

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

ShortTerm Actuarial Math Old Exam C Forum 

Thread Tools  Search this Thread  Display Modes 
#11




Wow, the very first part of MLE seemed easy, 4 easy to follow steps or MOM, but the whole variance part is really confusing me. I'm struggling to know how to get the variance of one MLE when the distribution has multiple parameters or if I need to use delta method, or sometimes both in the same problem.
Anyone have a clear, concise breakdown? Otherwise, I'm going to try to come up with something to outline the strategies for the type of questions as previous two posts have said that this is pretty important to the exam. 
#13




Make sure you see examples of chisquared with years instead of other values.
(not Poisson basis, variance in the denominator) Get accustomed to enter values directly to calculator and use the data columns to your advantage. It's important you understand when to reject and when not to (accept to the right). Chisquared also plays a role with loglikelihood later. Learn when you can use one and the other (continuous/discrete, grouped or not).
__________________
German ______________ Prelims: VEE: 
#14




Just finished up section 2/4. Moving onto Credibility Monday. I’ve been doing 10 question quizzes after each subsection and have been doing pretty well on them.
__________________
Introductory I: Introductory II: Actuarial: LTAM  Advanced: FAP  PA Professionalism: APC 
#15




Ugh the CA video on chisquared with years made no sense to me and there was only one example.
Quote:
__________________
Introductory I: Introductory II: Actuarial: LTAM  Advanced: FAP  PA Professionalism: APC 
#16




CA uses a really fast trick to evaluate Gamma CDF. If X is Gamma with parameters alpha and theta, then F(x) = 1  P(N<alpha) where N is Poisson with lambda = x/theta. Since alpha is usually small, you only have to add a couple of Poisson probabilities to get the answer. For example, let’s say X is Gamma with alpha=2 and theta=100. Then F(50) = 1  P(N<2) where N is Poisson with lambda=50/100=0.5. So the final answer would be 1  [P(N=0) + P(N=1)] and you can get the Poisson probabilities at 0 and 1 from the exam tables. Once you get the hang of it it speeds things up TREMENDOUSLY and lets you avoid integration by parts with the Gamma PDF.
Quote:
__________________
Introductory I: Introductory II: Actuarial: LTAM  Advanced: FAP  PA Professionalism: APC 
#18




Facetoface seminar a definite GO
My January examprep seminar for STAM now has sufficient enrollment to be guaranteed to be offered. For information and registration, see my website http://www.actuarialseminars.com .
__________________
Jim Daniel Jim Daniel's Actuarial Seminars www.actuarialseminars.com jimdaniel@actuarialseminars.com 
#19




Practicing Bayesian Credibility, loosing momentum and motivation, but I definitely need lots of practice even though it has a nice general sequence for solving problems. Hoping to get to 3.3 Buhlman credibility tomorrow and maybe take a few days off for Thanksgiving.

#20




Quote:
__________________
Introductory I: Introductory II: Actuarial: LTAM  Advanced: FAP  PA Professionalism: APC 
Thread Tools  Search this Thread 
Display Modes  

