The code is involved for the statistical analysis section of the LASSO for this article

The code was run using R, software version 4.1.2

The packages you need to install before running the software:

"broom"1.0.2
"dplyr" 1.0.10
"ggplot2" 3.4.0
"openxlsx" 4.2.5.2
"readr" 2.1.3
"sampling" 2.9
"showtext" 0.9.5
"tidymodels" 1.0.0

First run the file "LASSO SCI.R"

Line12: Change CFPD into the dataset of your training and testing set. The input form is ".csv".

Line13: Change CFPD into the dataset of your training and testing set. The input form is ".csv".

Line75-81: You can change best_penalty or best_se_penalty into the penalty you want. The penalty before the change is the best penalty automatically selected by the code.

Line136: The code "var_imf" give the Variables importance score of all variables.

Line 157: The name of the variable selected by the LASSO needs to be filled in after "lambda", otherwise this part of the code will not run.

Line179-189: "lasso_mod_out" in the code gives the predicted Y value of the training and testing set; "lasso_mod_out2" gives the predicted Y value of the validation set.

Then run the file "1000penalties train.R" if you need to calculate a large number of penalties at once and summarize the number of variables and R^2 of the model