Let's imagine we have a system/equipment/process which is required to be managed based on data. The system has sensors which measures system parameters. Suppose, we can adjust each measured system parameter except t5, which could be observed and could be changed by other parameters indirectly.
Goal:
How to control t5 using other parameters (t1-t4, t6-t13) and to make the model interpretable. Which values should be set for such parameters to achive specific value of t5?
About data:
Placed in ./data/raw/raw_data.csv
- columns - 13 parameters
- rows - measurements
The main investigation file is ./notebooks/0.0._parshin_explore.ipynb. Detailed descriptions are omitted for some steps. The notebook is mostly aimed to show approaches and logic. Some part of code could be (required t be) refactored.
Parshin Sergei / @ParshinSA / Sergei.A.P@yandex.com