Estimating Linear Models with brms


#1

I read the non-linear models with brms and it seems that you can write out formulas for calculating the response variable. I am trying to do this with calculating energy consumption. A simple equation would be:

use = baseload + (((floorarea*(To-Ti))/Rvalue)/SEER/4000)

use: Energy Consumption
To: Outdoor Temperature
Ti: Indoor Temperature

I only have outdoor temperature and energy consumption in the dataset.

I was hoping to incorporate a distribution of reasonable values for the other variables with the goal of estimating the impact of outdoor temperature on energy use.

However, I am at a loss on how to write the model to achieve this or if it is possible…

  • Operating System: Windows 10
  • brms Version: 2.5

#2

You can of course try to model the other variables as parameters and let the model estimate them as well. I would, however, doubt that such a procedure would make sense in your case, as the model is likely not well identified if you don’t have both outdoor and indoor temperature. So this will likely not work reasonably well.


#3

My plan was to assume a indoor temperature for the home. According to literature, homes indoor temperature set point tends to be within a 2 degree Celsius. So, my guess shouldn’t be to far off from reality.