Nonlinear PCA vs. standard linear PCA
→ How to get the variance that is explained by each component?
→ What about data preprocessing and normalization?
Model complexity
→ How to find the optimal number k of components?
→ How is the reproducibility of NLPCA results?
