4 ms·
Reminds me of the failing of the k-nearest neighbor algorithm. It works extremely well in low dimensionality, but given high dimensionality, its astoundingly ba
by ldjkfkdsjnv 3y ago
Reminds me of the failing of the k-nearest neighbor algorithm. It works extremely well in low dimensionality, but given high dimensionality, its astoundingly bad. As you scale up the space and number of dimensions, the distances between individual chunks of data gets so large its hard to draw meaningful conclusions.
- nmca 3y agoI dunno I've done a lot of high-dim knn that were quite pragmatically useful
- ldjkfkdsjnv 3y agohttps://en.wikipedia.org/wiki/Curse_of_dimensionality https://en.wikipedia.org/wiki/Curse_of_dimensionality
- nmca 3y agoMaybe we're talking at cross purposes; what I mean is that PCA+KNN or other tricks mean that for high dim and large amounts of data you can get useful results; with the enlarged view of KNN as "anything where the last step is KNN" your statement that it's "astoundingly bad" is wrong. But for the narrow view of just KNN then sure cod applies.
- natch 3y agoHow high is high?
- ohazi 3y agoI had a professor who would jokingly say "It's lonely in 128 dimensions"
- galaxyLogic 3y agoRight, nobody's at those other dimensions at the same time as you are.