Is the project still running?

[pingfreud]

I used abrem to perform Weibull analysis and I noticed it didn't have tools to compare samples directly.

I implemented one functin for this: https://github.com/pingfreud/abrem/blob/patch-1/R/abrem.compare.R

With methods suggested by Gerald G. Brown and Hebert C. Rutemiller as described in http://reliawiki.org/index.php/Comparing_Life_Data_Sets

Could this be merged to the dev-public branch?

[jto888]

The abrem project has been dormant for almost 2 years now. There was some academic challenge to some of Dr. Abernethy's approach, so we determined not to release the packages to CRAN.

As I recall the abrem.compare was just a testing function to demonstrate comparative calculations from other packages, most notably survival.

The comparative analysis that you wish to perform has been developed in the debias package as MLEw3pContour. I thought this had been upgraded to the abremDebias package but apparently not.

The layout of the packages are that abrem is a user interface with nice graphic feedback for the technical code. Technical code was originally prototyped in packages pivotals and debias and those packages were in the process of improvement and completion as abremPivotals and abremDebias. The plan has sort of been to perhaps create a single package like Weibull-R and not give as much attention to Abernethy's use of the pivotal analysis on median rank regression. This point seems to be the most in contention.

Work at OpenReliability is presently focused on FaultTree. This is found at

https://github.com/jto888/FaultTree

https://github.com/jto888/FaultTree.user_tutorial

You are welcome to participate in advances to the Weibull-R concept if you choose. Formative plans are to generate graphics using the D3 javascript library. This is the way FaultTree presents its graphics and will become an htmlwidget before CRAN release, which is expected soon.

On 8/22/2016 2:01 PM, pingfreud wrote:

I used abrem to perform Weibull analysis and I noticed it didn't have tools to compare samples directly.

I implemented one functin for this: https://github.com/pingfreud/abrem/blob/patch-1/R/abrem.compare.R

With methods suggested by Gerald G. Brown and Hebert C. Rutemiller as described in http://reliawiki.org/index.php/Comparing_Life_Data_Sets

Could this be merged to the dev-public branch?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Abernethy/abrem/issues/3, or mute the thread https://github.com/notifications/unsubscribe-auth/AIc65AZCvU7gtmhCe4RAcDq8aVt145waks5qieQXgaJpZM4JqI3B.

[jto888]

Sorry I did not notice the full significance of this back in August. I was just posting your comment to the web site, when I realized this. I will look into what you have here with interest.

On 8/22/2016 2:01 PM, pingfreud wrote:

I used abrem to perform Weibull analysis and I noticed it didn't have tools to compare samples directly.

I implemented one functin for this: https://github.com/pingfreud/abrem/blob/patch-1/R/abrem.compare.R

With methods suggested by Gerald G. Brown and Hebert C. Rutemiller as described in http://reliawiki.org/index.php/Comparing_Life_Data_Sets

Could this be merged to the dev-public branch?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Abernethy/abrem/issues/3, or mute the thread https://github.com/notifications/unsubscribe-auth/AIc65AZCvU7gtmhCe4RAcDq8aVt145waks5qieQXgaJpZM4JqI3B.

[AhuPersonal]

Hi I am trying to do a Weibull analysis similar to Reliasoft's Weibull ++ sofware for Weibull 2P MLE method. The graphs in your package look like very useful but I needed a vignette or documentation to see how to use your package. Do you have any suggestion for that ?

Thanks

[CarlesCG]

For more documentation have a look here: http://www.openreliability.org/weibull-r-weibull-analysis-on-r/

[AhuPersonal]

It seems that the website doesn't work ?

Actually I am having difficulties in installing abrem. I get a dependency error.

Error: package ‘pivotals’ required by ‘abrem’ could not be found. But pivotals also is not available. I am using R 3.4.1

Thanks in advance

[AhuPersonal]

Update: I solved it by directly downloading from https://r-forge.r-project.org.

Is there a way to extract eta and beta values from the fitted object in abrem package?

The plots include the eta and beta on the chart and I'd like to attain them via a function in abrem. Currently, I am using only the following functionality:

abr_obj <- Abrem(time=df$lifetime,event=df$status)   
fit1 <- abrem.fit(abr_obj,dist="weibull2p",method.fit=c("mle"))  
plot.abrem(fit1)

Update: found it here: fit1$fit[[1]]$beta

[AhuPersonal]

Might there be a bug with abrem.fit function? I have many censored data sets. The function works well with most of them. But I get the following error in some cases:

Error in prrval[[2]] : subscript out of bounds

Here is the dataset, there is 1 failure status out of 112.

Do you know how to handle this error? Thanks

[jto888]

With one failure status out of 112 a technique called Weibayes is required. I haven't implemented yet, you are first to comment. Will not be hard to do, just priorities, priorities you know.

