Hi - One of the business requirements will require me to solve the linear equations with multiple variables - does Anaplan has any functionality of how that can be achieved? To give more details, our client has the requirement of regression with multiple causals (Say 5). This means we need to solve for 5 independent variables and a constant. The linear equation for the dependent variable would be y = a0 + a1 x1 + a2 x2 + a3 x3 + a4 x4 + a5 x5.
To get the value of y i.e. forecast - we need to solve for the six variables in total: a0 to a6. Mathematically, this will give me six equations (say we have derived the six equations in Anaplan).
The question now is to solve all these six equations and get the values of these variables a0 to a6 so I can generate the forecast. I would like to seek help from the community how this can be done.
We did something similar to this to solve some complicated formulas that forecast the inventory needed by health care facilities to support the COVID-19 challenge.
The way we solved this was we first create a list for for each of the product types: replenishable (e.g., masks), reusable (e.g., ventilators), and staff (e.g., ventilator technicians).
Then we built a system module for the products and one of the properties was the product type.
Each product type had a process for generating the quantities needed based on their respective independent variables, or drivers, as we called them. For example, infection rate, recovery rate, avg time on ventilator, average use of mask, etc.
The "X" values in your example of the linear equation would be calculated in separate calculation modules for each product type then "lookup" in the consolidated calc module.
The big assumption here is that the coefficients of the linear equation, or the "a values in your example, are calculated outside of Anaplan, perhaps in Knime or Alteryx and imported into Anaplan as the model learns.
Anyway, let me know specifically the type of equations you're working on, or a made up example, and I'll map out a data model for you.