Actuarial Outpost
 
Go Back   Actuarial Outpost > Exams - Please Limit Discussion to Exam-Related Topics > SoA/CAS Preliminary Exams > Exam PA: Predictive Analytics
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions



Reply
 
Thread Tools Search this Thread Display Modes
  #91  
Old 04-11-2019, 01:06 PM
DjPim's Avatar
DjPim DjPim is offline
Member
SOA
 
Join Date: Nov 2015
Location: SoCal
Posts: 743
Default

Quote:
Originally Posted by gaudettj View Post
I'm debating doing this in June. I studied GLMs at University, did pretty well on top of finding it interesting, so I'm not overly concerned about the technical side of it. Anybody think this is too short of a time frame to get through the modules and absorb the material?
It depends on how much you were taught at Uni. Do you know R / was your class using it? Was it just OLS/GLM, or did you also go into decision trees, random forests, and other methods (GBM / PCA)? This exam does kind of go over a wider range of topics and methods than most classes would, but I would say your comfort in R and your writing ability might be a more important point if you are coming in with some background
__________________
Quote:
Originally Posted by Dr T Non-Fan View Post
"Cali" SMH.
Reply With Quote
  #92  
Old 04-12-2019, 12:37 AM
TranceBrah's Avatar
TranceBrah TranceBrah is offline
Member
SOA
 
Join Date: Mar 2014
Location: Best Coast
Posts: 255
Default

Quote:
Originally Posted by gaudettj View Post
I'm debating doing this in June. I studied GLMs at University, did pretty well on top of finding it interesting, so I'm not overly concerned about the technical side of it. Anybody think this is too short of a time frame to get through the modules and absorb the material?
Easily doable. There's still 2 months until the exam, which is more than enough time. If you look through the sample reports and December solutions you will see you don't really need too much technical knowledge for this.
Reply With Quote
  #93  
Old 04-12-2019, 12:20 PM
Squeenasaurus Squeenasaurus is offline
Member
SOA
 
Join Date: Jul 2016
College: Illinois State University
Favorite beer: Lagunitas
Posts: 185
Default

I sat in December and I will say it was pretty awful. But I'm back again ready for round two. Here's my plan of attack:
  • TIA R course
I've spent the past two weeks running through that a second time. They added additional material specific to Exam PA but nothing you would miss out on from the modules. Good for those of you who are beginners to R but not necessary IMO.

  • SOA Modules
I ran through these twice last sitting so I'll probably glaze over the easy modules but spend more time on the tougher ones. I'm telling you, last sitting I could answer almost any question you had about the models they introduce in the modules. I spent hours on Wikipedia and Youtube looking up random stuff about GBMs and random forests because I found that stuff very interesting. I practiced some complicated R code to learn what the different parameters did and to see how accurate I could make a model.

But knowledge isn't really what is being tested on this exam...

  • Practice!!
I can't emphasis this step enough. In December, I didn't sit down and practice writing a full report in one 5 hour sitting. I'd usually break it up over a couple days and then compare what I did to what the SOA did in their sample project. Most of our concepts aligned pretty well so I'd shrug my shoulders and say "ehh I did pretty good. I can probably do that in 5 hours if I needed to." I was mistaken.

The hardest part for me on exam day was time management. I spent way too much time coding (especially cleaning and analyzing the data) and not enough time putting together a well constructed report. Don't be like me. Actually practice a couple full 5 hour long sample projects, even if you just make one up off of a data set you find online somewhere.

This exam is all about communicating. If you've done the FAP modules (or at least some of them), I think you're in a better position than someone who hasn't. I started the FAP after exam PA last December and got through the first four EOMs. I will say they have tremendously helped me understand what kind of memo the SOA is looking for. Anyways, there's my two cents. Best of luck to everyone!
Reply With Quote
  #94  
Old 04-12-2019, 03:32 PM
avocado avocado is offline
Member
SOA
 
Join Date: Apr 2018
Posts: 106
Default

Quote:
Originally Posted by Squeenasaurus View Post
This exam is all about communicating. If you've done the FAP modules (or at least some of them), I think you're in a better position than someone who hasn't. I started the FAP after exam PA last December and got through the first four EOMs. I will say they have tremendously helped me understand what kind of memo the SOA is looking for. Anyways, there's my two cents. Best of luck to everyone!
Can you share some of your insights as to what kind of memo the SOA is looking for? I won't be able to start the FAP before the exam this June.
Thanks!
Reply With Quote
  #95  
Old 04-12-2019, 03:51 PM
Whoaminoneofyourbusiness's Avatar
Whoaminoneofyourbusiness Whoaminoneofyourbusiness is offline
Member
SOA
 
Join Date: Jan 2017
Location: The Grand Tournament
Studying for DP
Posts: 945
Default

Just outline what the sample solution memos are saying and regurgitate the format
__________________
Spoiler:
Reply With Quote
  #96  
Old 04-13-2019, 08:00 AM
NchooseK NchooseK is offline
Member
SOA Non-Actuary
 
Join Date: Nov 2012
Location: Philly area
College: Swarthmore College (BA Mathematics), Villanova University (MS Applied Stat)
Posts: 353
Default

**Update, with questions for the allstars:**
  • Regression Trees (ANOVA) are not covered in any organized format. SOA 7.2 RMD has a few chunks, not much interpretation, no testing/validation, no interpretation. I'm left with Google, which isn't foreign to any of us.
  • What is regression trees' analog to what classification models pull out of confusion matrices? RMSE? RMSE(Cross Validation)? Sure, there is the MSE of the prediction the node the node makes, which is the error if all obs. at the node were predicted to have the mean value for that node. How do I access accuracy for regression trees? THANK YOU.
  • The sample solution to Sample Project 2: Student Success is completely inadequate. The narrative flows, but many many items we are required to address are flat out missing. I almost get no value out of that solution. Thoughts? Agree? Not that important.

