In lots of knowledge science-related duties, we need to understand how sure we’re in regards to the end result. Realizing how a lot we will belief a end result helps us to make higher selections.
As soon as we’ve got quantified the extent of uncertainty that comes with a end result we will use it for:
- situation planning to guage a best-case and worst-case situation
- threat evaluation to guage the influence on selections
- mannequin analysis to match completely different fashions and mannequin efficiency
- communication with decision-makers about how a lot they need to belief the outcomes
The place does the uncertainty come from?
Let’s take a look at a easy instance. We need to estimate the imply worth of a 300-square-meter home in Germany. Accumulating the info for all 300-square-meter homes just isn’t viable. As a substitute, we’ll calculate the imply worth primarily based on a consultant subset.