Time series forecasting with different starting conditions

Question

Suppose we have two components/materials, say 'A' and 'B'. When both of these are combined in different ratios it results in a certain property which can be modelled as a time series (Property C).

Cycle (Time)	Component A	Component B	Property C
1	80	20	1.078
2	80	20	1.035
2	80	20	0.987

To predict the property C, I viewed it as a time series forecasting problem. To solve it, I have used ARIMA model which yields good results. How can I incorporate the ratio of 'A' and 'B' into modelling the time series? Ultimately, I want to predict 'Property C' with different ratios of 'A' and 'B'.

This question is unrelated to programming and debugging. It’s about statistics and ML. It belongs on [stats.SE] instead of here, so I’m voting to close it. — Arya McCarthy, Feb 15 '21 at 04:45

score 0 · Answer 1 · answered Feb 16 '21 at 23:14

Looks like you have panel data

So you maybe incorporate using ARIMAX models. Which ARIMA models with exogenous variables, in your case A and B or A/B.

For examples:

Time series forecasting with different starting conditions

1 Answers1