search

vegarsti / fhtboost

R package in development for use in my master thesis. Boosting of FHT in survival analysis.
1 stars 0 forks source link

How to cite this package? #1

Open liupei101 opened 5 years ago

liupei101 commented 5 years ago

Hi! vegarsti!

I am writing my paper about survival analysis. And I found your research work about fhtboost (the R package and master thesis, maybe...) in this repo.

I wonder how to cite your work about fhtboost in my paper? Something like this?

Lee, M.-L. T., and Whitmore, G. A. 2006. Threshold regression for survival analysis: Modeling event times by astochastic process reaching a boundary. Statistical Science 21(4):501–513.

Look forward to your reply! Thx!

vegarsti commented 5 years ago

Hi! Oh my, how did you even find this package? Have you found any use for it? I'm sure it's really hard to figure out how to use it since I haven't documented things, and not sure the structure of the functions make sense yet.

To answer your question, I honestly don't know. I'm sure a proper citation would be if the package was on CRAN. I'm not sure what the convention for citing a package on Github is, but possibly something like this?

Stikbakke, Vegard, fhtboost, (2019), GitHub repository, https://github.com/vegarsti/fhtboost

(inspired by this answer on StackOverflow)

liupei101 commented 5 years ago

Yes, as you see, I really found it. Such a nice job! ^_^

Cause I am investigating some existing or latest work about fht model. If I don’t understand it wrong, your proposed fhtboost improves the traditional fht model(lee and whitmore 2006) by using a boosting method to estimate the parameter of the model. And your published R package is also an important contribution. I prepare to add your research work about fhtboost into the introduction part in my draft article (titled "Survival Analysis using A Multi-output Gradient Boosting Decision Tree Method").

Thank u so much for your citation example! Maybe you can consider publishing your completed master thesis on arxiv. It will make the work known to more researchers, and also be more convenient to cite your thesis.

vegarsti commented 5 years ago

Indeed, currently there doesn't exist any method to estimate high-dimensional models, even regularized low-dimensional models, for the First Hitting Time models. So I'm exploring whether multidimensional gradient boosting will work.

I will indeed publish that master's thesis on ArXiV, thanks for the encouragement! I will hand it in by May 15, will let you know!