Fig. 2
From: A quantitative prediction method utilizing whole omics data for biosensing
Performance of prediction models based on OmicSense and six commonly used methods evaluated using a transcriptome dataset. The upper two graphs show the prediction errors (mean absolute error; MAE) of the models. The middle and bottom figures show the increase in prediction errors upon application of OmicSense to validation data (∆MAE overfit ) and upon the addition of noise (∆MAE noise ), respectively. The left and right figures show the results using 30 and 300 genes in the transcriptomes, respectively. Different colors indicate different prediction methods, and different letters indicate a significant difference between group means tested using the Tukey’s honest significant difference test (P < 0.05).
