Table 2 Difference between FA and NMF

Works with negative data

Does not work with negative data

Same word in different factors makes problem

Can handle polysemy

Not intuitive results interpretation

Due to non-negativity it is easier to infer the results

Reduces data dimensionality by identifying factors

Reduces data by spliting it into smaller subsets

Factor identification is found with high correlation variables

Splitting the data by finding the minimum distance