0

I have an algorithm that uses four operators, each operator has two or three possible values. I want to test the influence of each value on the performance of the algorithm. In total, 36 variants can be derived by varying the values of the operators.

I run the 36 variants on 3000 problem instances, and calculated the average deviation from the best known solutions.

I gathered the results of the values of each parameters similar to Taguchi Design experiments (Operators are the Parameters and values are the Levels).

My question is: what is the most suitable statistical test that can be used to see if there is a statistically significant impact of a value among the other values of each operator?

Farah Mind
  • 143
  • 5
  • 1
    suppose, if your equation of graph or whatever result you have is 4x + 2y + 3z + 9z = r , you can easily find out how much percent of r is 3z – Frost Jan 29 '23 at 10:43
  • In my question I mean which test to use: t-test, F-test or another test? Can we do an ANOVA for the parameters at once? – Farah Mind Jan 29 '23 at 13:03
  • i don't know much about the topic in details. However. if you application is ok with it, why not? I found someone did ANOVA https://www.researchgate.net/figure/Analysis-of-variance-ANOVA-of-the-5-parameters-Source-of-variation-Degrees-of-freedom_tbl4_267509045 – Frost Jan 29 '23 at 14:15
  • I saw papers in my topic using also ANOVA (similarly to the paper you mentioned) with all the parameters in the same table, and then the Errors and Total in the last two rows. My question may be basic but did they call ANOVA for all the parameters at the same time? Or they called ANOVA for each parameter separately? In the later case, how they obtained the Errors and Total of the last two rows? – Farah Mind Jan 29 '23 at 15:07

0 Answers0