Factor Analysis
Statistical method that models observed variables as linear combinations of latent factors plus noise
Factor Analysis
Statistical method that models observed variables as linear combinations of latent factors plus noise.
When to use:
- Believe data has underlying latent factors
- Need probabilistic model
- Want to model noise explicitly
- Social science / psychometric data
- Need factor interpretation
Strengths: Probabilistic model, models noise, interpretable factors, handles missing values Weaknesses: Assumes Gaussian noise, slower than PCA, can be unstable
Model Parameters
N Components (default: 2, required) Number of latent factors.
Max Iterations (default: 1000) Maximum EM algorithm iterations.
- 1000: Usually sufficient
- 2000+: For difficult convergence
Tolerance (default: 0.01) Convergence threshold.
- 0.01: Standard (default)
- 0.001: Stricter convergence
Random State (default: 42) Seed for reproducibility.