Table 2.

Descriptive data of the performance characteristics for the feedforward neural network FFNN and the multilinear regression model MLRM.

Mean squared prediction error msPE (N = 250)	Feedforward neural network FFMM				Multilinear regression model MLRM
	Training data (N = 150)	Validation data (N = 50)	Test data (N = 50)	Iterations (epochs)	Training data (N = 150)	Validation data (N = 50)	Test data (N = 50)	R²
Prediction of IOL decentration	0.0170 mm²	0.0156 mm²	0.0151 mm²	2	0.0182 mm²	0.0199 mm²	0.0266 mm²	0.265
Prediction of IOL tilt	0.6754°²	0.9409°²	0.8004°²	5	0.8529 °²	1.1319 °²	1.1278 °²	0.391
Prediction of IOL axial position	0.0415 mm²	0.0432 mm²	0.0405 mm²	2	0.0407 mm²	0.0499 mm²	0.0439 mm²	0.671

All models were trained on the training set and tested on the test set. The validation set was used with the FFNN for backpropagation. On the left side, the performance characteristics of the FFNN are listed in terms of mean squared prediction error for the training data, validation data and test data together with the optimal number of iteration cycles (epochs) during training (performance data refer to this iteration). On the right side, the performance characteristics of the MLRM are listed in terms of mean squared prediction error for the training data, validation data and test data together with the coefficient of determination (R²) derived from the training set