This repository provides the data and implementations of the methods described in Tuning parameter calibration for prediction in personalized medicine.
The file PAV.md contains a function PAVedr to select a tuning parameter within a set of tuning parameters for ridge regression that minimizes the prediction error for an specific individual covariate. The aforementioned paper contains detailed descriptions of these methods.
We provide an example code in SimulationStudy.md for a comparison of averaged individual prediction errors with 10-fold cross-validation. Developed for R 3.6.1.
Kidney Transplant Data Analysis :
The processed data is in personalized_medicine/data/E-GEOD-33070_clinical.RData,
personalized_medicine/data/E-GEOD-33070_gene.RData, and
personalized_medicine/data/index_for_parameters.RData, which are preprocessed by Kristoffer Herland Hellton and Shih-Ting Huang. The original raw kidney transplant data was collected and provided by Expression Levels of Obesity-Related Genes Are Associated with Weight Change in Kidney Transplant Recipients. The programs for our statistical analysis is in RealDataAnalysis.md.
-
Shih-Ting Huang, Ph.D. student in Mathematical Statistics, Ruhr-University Bochum
-
Yannick Düren, Ph.D. student in Mathematical Statistics, Ruhr-University Bochum
-
Kristoffer Herland Hellton, Researcher, Norwegian Computing Center
-
Johannes Lederer, Professor in Mathematical Statistics, Ruhr-University Bochum
RidgeCv.md : K-fold cross-validation for ridge regression.
RidgeEstimator.md : Computing ridge estimator.
All of the codes in this repository are written in R and supports all plarforms which are supported by R itself.
This repository does not depend on any R libraries or external sources.
The HDIM package is licensed under the MIT license. To
view the MIT license please consult LICENSE.txt.
Tuning parameter calibration for prediction in personalized medicine