PCA - NIPALS Tab

Select the NIPALS tab of the PCA dialog box to access options to configure the settings of the NIPALS algorithm (Geladi and Kowalski, 1986), including number of iterations and the convergence criterion. The NIPALS algorithm (see PCA and PLS Technical Details) extracts the principal components in a sequential manner and is particularly suited for handling problems with a large number of variables.

See also, NIPALS Overview and NIPALS Technical Notes.

Element Name Description
NIPALS algorithm The following options are in the NIPALS algorithm group box:
Maximum number of iterations STATISTICA PCA uses a state-of-the-art algorithm known as NIPALS, which extracts the principal components sequentially in a finite number of steps. With this option, specify the maximum number of iterations the NIPALS algorithm is allowed to take for extracting a single component. Note that the actual number of iterations the algorithm may take to find a component might actually be less than the maximum number specified depending on the setting of the Convergence criterion option (see below).
Convergence criterion Specify a minimum improvement (drop) in error that must be made; if the rate of improvement drops below this level, iterations are terminated. The default value is 0.0001, but you can change this setting to suit your analysis best. Note that if you set the convergence to zero, the algorithm will continue until the maximum number of iterations is reached.