Model Averaging for Predicting the Exposure to Aflatoxin B1 Using DNA Methylation in White Blood Cells of Infants

3rd International Seminar On Sciences "Sciences On Precision And Sustainable Agriculture"

Rahardiantoro, S.^1 ; Sartono, B.^1 ; Kurnia, A.^1

Department of Statistics, Bogor Agricultural University, Indonesia^1

关键词: Aflatoxin b1 (AFB1);    Classical approach;    DNA Methylation;    Ensemble modeling;    High dimensions;    Human disease;    Model averaging;    White blood cells;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/58/1/012019/pdf
DOI  :  10.1088/1755-1315/58/1/012019

来源: IOP

【 摘 要 】

In recent years, DNA methylation has been the special issue to reveal the pattern of a lot of human diseases. Huge amount of data would be the inescapable phenomenon in this case. In addition, some researchers interesting to take some predictions based on these huge data, especially using regression analysis. The classical approach would be failed to take the task. Model averaging by Ando and Li [1] could be an alternative approach to face this problem. This research applied the model averaging to get the best prediction in high dimension of data. In the practice, the case study by Vargas et al [3], data of exposure to aflatoxin B1(AFB1) and DNA methylation in white blood cells of infants in The Gambia, take the implementation of model averaging. The best ensemble model selected based on the minimum of MAPE, MAE, and MSE of predictions. The result is ensemble model by model averaging with number of predictors in model candidate is 15.