10.2 Summary

Summed up, the usage of the factor_analyzer package is similar to previously introduced workflows for statistical modeling. Please read through the documentation for a detailed overview.

https://factor-analyzer.readthedocs.io/en/latest/factor_analyzer.html

fa_object = FactorAnalyzer(n_factors=3,
                           rotation='promax',
                           method='minres',
                           use_smc=True,
                           is_corr_matrix=False,
                           bounds=(0.005, 1),
                           impute='median',
                           svd_method='randomized',
                           rotation_kwargs=None)

For the FactorAnalyzer object, we have several options as described in the documentation. The most important ones are:

• n_factors: The number of factors

• rotation: The type of rotation to perform after fitting the factor analysis model

• method: The fitting method to use

• is_corr_matrix can be set to Tue if the data is already a correlation matrix

We can then fit the model and extract its estimates such as eigenvalues, loadings, and communalities:

fa_object.fit(data)

ev, cfev = fa.get_eigenvalues()
l = fa2.loadings_
c = fa2.get_communalities()