It's interesting how it basically generates a larger sample size to create a regression against. The larger the input, the larger the surface area it can compar

It's interesting how it basically generates a larger sample size to create a regression against. The larger the input, the larger the surface area it can compare against existing training data (implicitly through regression of course).