Thank you in advance. I love searching instead of learning :/
__________________
Exams: P | FM | C | MFE | LTAM | SRM Credit | PA

VEE: Statistics | Finance | Economics

FAP: 1 | 2 | 3 | 4 | IA | 6 | 7 | FA

Conferences: APC
Reply With Quote
  #97  
Old 04-13-2019, 12:57 PM
ActuariallyDecentAtBest ActuariallyDecentAtBest is offline
Member
SOA
 
Join Date: Dec 2016
Posts: 409
Default

So when can we actually register for the exam??
Reply With Quote
  #98  
Old 04-13-2019, 02:10 PM
ShannonB1 ShannonB1 is offline
SOA
 
Join Date: Mar 2019
Posts: 10
Default

Quote:
Originally Posted by ActuariallyDecentAtBest View Post
So when can we actually register for the exam??
You can register now. https://www.soa.org/Education/Exam-R...analytics.aspx
Reply With Quote
  #99  
Old 04-13-2019, 02:55 PM
TranceBrah's Avatar
TranceBrah TranceBrah is offline
Member
SOA
 
Join Date: Mar 2014
Location: Best Coast
Posts: 255
Default

Quote:
Originally Posted by NchooseK View Post
**Update, with questions for the allstars:**
  • Regression Trees (ANOVA) are not covered in any organized format. SOA 7.2 RMD has a few chunks, not much interpretation, no testing/validation, no interpretation. I'm left with Google, which isn't foreign to any of us.
  • What is regression trees' analog to what classification models pull out of confusion matrices? RMSE? RMSE(Cross Validation)? Sure, there is the MSE of the prediction the node the node makes, which is the error if all obs. at the node were predicted to have the mean value for that node. How do I access accuracy for regression trees? THANK YOU.
  • The sample solution to Sample Project 2: Student Success is completely inadequate. The narrative flows, but many many items we are required to address are flat out missing. I almost get no value out of that solution. Thoughts? Agree? Not that important.

Thank you in advance. I love searching instead of learning :/
  • there is no analog. we would only use confusion matrixes for classifaction problems. the accuracy assessment would be the same as a linear model - simplest is MSE
  • what are some things you feel are inadequate? It seems to answer everything fine to me?
Reply With Quote
  #100  
Old 04-13-2019, 06:34 PM
XFENG01 XFENG01 is offline
SOA
 
Join Date: Dec 2018
Posts: 7
Default

Quote:
Originally Posted by NchooseK View Post
Similarly, I have an MS in Applied Stat. Two grad classes routinely worked in R, but I had/have a decent amount of programming experience. An elective course in sampling methods has been slightly helpful toward the end of the modules.

The stat coverage is not difficult. Instead of writing R code, you largely just need to understand and be able to modify or add to it. I confidently enjoy writing. These tasks are not difficult individually. However, interactively applying my stat, programming and writing brains on one large projects is somewhat daunting; I've never been required to do math, program and write concurrently in an exam setting. Furthermore, Prometric has bad slow computers and the user experience is a mortal lock to be unfriendly. Having been administered only once previously, we are also vulnerable to a range of the unexpected from SOA.

You are clearly an R pro, so you might be able to shut off your programming brain. 9 total modules. IIRC, Modules 1 and 3 are exclusively introductory R. Modules 2 and 9 discuss how SOA wants you to structure your writing.
Modules 4 and 5. Basic stat is covered in Modules 3 and 5--you already know it if you have an MS. I forget Module 8--on principle components and stuff. The heart and soul--like 98% of your study time--are Modules 6 and 7. With all my stat experience, I was introduced to several new topics in these modules. They aren't terrible at all--some are very interesting. There may be some memorization required here, though.

You may be on Module 6 after an hour or two. 6 is very long. 7 is somewhat long.

There are two practice projects and one prior exam, all with sample solutions. The second practice project, Student Data, is not touched upon in the Modules and doesn't follow the Module 9 heavily recommended writing outline in the least. The first sample project, Term Life, is covered in depth throughout the entire module set, but either skips several outline items or presents them in bizarre order. The actual November exam has VERY helpful commentary in the RMD and sample solution. The actual exam followed the outline closest.

Studying would be easier if I used multiple monitors, tbh.

Good luck!
I don't have any stats background and was very confused by module 6. Do you have any suggestions of where to look for help?
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


All times are GMT -4. The time now is 03:42 AM.


Powered by vBulletin®
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
*PLEASE NOTE: Posts are not checked for accuracy, and do not
represent the views of the Actuarial Outpost or its sponsors.
Page generated in 0.58171 seconds with 10 queries