Fixed Nonlinear Regression Startup Panel and Quick Tab
Ribbon Bar: Click the Statistics tab. In the Advanced/Multivariate group, click Advanced Models and select Fixed Nonlinear Regression to display the Fixed Nonlinear Regression Startup Panel.
Classic Menus: On the Advanced Linear/Nonlinear Models menu, select Fixed Nonlinear Regression to display the Fixed Nonlinear Regression Startup Panel.
The Startup Panel contains a Quick tab.
Use Fixed Nonlinear Regression to select various transformations of the independent variables; thus, new variables are created (in memory) as transformations of the raw data and included in the overall correlation matrix. A common application of this technique is to estimate the linear, squared, and cubic components of the relationship of an independent variable with the dependent variable. Note that Statistica also includes procedures for Nonlinear Estimation for estimating true nonlinear regression models, for example, probit, logit, discontinuous, and regression models.
For more information, see Fixed Nonlinear Regression Index, Overviews, and Example.
| Option | Description |
|---|---|
| Variables | Displays the standard Variable Selection dialog box in which you can select a maximum of four variables that are used in the regression analysis (a correlation matrix is calculated for all variables that are selected). |
| Extended precision computations | You can select the Extended precision computations check box to use the extended precision algorithm in generating the input correlation matrix that is used for regression calculations. This option is useful if the data follow a very rare pattern of small relative variance. The default double precision calculations performed in fixed nonlinear regression feature precision optimizations and offer superior accuracy and for almost all data sets obtained from measurements (not artificially created), produce results identical to those which can be obtained by using the extended precision option. |
| Review descriptive statistics, correlation matrix | Displays detailed descriptive statistics (and pairwise N, if Pairwise deletion of missing data was requested) for the variables selected. Click OK in the Review Descriptive Statistics dialog box to continue the analysis and display the Model Definition dialog box. |
| OK | Displays the standard Variable Selection dialog box or (if you have already chosen variables) the Nonlinear Components Regression dialog box, which is used to select various transformations of the independent variables. Thus, new variables are created (in memory) as transformations of the raw data and included in the overall correlation matrix. |
| Open Data | Used to choose the spreadsheet to be analysed. The Select Data Source dialog box contains a list of the spreadsheets that are currently active. |
| Select Cases | Displays the Analysis/Graph Case Selection Conditions dialog box, which is used to create conditions for which cases are to be included or excluded in the current analysis. |
| W | Displays the
Analysis/Graph Case Weights dialog box, which is used to adjust the contribution of individual cases to the outcome of the current analysis by weighting those cases in proportion to the values of a selected variable.
Weighted Moments You can select the Weighted moments check box to specify that each observation contributes the value of weighting variable for that observation. The weight values need not be integers. The option is available only after you have defined a weight variable via the W option. This module can use fractional case weights in most computations. Some other modules use case weights as integer case multipliers or frequency values. |
| DF = W-1 / N-1 | Bases some statistics related to the moments, for example, standard deviations and variances, skewness, kurtosis, on the sum of the weight values for the weighting variable if the
W-1 is set or on the number of unweighted observations if the N-1 is set. The sums (and means), and sums of squares and cross products are always based on the weighted values of the respective observations. However, in computations requiring the degrees of freedom, for example, ANOVA table, statistical significance of parameter estimates, the value for the degrees of freedom can either be computed as the sum of the weight values minus one (W-1), or as the number of observations minus one (N-1). When the Weighted moments option is in effect, some graphics options might not be available. Note that these options are only available if Weighted moments is selected.
Weighted Least Squares In some cases you must apply differential weights to the observations in a regression analysis, and to compute the weighted least squares regression estimates. This method is commonly applied when the variances of the residuals are not constant over the range of the independent variable values. In that case, you can apply the inverse values of the variances for the residuals as weights and compute weighted least squares estimates. In practice, these variances are usually not known, however, they are often proportional to the values of the independent variables, and this proportionality can be exploited to compute appropriate case weights. Neter, Wasserman, and Kutner describe an example of such an analysis, which is also discussed in the Nonlinear Estimation Examples. To compute weighted least squares estimates, choose the weight variable, and then select the Weighted moments and N-1 options on the Fixed Nonlinear Regression Startup Panel. |
| MD deletion | Deletes missing data depending on the selection in the MD deletion groupbox. |
| Casewise | Includes only the cases that do not contain any missing data for any of the selected variables in the analysis. |
| Pairwise | Excludes cases from the calculation of correlations involving variables for which they have missing data. In subsequent analyses, all tests of statistical significance in that instance are based on the smallest number of valid cases found in any of the selected variables. |
| Mean substitution | Replaces missing data by the means for the respective variables (for this analysis only, not in the data file). |