Many other problems with mle fitting in abrem were caused by the starting estimate for the likelihood optimization routine. This was corrected in underlying code in WeibullR, which now uses the lslr fit as a start. This fix would not have been as easy in abrem due to the package structure that was separating mle from lslr code.

By the way, WeibullR has been submitted to CRAN now, so you should consider moving your work to this package. See https://github.com/Weibull-R/WeibullR. Same concept except names are wblr, wblr.fit, and wblr.conf. Many enhancements have been implemented over abrem. Also, for your original interest there are 'Quick Fit' functions MLEw2p, . . . that will simply return the fit vector based on a number of default selections.

I always appreciate your involvement. I hope to hear from you more. On 3/15/2018 5:03 PM, AhuPersonal wrote:

Might there be a bug with abrem.fit function? I have many censored data sets. The function works well with most of them. But I get the following error in some cases

Error in prrval[[2]] : subscript out of bounds

Here is the dataset, there is 1 failure status out of 112. capture3 - copy https://user-images.githubusercontent.com/22137155/37496926-91bdb300-2872-11e8-9374-ea196a20820d.PNG

Do you know how to handle this error? Thanks

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Abernethy/abrem/issues/3#issuecomment-373560994, or mute the thread https://github.com/notifications/unsubscribe-auth/AIc65HSrXYcCvjPvlIz-4_i0XjuXne8Lks5tewFagaJpZM4JqI3B.

[AhuPersonal]

Thanks @jto888 , I will definitely check WeibullR soon. The main reason I chose abrem at the beginning was that its plotting functionality was comprehensive and similar to Weibull ++ software. If the abrem plotting functions were migrated to WeibullR too, I 'd prefer using the new package with the enhancements.

[jto888]

Oh yes, graphics are the same in WeibullR, thanks to Jurgen Symynck. You may also contact me personally as my contact is referenced as administrator in the package DESCRIPTION file.

We are still open for comments on how to improve. But I expect versioning to slow down to once per month at most on CRAN. Intermediate development versions may appear on github and/or R-Forge. On 3/16/2018 10:12 AM, AhuPersonal wrote:

Thanks @jto888 https://github.com/jto888 , I will definitely check WeibullR soon. The main reason I chose abrem at the beginning was that its plotting functionality was comprehensive and similar to Weibull ++ software. If the abrem plotting functions were migrated to WeibullR too, I 'd prefer using it with the new enhancements.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Abernethy/abrem/issues/3#issuecomment-373782358, or mute the thread https://github.com/notifications/unsubscribe-auth/AIc65G8vuuvTRLeZu4ACfvpCy8kFHxhAks5te_KDgaJpZM4JqI3B.

[fargolo]

Sorry I did not notice the full significance of this back in August. I was just posting your comment to the web site, when I realized this. I will look into what you have here with interest.

On 8/22/2016 2:01 PM, pingfreud wrote:

I used abrem to perform Weibull analysis and I noticed it didn't have tools to compare samples directly. I implemented one functin for this: https://github.com/pingfreud/abrem/blob/patch-1/R/abrem.compare.R With methods suggested by Gerald G. Brown and Hebert C. Rutemiller as described in http://reliawiki.org/index.php/Comparing_Life_Data_Sets Could this be merged to the dev-public branch? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub

3, or mute the thread

https://github.com/notifications/unsubscribe-auth/AIc65AZCvU7gtmhCe4RAcDq8aVt145waks5qieQXgaJpZM4JqI3B.

Sorry, but I did not catch this back in 2018, as the issue as posted in 2016. @pingfreud was my github nickname before.

[jto888]

Sorry I did not notice the full significance of this back in August. I was just posting your comment to the web site, when I realized this. I will look into what you have here with interest. On 8/22/2016 2:01 PM, pingfreud wrote:

I used abrem to perform Weibull analysis and I noticed it didn't have tools to compare samples directly. I implemented one functin for this: https://github.com/pingfreud/abrem/blob/patch-1/R/abrem.compare.R With methods suggested by Gerald G. Brown and Hebert C. Rutemiller as described in http://reliawiki.org/index.php/Comparing_Life_Data_Sets Could this be merged to the dev-public branch? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub

3, or mute the thread

https://github.com/notifications/unsubscribe-auth/AIc65AZCvU7gtmhCe4RAcDq8aVt145waks5qieQXgaJpZM4JqI3B.

Sorry, but I did not catch this back in 2018, as the issue as posted in 2016. @pingfreud was my github nickname before.

It is hard for me to tell whether you identified the successor project as WeibullR. This is alive and kicking at https://github.com/Weibull-R/WeibullR