An off-topic spot to chat about your musings of the day

PG's law of environmental modelling

Reply
Respected Advisor
Posts: 4,606

PG's law of environmental modelling

"Time required to build a statistical model is inversely related to the number of observations"

 

Although not intuitive, experience with datasets of all sizes (5e1 to 5e6 obs.) provided me with plenty of empirical evidence for this inverse relationship.

PG
Regular Contributor
Posts: 153

Re: PG's law of environmental modelling

I would be interested in reading more on this. Do you have references, or is it more based on experience? Is the amount of time primarily relating to finding a good type of model?

Respected Advisor
Posts: 4,606

Re: PG's law of environmental modelling

The main problem with small datasets in environmental modelling is the abundance of correlated variables covering only a small portion of parameter space. When adding or removing a couple of data points changes your choice of explanatory variables, you know you are in trouble. A lot of time is spent pruning and cross-validating. Most often, the resulting model is deceptively small and simple with modest predictive power.

PG
This widget could not be displayed.
This widget could not be displayed.
Ask a Question
Discussion stats
  • 2 replies
  • 259 views
  • 4 likes
  • 2 in conversation