PCA/Factor Node Model Options
Model name You can generate the model name automatically based on the target or ID field (or model type in cases where no such field is specified) or specify a custom name.
Use partitioned data. If a partition field is defined, this option ensures that data from only the training partition is used to build the model.
Extraction Method. Specify the method to be used for data reduction.
- Principal Components. This is the default method, which uses PCA to find components that summarize the input fields.
- Unweighted Least Squares. This factor analysis method works by finding the set of factors that is best able to reproduce the pattern of relationships (correlations) among the input fields.
- Generalized Least Squares. This factor analysis method is similar to unweighted least squares, except that it uses weighting to de-emphasize fields with a lot of unique (unshared) variance.
- Maximum Likelihood. This factor analysis method produces factor equations that are most likely to have produced the observed pattern of relationships (correlations) in the input fields, based on assumptions about the form of those relationships.
- Principal Axis Factoring. This factor analysis method is very similar to the principal components method, except that it focuses on shared variance only.
- Alpha Factoring. This factor analysis method considers the fields in the analysis to be a sample from the universe of potential input fields. It maximizes the statistical reliability of the factors.
- Image Factoring. This factor analysis method uses data estimation to isolate the common variance and find factors that describe it.