LDvine: Longitudinal D-vine copula model

The codes in this repository implement time-heterogeneous D-vine model (HET--P) presented in the paper

Hoque, M. E., Acar, E. F., and Torabi, M. (2022). A time-heterogeneous D-vine copula model for unbalanced and unequally spaced longitudinal data. Biometrics.

The repository is structured as follows:

-data: Contains csv/rda/rds/txt files that are used as data into R in order to do the analysis.

We used a subset of Manitoba Follow-up Study (MFUS) data. Because of data confidentiality, the data set file is not included in this R project. 
However, the data descriptions and structures are presented below:   

VARIABLE NAME AND DESCRIPTIONS:

COL 01 -- id = individual id number

COL 02 -- followup = follow-up number of each individual

COL 03 -- px_age = age at the time of experiment

COL 04 -- bmi = body mass index in each follow-up

COL 05 -- IHD = ischemic heart disease status

COL 06 -- DBP = Diastolic Blood Pressure

Finally, the data set has 462 subjects with at most 5 follow-ups. In total, the data has 1713 rows and 6 columns.

• src: Contains R functions (.R) that implement the methods in this paper.

• examples: Contains RMarkdown (.Rmd) files with the complied pdf versions. This folder contains two sub-folders:

  • Simulation: this sub-folder contains .Rmd files with the complied pdf versions for simulation study. The R-code for the simulation example is in the R Markdown file "LD-vine_Sim.Rmd" (use RStudio for opening/modifying/compiling). The compiled version is available as a pdf ("LD-vine_Sim.pdf"). The code is commented and directly evaluated (see "LD-vine_Sim.pdf"). To reproduce the simulated example, simply run the R code line by line (after having installed all R-packages that are loaded at the very beginning - if needed).
  • Data Analysis: this sub-folder contains the data application code in RMarkdown. Only the complied version is available as a pdf file ("LD-vine_Data.pdf"), which includes the partial results of data analysis presented in Section 5 of manuscript. However, since the real data is not provided because of data confidentiality, this RMarkdown file is not reproducible.

• results: Save all printed project outputs here, including plots, HTML, and data exports.

First time only: Setting up the directory

To start you need to open the project file (LDvine.Rproj) in RStudio which will bring RStudio to the root directory.

References

• Hoque, M.E., Acar, E.F. and Torabi, M. (2022), A time-heterogeneous D-vine copula model for unbalanced and unequally spaced longitudinal data. Biometrics. https://doi.org/10.1111/biom.13652

Contact

e-mail: hoqueme@myumanitoba.ca, elif.acar@umanitoba.ca, mahmoud.torabi@umanitoba.ca