Actuarial Outpost
 
Go Back   Actuarial Outpost > Actuarial Discussion Forum > Property - Casualty / General Insurance
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions



Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 01-18-2017, 11:11 AM
yerromnitsuj yerromnitsuj is online now
Member
CAS
 
Join Date: Jan 2010
Studying for 7
Posts: 305
Default Practical Use of R/Python for the Ratemaking Actuary

For those of you who are R/Python users, what are some ways in which you use them to enhance your ratemaking analysis? For example, say you are doing a standard state review for a certain product. In what ways might you use R/Python to enhance your analysis?
Reply With Quote
  #2  
Old 01-22-2017, 02:07 AM
Marcie's Avatar
Marcie Marcie is offline
Member
CAS
 
Join Date: Feb 2015
Posts: 7,540
Default

Quote:
Originally Posted by yerromnitsuj View Post
For those of you who are R/Python users, what are some ways in which you use them to enhance your ratemaking analysis? For example, say you are doing a standard state review for a certain product. In what ways might you use R/Python to enhance your analysis?
What do you mean by a "standard state review," nitsuJ?
Reply With Quote
  #3  
Old 01-22-2017, 09:51 AM
Guinness Guinness is offline
Member
SOA
 
Join Date: Jul 2016
Posts: 41
Default

What does "enhance" mean? Enhancing your data analysis or looking at more sophisticated stastical/modeling procedures.
Reply With Quote
  #4  
Old 01-22-2017, 07:46 PM
Ron Weasley's Avatar
Ron Weasley Ron Weasley is offline
Member
CAS AAA
 
Join Date: Oct 2001
Studying for naught.
Favorite beer: Butterbeer
Posts: 8,623
Default

Why the focus on R and Python?
__________________
“Fantasy is escapist, and that is its glory. If a soldier is imprisioned by the enemy, don't we consider it his duty to escape?. . .If we value the freedom of mind and soul, if we're partisans of liberty, then it's our plain duty to escape, and to take as many people with us as we can!”
― J.R.R. Tolkien
Reply With Quote
  #5  
Old 01-22-2017, 07:55 PM
Vorian Atreides's Avatar
Vorian Atreides Vorian Atreides is online now
Wiki/Note Contributor
CAS
 
Join Date: Apr 2005
Location: Hitler's Secret Bunker
Studying for ACAS
College: Hard Knocks
Favorite beer: Sam Adams Cherry Wheat
Posts: 61,832
Default

If I'm doing a "standard rate review" . . . the starting point is to understand the tools currently used for the various parts of the rate review.

Once you have that knowledge, it should be clear how you can enhance the process using other means (like R or Python).

So, bottom line, whatever you decide you might try to do to "enhance" the current process, discuss it first with your supervisor.
__________________
I find your lack of faith disturbing

Why should I worry about dying? It’s not going to happen in my lifetime!


Freedom of speech is not a license to discourtesy

#BLACKMATTERLIVES
Reply With Quote
  #6  
Old 01-23-2017, 12:51 AM
extrovertedactuary extrovertedactuary is offline
Member
CAS SOA
 
Join Date: May 2012
Location: West of Nowhere
College: State
Posts: 1,241
Default

Quote:
Originally Posted by yerromnitsuj View Post
For those of you who are R/Python users, what are some ways in which you use them to enhance your ratemaking analysis? For example, say you are doing a standard state review for a certain product. In what ways might you use R/Python to enhance your analysis?
I think you could use those two programs to construct a GLM approach to trending. Also, you might use one of the two to examine patterns over time in individual claim data, especially large claims. An example of examining large claim data is developing a large loss load to stabilize your adjusted loss ratio used to determine the indication. You perhaps could do a cluster analysis if the individual program does not have sufficient claim volume to determine this excess load on its own. If you write 30 programs countrywide, then a k-means clustering algorithm could be used to figure out the optimal assignments of programs to clusters to be used when determining the large loss load.

Good luck!
Reply With Quote
  #7  
Old 01-23-2017, 08:49 AM
Klaymen's Avatar
Klaymen Klaymen is online now
Member
CAS
 
Join Date: Oct 2001
Studying for the Last Reformation
Posts: 19,006
Default

Those sound like interesting projects, but certainly not the sort of thing one would normally expect in a "standard rate review". More likely some person would do the "interesting" tasks and then others involved in the standard rate review would follow through. Not that it has to be that way, just guessing that people using R/Python at work are probably less likely to be involved in rate reviews.

Having said that, I occasionally use R and work on rate reviews a portion of the time, so my own experience contradicts what I just said. Cause I'm special.
__________________

You make known to me the path of life; you will fill me with joy in your presence, with eternal pleasures at your right hand. (Psalm 16:11)
Reply With Quote
  #8  
Old 01-23-2017, 08:54 AM
Marcie's Avatar
Marcie Marcie is offline
Member
CAS
 
Join Date: Feb 2015
Posts: 7,540
Default

Quote:
Originally Posted by yerromnitsuj View Post
For those of you who are R/Python users, what are some ways in which you use them to enhance your ratemaking analysis? For example, say you are doing a standard state review for a certain product. In what ways might you use R/Python to enhance your analysis?
If your standard rate review involves conditional sums over hundreds of thousands of rows in Excel, this could be 'enhanced' by moving to R/Python.

But beware: your analysis will run much, much faster, which might cut into your AO posting time.
Reply With Quote
  #9  
Old 01-23-2017, 09:36 AM
Ron Weasley's Avatar
Ron Weasley Ron Weasley is offline
Member
CAS AAA
 
Join Date: Oct 2001
Studying for naught.
Favorite beer: Butterbeer
Posts: 8,623
Default

If you have a reason to learn R and/or Python, this is a reasonable approach, because you will be working with business problems you are already familiar with and trying to learn new skills within the context of otherwise familiar territory.

If the primary objective is to improve your pricing methodologies, an approach that may be more fruitful is to determine how you want to improve your currentl methodologies and then finding out what tools are readily available to you for making these improvements rather than starting out with the answer of "using R" or "using Python". You can come up with completely nonsensical rate indications using R and Python. There is nothing magic about them that will make your rating practices better.
__________________
“Fantasy is escapist, and that is its glory. If a soldier is imprisioned by the enemy, don't we consider it his duty to escape?. . .If we value the freedom of mind and soul, if we're partisans of liberty, then it's our plain duty to escape, and to take as many people with us as we can!”
― J.R.R. Tolkien
Reply With Quote
  #10  
Old 10-05-2017, 03:40 PM
yerromnitsuj yerromnitsuj is online now
Member
CAS
 
Join Date: Jan 2010
Studying for 7
Posts: 305
Default

I never followed up on this from a while ago. Thanks for the input. I decided recently to look into how appropriate current groupings for large loss loads are using k means. Also looking at some other clustering methods.

To clarify, I am mainly trying o improve R skills within the context of something I am familiar with.
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 03:06 PM.


Powered by vBulletin®
Copyright ©2000 - 2018, 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 0.24031 seconds with 9 queries