Finding the right attribute

Hi, I am new to new to KNIME and are trying to find the most relevant input factor. I have a table with columns of different food ingredients with values 0 and 1 and allergic symptoms with a scale from 0-10 and an additional column with the sum of the symptom values. Which method or function can I use to find the most relevant food ingredient triggering the allergy symptoms? Thanks for your help.

Hi @RemoW and welcome to the Knime Community.

It might be easier to show us or share some sample data and also show us what the expected output should be. Of course, if there are any rules to follow to reach the expected output, you can share these as well.

We’ll then be able to help. There are usually various ways to do things in Knime, the best way will always depend on what you are trying to achieve.

Data sample.xlsx (24.5 KB)
**[quote=“RemoW, post:1, topic:37563, full:true”]
Hi, I am new to new to KNIME and are trying to find the most relevant input factor. I have a table with columns of different food ingredients with values 0 and 1 and allergic symptoms with a scale from 0-10 and an additional column with the sum of the symptom values. Which method or function can I use to find the most relevant food ingredient triggering the allergy symptoms? Thanks for your help.
[/quote]

**Hi @bruno29a Data sample.xlsx (24.5 KB)
thanks for your quick response! I uploaded a sample data file. I am trying to find out which of the food ingredients starting from column AJ triggers the value in column E, which represents the allergic symptoms. Many thanks for any ideas.

Hi @RemoW , thank you for sharing the data.

I don’t quite understand the rules. Can you elaborate further on the rules? For example, can you give a use case? Like if column AJ has value X then what is the action, or if column E has X value, what is the action? Like what is the relation between column AJ and E?

What are you trying to achieve?

1 Like

More info is certainly the key. When I should guess, I would say you want to unpivot the food ingredient columns and see whether there is a relation between the unpivoted ingredient column values and the sickness
br

1 Like

The goal is to find out which food ingredients trigger the allergy symptoms. Values from column AJ to GC represent the food ingredients in a meal. If the meal contains the ingredient, the value is 1. If not, the cell is blank. Columns I to V are the allergy-related symptoms on a scale from 1-10. And E to G are sums and averages of the symptom values. Again, goal is to find out which ingredient from AJ go GC triggers the symptoms.

Hi @RemoW , thanks for the high level explanation. How would you translate “goal is to find out which ingredient from AJ go GC triggers the symptoms” in terms of rules?

Can you give an example?

Hi @RemoW,

in my opinion for this case using a principal component analysis (PCA) would be the best way to find out which ingredients or combination of ingedients could trigger allergy symptoms.

BR

1 Like

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.