faNFactors - WeAreCAS

Q: What is the purpose of the faNFactors action?

The faNFactors action is designed to determine the number of factors for factor analysis.

Q: What criteria can be used to determine the number of factors?

You can specify criteria using the "type" parameter within the "criteria" list. Valid types include "EIGENVALUE" (minimum eigenvalue), "MAP2" (minimum average squared partial correlation), "MAP4" (minimum average fourth-power partial correlation), "PARALLEL" (parallel analysis), and "PROPORTION" (proportion of variance explained).

Q: How does the action decide the final number of factors if multiple criteria are active?

The "nFactors" parameter controls this behavior. It can be set to "MAX", "MEAN", "MEDIAN", or "MIN" (the default) to calculate the final number based on the results of all active criteria.

Q: How can I customize the parallel analysis simulation?

Within the "criteria" parameter list for the "PARALLEL" type, you can set "nSimulations" (defaults to 10,000), "alpha" (critical value, defaults to 0.05), and "seed" for the pseudorandom number generator.

Q: Can I provide my own prior communality estimates?

Yes, by setting the "priors" parameter type to "INPUT". You must then provide the specific estimates using the "values" subparameter.

Q: What does the "threshold" parameter do in the criteria configuration?

The "threshold" parameter specifies cutoff values used only by the "EIGENVALUE" and "PROPORTION" criteria. If not specified, the default value is 1.0.

Q: How can I perform a weighted analysis?

You can specify the "weight" parameter with the name of a numeric variable to use as a weight. Additionally, the "varianceDivisor" parameter can be adjusted to "DF", "N", "WDF", or "WEIGHT" to control how variances and covariances are calculated.

Description

The faNFactors action determines the number of factors to be retained in a factor analysis. It allows you to specify various criteria such as minimum eigenvalue, proportion of variance explained, minimum average partial correlation (MAP), and parallel analysis. The action can combine these criteria using summary statistics like MIN, MAX, MEAN, or MEDIAN to suggest a final number of factors.

Settings

Parameter	Description
attributes	Changes the attributes of variables used in this action (e.g., formats, labels).
corrOut	Specifies an output table to contain the correlation matrix, summary statistics, and number of observations.
criteria	Specifies one or more criteria (EIGENVALUE, MAP2, MAP4, PARALLEL, PROPORTION) to determine the number of factors.
display	Specifies a list of results tables to send to the client for display.
freq	Specifies a numeric variable that contains the frequency of occurrence of each observation.
inputs	Specifies the input variables to use for the analysis.
nFactors	Specifies how to combine multiple criteria to determine the final number of factors (MIN, MAX, MEAN, MEDIAN).
outputTables	Lists the names of results tables to save as CAS tables on the server.
priors	Specifies the method of computing prior communality estimates (e.g., SMC, MAX, ONE, RANDOM).
table	Specifies the settings for the input table.
varianceDivisor	Specifies the divisor used for calculating variances and covariances (DF, N, WDF, WEIGHT).
weight	Specifies a numeric variable to use as a weight to perform a weighted analysis.

Data Preparation View data prep sheet

Create Data for Factor Analysis

Creates a sample dataset 'analysisData' in the 'mycas' library with simulated numeric variables.

Copied!

1	cas;
2	LIBNAME mycas cas;
3
4	DATA mycas.analysisData;
5	array x{10};
6	DO i=1 to 1000;
7	DO j=1 to 10;
8	x{j} = rannor(123) + (i/1000);
9	END;
10	OUTPUT;
11	END;
12
13	RUN;
14

Examples

Determines the number of factors using the default settings based on the input variables x1 through x10.

SAS® / CAS Code Code awaiting community validation

Copied!

1
2	PROC CAS;
3	factorAnalysis.faNFactors TABLE={name="analysisData"} inputs={"x1", "x2", "x3", "x4", "x5", "x6", "x7", "x8", "x9", "x10"};
4
5	RUN;
6

Result :
A table displaying the suggested number of factors based on default criteria.

Uses Eigenvalue and Proportion criteria, setting the final factor count to the minimum suggested by active criteria, and specifies Squared Multiple Correlations (SMC) for priors.

SAS® / CAS Code Code awaiting community validation

Copied!

1
2	PROC CAS;
3	factorAnalysis.faNFactors TABLE={name="analysisData"} inputs={"x1", "x2", "x3", "x4", "x5"} criteria={{type="EIGENVALUE", threshold=1.0}, {type="PROPORTION", threshold=0.75}} nFactors="MIN" priors={type="SMC"};
4
5	RUN;
6

Result :
Detailed tables showing the number of factors suggested by each criterion and the final determined number of factors.

FAQ

What is the purpose of the faNFactors action?

What criteria can be used to determine the number of factors?

How does the action decide the final number of factors if multiple criteria are active?

How can I customize the parallel analysis simulation?

Can I provide my own prior communality estimates?

What does the "threshold" parameter do in the criteria configuration?

How can I perform a weighted analysis?

Actions associées

factorAnalysis

faExtract

Extracts common factors. The faExtract action is part of the Factor Analysis ...

Table of Contents

Description

Create Data for Factor Analysis

Examples

Basic Factor Determination

Factor Determination with Multiple Criteria

FAQ

Actions associées

faExtract