive and negative residuals. The smaller the sum the better the data fit the predicted curve.
To check the predicted fit of the line one usually calculates all the residuals observed - predicted and sums all the differences.
The goal is to determine values of m and c which minimize the differences residuals between the observed values i. Once the standard curve is generated it is relatively easy to see where on the curve your sample lies and interpolate a value. You can think of the standard curve as the ideal data for your assay. The standards, in your assay, should be tested at a range of concentrations that yields results from essentially undetectable to maximum signal. In order to determine a quantity of something you will need to compare your sample results to those of a set of standards of known quantities. This is where things can get interesting. Maybe you will even develop your own assay. All you have to do is test the sample using any number of commercially available kits. Please enable JavaScript before continuing otherwise certain functions will not operate correctly.
0 kommentar(er)
