Like Principal Components Analysis? New Paper Reports It Can Produce “Phantom Oscillation” Artifacts

Author(s): LucianoSphere (Luciano Abriata, PhD)

Originally published on Towards AI.

Principal Component Analysis (PCA), a widely used statistical method for simplifying complex datasets, has been found to produce “phantom oscillations” — patterns that appear in the data although they don’t exist in the original data set. Read on to know more about this, of special relevance to you if you are used to applying PCA on datasets with the features discussed. This also constitutes a chance to overview other limitations and disadvantages of PCA.

Figure composed by the author from his own PCA tool (here).

Principal Component Analysis (PCA) is a dimensionality reduction technique that projects the input variables that describe a set of objects into linear combinations of these variables to attempt maximize explanation of variance in as few variables as possible. PCA is very widely used to simplify complex data sets.

To know how exactly PCA works, check this out:

A tutorial stripping down low-level code that you can edit and run in your browser to understand PCA once and forever…

towardsdatascience.com

But no wonder the technique comes with its flaws. You probably already know or at least are unconsciously aware of the low interpretability of principal components (they are linear combinations of the features from the original data, but these combinations are not… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI