Closed cangubur closed 5 years ago
We must have had a miscommunication. What you describe would indeed be wrong. I think I was discussing two different approaches you could take when you code up the log-likelihood function: you could write a "for"-loop through each observation. Or you could vectorize some of the operations. For example, give parameters theta
, you could compute S,V, and F for each row in single, vectorized operations.
When we talked about what would be an easy way to approach Q1, you said find the average profits for N=0 to 6, but isn't that wrong?
Since different markets (counties) have different populations and different population variables, shouldn't we just loop through every row and calculate profits for each county and etc.