search

erdogant / pca

pca: A Python Package for Principal Component Analysis.
https://erdogant.github.io/pca
MIT License
284 stars 42 forks source link

Issues with 2 principal components #25

Closed acmoudleysa closed 1 year ago

acmoudleysa commented 1 year ago

image

How do I fix this?

erdogant commented 1 year ago

The first component seems to explains 96% of all variation in your data (thats a lot). The second 4%. Thus all features that are horizontal-ish have the most weight. Did you normalize the data? Maybe that can help. Alternatively, Have you checked whether you have highly correlated features? You can try to remove the multicollinearity at start and then re-run the analysis. You can also limit the number of loadings but that is only for visualization purposes.