Actuarial Outpost
 
Go Back   Actuarial Outpost > Search Forums
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions


Showing results 1 to 15 of 15
Search took 0.02 seconds.
Search: Posts Made By: jericc1
Forum: Exam PA: Predictive Analytics 03-02-2020, 01:25 AM
Replies: 850
Views: 204,881
Posted By jericc1
Passed with a 9! So glad to finally be an ASA...

Passed with a 9! So glad to finally be an ASA (hopefully by the end of the month)!!
Forum: Exam PA: Predictive Analytics 12-10-2019, 03:01 AM
Replies: 16
Views: 1,806
Posted By jericc1
For decision trees, cp (the complexity parameter)...

For decision trees, cp (the complexity parameter) is a particularly important hyperparameter. Especially since rpart() has a built in cross validation procedure that facilitates cp tuning and pruning...
Forum: Exam PA: Predictive Analytics 12-09-2019, 06:40 PM
Replies: 850
Views: 204,881
Posted By jericc1
I believe I also read somewhere in the modules...

I believe I also read somewhere in the modules that loading in some packages will reset whatever seed you may have already initialized, so it's safer to just set the seed after loading a library each...
Forum: Exam PA: Predictive Analytics 12-08-2019, 05:20 PM
Replies: 278
Views: 39,653
Posted By jericc1
ACTEX PA Manual: First Component

It would be the cross-validation error + the standard error. The one standard error rule is an alternative for selecting a hyperparamter associated with a more parsimonious/simpler model than the one...
Forum: Exam PA: Predictive Analytics 12-08-2019, 05:16 AM
Replies: 4
Views: 977
Posted By jericc1
Personally, I think it's good practice to...

Personally, I think it's good practice to explicitly binarize your factor variables in the data you're planning to pass through modeling functions, such as lm()/glm(), even if they automatically...
Forum: Exam PA: Predictive Analytics 12-08-2019, 04:55 AM
Replies: 5
Views: 966
Posted By jericc1
Few pros/cons regarding boosting-- Pros: ...

Few pros/cons regarding boosting--

Pros:

Tends to achieve higher predictive accuracy (compared to bagging)
No need to balance the data if response skews mostly one way over another, since...
Forum: Exam PA: Predictive Analytics 12-04-2019, 05:30 PM
Replies: 91
Views: 13,820
Posted By jericc1
That's correct -- when one of your predictors is...

That's correct -- when one of your predictors is log-transformed and you're also using a logit link, then the exponentiated version would take the form you described where estimated odds =...
Forum: Exam PA: Predictive Analytics 12-04-2019, 03:24 PM
Replies: 850
Views: 204,881
Posted By jericc1
The logit link is arguably as easy to interpret...

The logit link is arguably as easy to interpret as the log link. Since the logit function represents log-odds, the way you'd interpret coefficients is very similar to interpreting coefficients with a...
Forum: Exam PA: Predictive Analytics 12-03-2019, 03:46 PM
Replies: 278
Views: 39,653
Posted By jericc1
I haven't actually done any work with these...

I haven't actually done any work with these datasets yet, but a few that look like good candidates to me are:

Bank Marketing (the target being whether the client subscribed to a term deposit)
...
Forum: Exam PA: Predictive Analytics 12-02-2019, 05:00 PM
Replies: 278
Views: 39,653
Posted By jericc1
UCI's ICS department has a machine learning data...

UCI's ICS department has a machine learning data repository: https://archive.ics.uci.edu/ml/datasets.php
I think one of the modules also used a dataset from here.

Other than that, Base R has...
Forum: Exam PA: Predictive Analytics 12-01-2019, 02:43 AM
Replies: 91
Views: 13,820
Posted By jericc1
Regarding the interpretation of the logit link:...

Regarding the interpretation of the logit link: typically, coefficients in logistic regression are interpreted in terms of odds. So in your example, a coefficient of 0.115 can be interpreted as:...
Forum: Exam PA: Predictive Analytics 11-21-2019, 06:06 PM
Replies: 850
Views: 204,881
Posted By jericc1
You can use the stepAIC() function for both...

You can use the stepAIC() function for both forward and backward selection based on AIC. stepAIC() is part of the MASS library which is allowed on the exam.
Forum: Exam PA: Predictive Analytics 11-21-2019, 12:44 AM
Replies: 10
Views: 1,563
Posted By jericc1
This is a common misconception of what the link...

This is a common misconception of what the link function is. What you're describing is applying a transformation on the response variable.

A link function transforms the mean or E[] of your...
Forum: Modules 6-8 10-18-2019, 04:15 PM
Replies: 2
Views: 845
Posted By jericc1
I think the Model Results tab is meant to be...

I think the Model Results tab is meant to be hidden/inaccessible, presumably since it would only reflect projections for 1 simulated outcome.
Forum: Modules 1-5 09-15-2019, 03:16 PM
Replies: 5,666
Views: 2,252,427
Posted By jericc1
Submitted: 8/31/19 MMR: 9/15/19 15 days

Submitted: 8/31/19
MMR: 9/15/19

15 days
Showing results 1 to 15 of 15

 

All times are GMT -4. The time now is 10:31 PM.


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 0.33617 seconds with 7 queries