If your nonlinear model contains one predictor, Minitab displays the fitted line plot to show the relationship between the response and predictor data. The plot includes the regression line, which represents the regression equation. You can also choose to display the 95% confidence and prediction intervals on the plot.
Use the regression equation to describe the relationship between the response and the terms in the model. The regression equation is an algebraic representation of the regression line. Enter the value of each predictor into the equation to calculate the mean response value. Unlike linear regression, a nonlinear regression equation can take many forms.
For nonlinear equations, determining the effect that each predictor has on the response can be less intuitive than it is for linear equations. Unlike the parameter estimates in linear models, there is no consistent interpretation for the parameter estimates in nonlinear models. The correct interpretation for each parameter depends on the expectation function and the parameter's place in it. If your nonlinear model contains only one predictor, assess the fitted line plot to see the relationship between the predictor and response.
If you need to determine whether a parameter estimate is statistically significant, use the confidence intervals for the parameters. The parameter is statistically significant if the range excludes the null hypothesis value. Minitab cannot calculate p-values for parameters in nonlinear regression. For linear regression, the null hypothesis value for every parameter is zero, for no effect, and the p-value is based on this value. However, in nonlinear regression, the correct null hypothesis value for each parameter depends on the expectation function and the parameter's place in it.
For some data sets, expectation functions, and confidence levels, it is possible that one or both confidence bounds may not exist. Minitab indicates missing results with an asterisk. If the confidence interval has a missing bound, a lower confidence level might produce a two-sided interval.
Convergence on a solution does not necessarily guarantee that the model fit is optimal or that the sum of squared errors (SSE) are minimized. Convergence on incorrect parameter values can occur due to a local SSE minimum or an incorrect expectation function. Therefore, it is crucial to examine the parameter values, fitted line plot, and residual plots, to determine if the model fit and parameter values are reasonable.
To determine how well the model fits your data, examine the statistics in the Model Summary table and the Lack of Fit table.
Use S to assess how well the model describes the response.
S is measured in the units of the response variable and represents how far the data values fall from the fitted values. The lower the value of S, the better the model describes the response. However, a low S value by itself does not indicate that the model meets the model assumptions. You should check the residual plots to verify the assumptions.
Minitab automatically displays the Lack of Fit table when your data contain replicates. Replicates are multiple observations with identical predictor values. If your data do not contain replicates, it is impossible to calculate the pure error that is required to perform this test. Different response values for replicates represent pure error because only random variation can cause differences between the observed response values.
If the p-value is larger than the significance level, the test does not detect any lack-of-fit.
Source | DF | SS | MS | F | P |
---|---|---|---|---|---|
Error | 229 | 1.53244 | 0.0066919 | ||
Lack of Fit | 228 | 1.52583 | 0.0066922 | 1.01 | 0.679 |
Pure Error | 1 | 0.00661 | 0.0066125 |
Iterations | 15 |
---|---|
Final SSE | 1.53244 |
DFE | 229 |
MSE | 0.0066919 |
S | 0.0818039 |
In these results, S indicates that the standard deviation of the distance between the data values and the fitted values is approximately 0.08 units. The p-value for the lack-of-fit test is 0.679, which provides no evidence that the model fits the data poorly.
Use the residual plots to help you determine whether the model is adequate and meets the assumptions of the analysis. If the assumptions are not met, the model may not fit the data well and you should use caution when you interpret the results.
For more information on how to handle patterns in the residual plots, go to Residual plots for Nonlinear Regression and click the name of the residual plot in the list at the top of the page.
Use the residuals versus fits plot to verify the assumption that the residuals are randomly distributed and have constant variance. Ideally, the points should fall randomly on both sides of 0, with no recognizable patterns in the points.
Pattern | What the pattern may indicate |
---|---|
Fanning or uneven spreading of residuals across fitted values | Nonconstant variance |
Curvilinear | A missing higher-order term |
A point that is far away from zero | An outlier |
A point that is far away from the other points in the x-direction | An influential point |
Use the normal probability plot of the residuals to verify the assumption that the residuals are normally distributed. The normal probability plot of the residuals should approximately follow a straight line.
Pattern | What the pattern may indicate |
---|---|
Not a straight line | Nonnormality |
A point that is far away from the line | An outlier |
Changing slope | An unidentified variable |