Diana Liang (STATA)
Sijun Zhang (dplyr_glmnet)
Wenjing Zhou (data.table_customized_cross-validation)
The completed report in .html and .Rmd files are stored in current folder.
We’re using a combination of 4 datasets (Demographics, Blood Pressure, Total Nutrients day 1, Total Nutrients day 2) from the 2015-2016 NHANES, which are stored in DATA folder.
We are working on both the issues and commit code reviews. Here some instances about code reviews README.md, STATA_LASSO_penalty, README.md_formula_editing and Interactions_glmnet.
The final scripts path are linked in the Author part and you can find running guidence in each folder's readme.md.
We choose to investigate whether both genders will react the same under the the effect of consumption habits on blood pressure, a known symptom of different chronic diseases. Will the factors that are most important in determining blood pressure be different between males the females?
Whether both genders will react the same under the the effect of consumption habits on blood pressure
We will show in the following analysis that there are certain foods that affect one gender more than the other, and that these foods change depending on the type of blood pressure being measured.
Since our purpose is to find how gender effect the relationship between the nutrition intake and the blood pressure measurement, we only penalize the interaction terms by setting the penalty.factor of the interaction terms with a same positive value and let other terms' zero. The explict form for LASSO minizing goal is shaped to
where is the penalty factor we used for each term
To run the group_2_final.Rmd file in Draft folder, the following packages should be pre-installed in the IDE.
| Version | Package |
|---|---|
| 3.6.1 | dplyr |
| 3.6.1 | ggplot2 |
| 2.1.3 | tibble |
| 1.3.1 | readr |
| 1.0.0 | tidyr |
| 1.4.0 | stringr |
| 0.4.0 | forcats |
| 4.3-0 | Hmisc |
| 1.6.0 | SASxport |
| 2.0-18 | glmnet |
| 1.12.2 | data.table |
| 0.4.0 | Statamarkdown |
| 1.1.0 | kableExtra |
| x.x.x | doMC |
doMC is only available in UNIX-like System. The installation command has been embedded in the group_2_final.Rmd file, thus the UNIX-like System user can knit the group_2_final.Rmd directly after meeting the requirements other than doMC.
As Statamarkdown doesn't support direct downloading using install.packages(), we can use the devtools package to install it from github.com.
library(devtools) # before this you may need to install devtools
install_github("hemken/Statamarkdown")
If for some reason that gives you problems, you can also install from this website
# For Windows
install.packages("https://www.ssc.wisc.edu/~hemken/Stataworkshops/Stata%20and%20R%20Markdown/Statamarkdown_0.3.9.zip", repos=NULL)
# For linux or Mac
install.packages("https://www.ssc.wisc.edu/~hemken/Stataworkshops/Stata%20and%20R%20Markdown/Statamarkdown_0.3.9.tar.gz", type="source", repos=NULL)