abrem.fit weird behaviour

[CarlesCG]

Seems that the function abr.fit is not getting the fit.
From the first example at the vignettes:

library(abrem)
Prob2.1Data   <- c(150,85,250,240,135,200,190)
Prob2.1Object <- Abrem(Prob2.1Data)
Prob2.1Object <- abrem.fit(Prob2.1Object)

Then this error pops up:

Error in .Call("LSLR", na.omit(x)$time, na.omit(x)[, which(where)[1]],  : 
  "LSLR" not resolved from current namespace (abremPivotals)
*** calculateSingleFit: Nothing has been fitted. ***
*** Does "method.fit" include sensible options? ***

Naming some defaults doesn't help either: abrem.fit(Prob2.1Object,dist="weibull2p", method.fit= c("rr","xony")) but MLE gets the fit done: abrem.fit(Prob2.1Object, method.fit="mle")

• Could be a problem with the abremPivotals dependencies?
• Is this the expected behavior?

Just in case my session info:

> sessionInfo()
R version 3.2.0 (2015-04-16)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: OS X 10.10.3 (Yosemite)

locale:
[1] es_ES.UTF-8/es_ES.UTF-8/es_ES.UTF-8/C/es_ES.UTF-8/es_ES.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] abrem_0.1.23         debias_0.1.9         abremPivotals_0.2.10

loaded via a namespace (and not attached):
[1] tools_3.2.0               RcppArmadillo_0.5.200.1.0 Rcpp_0.11.6

[jto888]

It has been a bit of time. Project Abrem has been on a hold. I had a complete package abrem/abremPivotals/abremDebias in a second preliminary fork back in October 2014.

The project was indeed vulnerable to breakage between abrem (the application controller) and abremPivotals and abremDebias (the model).
In particular the vignettes were created by Jurgen Symynck, author of abrem package. I have spent time assuring that the version of the "Using Abrem" that is posted at www.openreliability.org is functional by simple copy and paste into R console.

The project is not dead, only dormant. Impasse occurred due to problems with academic acceptance of some of Dr. Abernethy's presentation. I expect to resume with guidance from Bill Meeker, but have been working - longer than expected - on learning to build web applications for user front end. I am expecting some kind of name change when work resumes.
I am fond of something like Weibull-R. The goal is to provide a free user signup for access to this as a cloud app. Lots of things for me to learn about - AWS - security - scalability . . .

On 6/7/2015 5:59 AM, Carles C.G. wrote:

Seems that the function /abr.fit/ is not getting the fit.

From the first example at the vignettes:

library(abrem) Prob2.1Data <- c(150,85,250,240,135,200,190) Prob2.1Object <- Abrem(Prob2.1Data) Prob2.1Object <- abrem.fit(Prob2.1Object)

Then this error pops up:

Error in .Call("LSLR", na.omit(x)$time, na.omit(x)[, which(where)[1]], : "LSLR" not resolved from current namespace (abremPivotals) * calculateSingleFit: Nothing has been fitted. * * Does "method.fit" include sensible options? *

Naming some defaults doesn't help either: |abrem.fit(Prob2.1Object,dist="weibull2p", method.fit= c("rr","xony"))| but MLE gets the fit done: |abrem.fit(Prob2.1Object, method.fit="mle")|

• Could be a problem with the /abremPivotals/ dependencies?
• Is this the expected behavior?

Just in case my session info:

|> sessionInfo() R version 3.2.0 (2015-04-16) Platform: x86_64-apple-darwin13.4.0 (64-bit) Running under: OS X 10.10.3 (Yosemite)

locale: [1] es_ES.UTF-8/es_ES.UTF-8/es_ES.UTF-8/C/es_ES.UTF-8/es_ES.UTF-8

attached base packages: [1] stats graphics grDevices utils datasets methods base

other attached packages: [1] abrem_0.1.23 debias_0.1.9 abremPivotals_0.2.10

loaded via a namespace (and not attached): [1] tools_3.2.0 RcppArmadillo_0.5.200.1.0 Rcpp_0.11.6 |

— Reply to this email directly or view it on GitHub https://github.com/Abernethy/abrem/issues/1.

[CarlesCG]

• Sorry to hear about the breakage. Sounds like a deeper task that requires to have an understanding about package and specially abremDebias and abremPivotals.
• Are you open to (small for the moment) collaborations? I'll have a look to the vignettes, and see if I can correct the functions to force to use "mle" fit. At least the vignettes will run as an example.
• Looking forward to see this project active, I'm quite interesting.

[CarlesCG]

Well, after playing with the method of installation I manage to create a simple online App for the abrem package:

• You can check it out the app here
• A small elevator pitch here
• The repo with the code and the installation how to here

I hope this can inspire and give some publicity to this great package.