Survival Random Forest

[iuri_leite]

Dear Colleagues,

I am inquiring to know if it is possible to execute a survival random forest analysis using SAS. I have been examining the PROC HPFOREST procedure, but I have not encountered any features related to survival analysis within this context.

I have come across a paper wherein survival was considered using a discrete-time logistic regression model.

Thank you in advance.

Best regards,

Iuri Leite

[ballardw]

Since you "have come across a paper wherein survival was considered using a discrete-time logistic regression model." perhaps you  can share a link to that paper so we have some clue what you may be contemplating.

---

@iuri_leite wrote:

Dear Colleagues,

I am inquiring to know if it is possible to execute a survival random forest analysis using SAS. I have been examining the PROC HPFOREST procedure, but I have not encountered any features related to survival analysis within this context.

I have come across a paper wherein survival was considered using a discrete-time logistic regression model.

Thank you in advance.

Best regards,

Iuri Leite

---

[iuri_leite]

Dear @ballardw ,

here is the link of the application of random survival forest using the discrete-time hazard model.

Claim Risk Scoring Using Survival Analysis Framework and Machine Learning with Random Forest (sas.co...

There is also this paper with an application using R.

https://www.sciencedirect.com/science/article/pii/S2589537021003138

Best regards,

Iuri Leite

[sbxkoenk]

Hello @iuri_leite ,

I think you're confusing two things.

I use discrete-time Logistic Hazard Regression regularly and you do that with PROC LOGISTIC or PROC LOGSELECT (or PROC GENMOD or PROC HPGENSELECT).

Powerful technique, ... versatile and flexible.

That Random-Forest Machine Learning technique is a different approach.

Let us know if you would like more information about this PROC LOGISTIC approach to survival analysis. There is a complete course on this subject at SAS of a few hundred pages.

Greet, Koen

[iuri_leite]

Dear @sbxkoenk,

thanks for sending me an email.

I used the discrete-time  hazards model few times (with logistic and log-log link), but I am not interested in using it. I am interested in using the random forest survival model for variable selection.

In the paper I posted, It seems to me that it was used a survival forest for a data  structured for the implementation of a discrete-time logistic regression.

I will read the papers again.

Thanks a lot.

Regards,

Iuri Leite 

[fierceanalytics]

Hello @iuri_leite ,

I initially thought you were talking about Random Survival Forests scikit  recently released. That was purely technical variation. 

The claim paper you provided shows  typically how survival analysis has been done. As you see, there isn't  any special invention on the HPFOREST procedure part to extend the model from logistic approaches. As in all survival analysis, heavy duty is in data preparation. If you have analysis data ready for logistic regression, it is ready for HPfOREST. 

Jia