. 2017 Nov 11;7(11):386. doi: 10.3390/nano7110386

Table 1.

Predictive model based on Machine Learning and Perturbation Theory (PTML) statistics for the evaluation of mitochondrial oxygen flow modifications due to CNTs (10 random splits for each method).

Regression Method	Statistics	Training		Test
Regression Method	Statistics	R²_tr	RMSE_tr	R²_ts	RMSE_ts
Linear Multi-regression (LM)	Mean	0.358	0.0959	0.356	0.0954
	Min	0.349	0.0954	0.340	0.0932
	Max	0.363	0.0966	0.384	0.0969
Neural Network (NN)	Mean	0.645	0.0709	0.672	0.0681
	Min	0.626	0.0697	0.620	0.0613
	Max	0.659	0.0727	0.739	0.0738
Random Forest (RF)	Mean	0.855	0.0455	0.856	0.0452
	Min	0.851	0.0451	0.853	0.0431
	Max	0.858	0.0462	0.863	0.0461