What happens when you don’t have enough data?
Ans: You might see wheel going backwards when it is actually going forward, i.e: you’ll things which don’t exist.
There is underlying belief in the data-science community that given enough data any problem is solvable, but what happens when you don’t have enough data? Answer is you see patterns which don’t exist.
What pattern do you see in the image below? I see two lines diverging.
The points are actually on a sign curve!
The amount of data you need depends on the complexity of the underlying problem you are dealing with, when you don’t have enough data corresponding to the problem you see patterns which don’t exist and from the data itself it is not possible to tell if the pattern is false or not. Bigger problem, from the data itself you won’t be able to tell if you need more data or not.
A physical example of this phenomenon happens when it looks like wheels of a fast moving car are going in the wrong direction. Our eyes (and cameras) can take input with only limited frequency, when the speed of the rotation of the wheels increases a certain threshold, we start seeing wheels moving backwards even though they are moving forward faster.
