Is there a name for a type of algorithm or processing which uses only one source of data as input?

Question

Many algorithms in remote sensing take one or more different datasets as inputs and return some output. An example might be computing the amount of sulfur dioxide (SO2) injected into the stratosphere after a volcanic eruption, having two main input datasets: 1) the SO2 amount and layer altitude (e.g., Eumetsat IASI Level-2 SO2 data) and 2) the altitude of the tropopause (e.g. NCEP/NCAR reanalysis data). Even though each input dataset contains multiple fields which are used in the computation, there are only two sources of data.

I am working on a remote sensing tool that uses only one input dataset (containing multiple fields) which accomplishes a task that has previously required more than one dataset with quite a bit of uncertainty coming from the other datasets. I think that's kind of novel.

Consequently, I am wondering if there is an existing name for an algorithm which uses only one dataset that distinguishes it from previous algorithms using multiple input sets?

My instincts (read: "love of coffee") tell me to refer to it as a "single-origin" algorithm, but surely there must be a name for this already.

Any insight is greatly appreciated.

Answer 1

I think you are thinking of an injective function or mapping.

However this work you are doing sounds sort of suspect. Usually adding additional data (with appropriate cleaning) reduces uncertainty. If the additional data is adding to the uncertainty that might indicate that the original data set is noisy or unreliable.