I’m currently working with data from a chemical process that runs in steps (each step creates an “intermediate” which then is the starting raw material for the next step).
Each “intermediate” has an internal specification for quality: assay value, impurity X, etc…
I would like to know if it is possible to predict the final specification value, with the values of the intermediates.
Am I reaching to far or is this something that I can do with a workflow in KNIME?
Hi @slm1989 and welcome to the community!
Depending on the actual data, this could sound doable to me - probably a task for a regression algorithm. It depends on the data you have at hand, however, and what you want to predict exactly. Do you have/How much training data do you have to train a model? What does that data look like? Do you want to predict a class (good/bad) or numeric values (% pure, yield, …). Do you already have a ML model/an algorithm in mind you’d want to use?
The general process would then be to build a trainer-predictor construct, as described e.g. here.
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.