TransWikia.com

What are some ways to transition from additive to multiplicative(or vice-versa) changes in time-series quickly?

Data Science Asked on November 5, 2021

Let’s say I have a variable x and as time progresses, I want to update the value of x in additive/multiplicative manner using some attributes.

At t=t, new_x = x * some_percentage or new_x = x + some_absolute_number.

Is there some way to transition from additive change to multiplicative change or vice-versa and the number by which we can? It’s fine if we focus only on one of the additive or multiplicative by reactively changing the percentage_increase or absolute_change.


A toy example task (hypothetical)

For a salary of 100k, an increase of 50% is significant but for a salary of 100(could be pocket money of some child), a 50% increase might not be enough. When we are discussing a smaller number of salaries, we need to increase the salary by absolute number. For a salary of 100, we can increase it by 70 to make it 170)


Are there any ways or literature to handle this which can adapt quickly?

One Answer

The problem with your request is that both ways are always possible. For any pair of values x and new_x you can choose to represent the increase either as an addition or a multiplication:

  • additive increment: new_x -x
  • multiplicative coefficient: new_x / x

So you need a criterion to decide which one to choose, for instance a threshold: if x is lower than 100, use addition otherwise use multiplication. Additionally if any kind of change can happen at anytime it's completely arbitrary, there's no way to represent this in a systematic way.

Answered by Erwan on November 5, 2021

Add your own answers!

Ask a Question

Get help from others!

© 2024 TransWikia.com. All rights reserved. Sites we Love: PCI Database, UKBizDB, Menu Kuliner, Sharing RPP