1. In a multivariable scenario, a few potential influencing factors, such as service quality and commodity quality, can be portrayed by mining the influencing factors behind these problems.
2. Dimension reduction before mathematical modeling
Both factor analysis and principal component analysis can be used to reduce dimensions. The advantage of factor analysis is that as a new explanatory variable, factor modeling has better explanatory power.
Therefore, for some data modeling that needs business interpretation, we can extract key factors through factor analysis before modeling, and then use factor scores as explanatory variables to model through classification models such as regression or decision tree.
3 algorithm implementation steps
First of all, it should be explained that, like principal component analysis, the purpose of both methods is to reduce dimensions, so the premise of both methods is incomplete interaction of features.
Factor analysis is to find the linear combination of nonlinear related "variables" to represent the original variables. These "variables" are called factors. .