

FlashChat  Actuarial Discussion  Preliminary Exams  CAS/SOA Exams  Cyberchat  Around the World  Suggestions 

Thread Tools  Search this Thread  Display Modes 
#1




Some Tips with the Calculations?
Some of the calculations in exam C, in my opinion, are getting messy. So, to speed things up, I'm beginning to use shortcuts and memorize some formulas. However, the exam just seems to find a way to trip you up. I'm wondering if anyone had a similar experience and can comment on how to strike some balance.
For example, in the SOA sample exam, there was a question (#256) finding the value of the loglikelihood function with the MLE. Instead of writing out the entire function, I omitted the terms that did not involve the parameter in question. I got the MLE quite quickly, but I forgot the omitted terms when I was plugging the MLE back into the loglikelihood function. Needless to say, this was frustrating beyond words .... I know that the best way is to write out every thing, but like I said, I am trying to speed things up too. 
#2




I run into this as well. I think the key is to do lots of problems of that type and essentially make it muscle memory. At least, that's what has worked for me on certain types of problems that are computationally intensive (Agg Deductibles, Variances, MLEs, Nonparametric Credibility, etc...)
Specifically for MLEs, I've noticed that I can memorize the general "form" of the loglikelihood function for a few distributions (Pareto, Exponential, Gamma, Poisson, Lognormal). This allows me to skip directly to writing l(parameter) instead of taking logs. Note that you do have to be careful and I often times go back and check that I did it correctly to make sure I can still get there from first principles. 
#4




Quote:
Edit: I'll also add that C is more segmented than the exams I've taken so far. So sometimes memorizing all the theories isn't useful or necessary. I had to just sit down and memorize the formulas for conjugate priors. And while practicing you can find and utilize shortcuts. Like MLE(\theta) for Exponential is equal to the sum of the losses divided by the number of uncensored losses. Or know the Weibull shortcut and you know you know the Exponential. I think the key to all of the prelims is practice.
__________________
Last edited by Totient(43); 05292018 at 04:32 PM.. 
#5




Quote:

Thread Tools  Search this Thread 
Display Modes  

