Is there any way to spot an abrupt (massive) change in the values of a list? For instance, I have a list of items (areas) 2, 4, 6, 8, 1000, 1100, 1500. As you can see the increase from 8 to 1000 is massive. Is there any node that can help extract the value at which this change takes place?
Thanks a lot!
You can compute the Relative Difference
Thanks Martin, it seems that in this solution one has to visually identify the difference. I would like to make it parametric. For instance, we know that index 4 has the (1st) largest relative difference so the node identifies it and gives me the output value (i.e. 1000). I am trying different combination of nodes to see if it works and if you could share some further insight, it’d be very helpful
Relative Difference + Sort
If you can test it against one fixed value (in the image : 63), then something like this?
Sort works for the largest relative difference but not for the first largest relative difference as between 5000 and 1001 but I want the difference between 1000 and 8 so I can get the first largest relative difference. I know it is not really relative if we do not compare it with another value (because the first largest (abrupt increase) number idea is really vague) and I have to identify some sort of value so that the relativity is established. I am thinking about that.
Hi Antonie, yes. The second number 63, how did you decide that? Was it trial and run?
Trial and run
If you are sure that there is a hard value you can find that will always works, I’d do that…
If you’re not sure about that, I think you’d have to compare values to the mean and the standard deviation.
And it might not detect everything without some manual tweaks
If you use the image I posted above, and if you can have also drop values, it will create negative spikes, so make sure to add an “absolute value” component after the relative difference !