How to predict change by looking at the impact of various attributes?

N_28
N_28 New Altair Community Member
edited November 5 in Community Q&A
I have the following attributes:
  • Country
  • Annual change rate 2014 - 2019
  • Change rate for 2020
  • Breakdown per sector and company size
  • Indicators (e.g. the percentage of companies with a website)

Purpose: I want to predict the differences between split by the following attributes:

  • Per sector/size
  • Per country
  • Per indicator

Since I have over 10 sector/sizes, 30+ countries and 20+ indicators, splitting this manually is a hard job since I need to do that for every combination. I was wondering if there exists a solution in RapidMiner that would allow to make a prediction based on the split attributes I have listed above? The ultimate purpose is if I can predict what will happen when a pandemic like COVID-19 will occur again.

To given a better idea, this is how my Rapidminer design view looks like right now trying to split my data in 70% training and 30% testing and run a decision tree model. However, as you can see below this does not go so well.




Tagged:

Categories