PCA - Advanced Tab

Select the Advanced tab of the PCA dialog box to access the options described here.

Element Name Description
Variable scale The following four options are in the Variable scale group box.
Do not scale variables Select this check box if you do not want to scale the continuous variables (generally not recommended). Whether you use scaling or not, continuous variables are always zero-mean centered since it is a requirement of the PCA analysis.
Unit standard deviations Select this option button to scale the continuous variables to unit standard deviation.
User-defined standard deviations Select this option button to define your own scaling parameter (i.e., standard deviation) for the analysis variables using the Define standard deviations button described below.
Define standard deviations Click this button to display the User-defined scale (standard deviation) dialog box where you can specify the individual scaling of each variable.
Note: It is good practice to scale your analysis variables, particularly when different variables have typical values that differ significantly. This is often the case when different variables pertain to different quantities (e.g., temperature and pressure) or when they are measured in different units (e.g., inches and miles). In cases such as this, independent variables with typically large values will bias the analysis outcome, which may then lead to poor predictions and performance. STATISTICA PCA resolves this problem by applying a linear transformation, which pre-processes the variables to zero mean and unit (default) or a user-defined standard deviation.
Variable blocks The following two options are in the Variable blocks group box. Blocks of data consist of a number of similar variables. Use the options in this group box to assign variables of similar type to the same block. Variables assigned to the same group will be scaled to the same standard deviation. You can use block scaling when variables can naturally be divided into different types. Thus, instead of scaling each variable individually, all member variables of a block will be scaled to a common standard deviation.
Variable blocks Click this button to display a dialog box for Variable block assignment. The dialog box displays variable indices, names, types, and block memberships. To assign a variable to one of the specified blocks, select the variable name, change the block number in the block box, and click the Apply button. You can also make multiple block assignments by selecting more than one variable and repeating these steps. Click OK to make your block assignment permanent. To reject it click the Cancel button.
Maximum number of user-defined blocks Specify here the maximum number of user-defined blocks that you can assign. By default, all selected variables will be assigned to the default block 0 (zero). Use the Variable blocks option to select the variables, from among those chosen for the analyses, that you want to assign to the same block. Note that for all variables in the same user-defined block (i.e., other than block 0), the total (combined) variance for the analyses is equal to 1.0, while for each variable in the default block 0, it is 1.0. These options are particularly useful in order to account for the indicator coding of categorical variables, which yields as many (0/1 - coded) indicator variables as there are categories or classes found in the respective variable. By assigning a categorical variable to a unique user-defined block, you in effect assign to the combination of all coded indicator variables the same "importance" for the analyses as that of each variable in the default block 0.