B.S. Applied Mathematics, Peking University, Beijing, China, 2015
The Impact of Batch Effect and Batch Adjustment on Predictive Model Performances
This project focuses on how batch effect and its adjustment influence the accuracy of classification models. I simulate expression datasets with differences in mean and variance across batches, and use ComBat for batch correction. Then I validate multiple commonly-used classification models on the data before and after the adjustment. I investigate their performance change on the data with and without batch correction, and try to search for or develop a model which stays robust to batch effect in its prediction performance.