Skip to main content

Table 2 Different hyperparameter settings for 2D-Vanilla-CNN model based on the trained and tested statistical measures. The final selected parameters are highlighted

From: Convolutional neural network models for cancer type prediction based on gene expression

HyperparametersLoss
dense layer sizefilterkernelstridemean train_scorestdev train_scoremean test_scorestdev test_score
12832(7, 7)(1, 1)20.99918.22821.28114.904
12832(7, 7)(2, 2)0.0050.0020.1920.022
12832(10, 10)(1, 1)21.39818.58221.77115.298
12832(10, 10)(2, 2)0.0090.0030.1870.008
12832(20, 20)(1, 1)0.0270.0040.2020.029
12832(20, 20)(2, 2)0.0430.0110.2060.009
12864(7, 7)(1, 1)10.21317.68810.56614.618
12864(7, 7)(2, 2)0.0040.0010.1870.018
12864(10, 10)(1, 1)31.4301.14931.6751.019
12864(10, 10)(2, 2)0.0120.0060.1770.014
12864(20, 20)(1, 1)12.02018.05212.14914.818
12864(20, 20)(2, 2)0.0550.0160.2040.020
51232(7, 7)(1, 1)21.24518.41921.17514.815
51232(7, 7)(2, 2)10.94418.95311.02215.306
51232(10, 10)(1, 1)10.96418.98711.14815.482
51232(10, 10)(2, 2)0.0030.0010.2130.025
51232(20, 20)(1, 1)10.98819.00211.13215.436
51232(20, 20)(2, 2)1.1101.8491.2711.397
51264(7, 7)(1, 1)31.4301.14931.6751.019
51264(7, 7)(2, 2)10.21317.68810.56014.622
51264(10, 10)(1, 1)31.4971.21131.6481.087
51264(10, 10)(2, 2)20.62817.85820.48114.363
51264(20, 20)(1, 1)11.29916.82511.56213.969
51264(20, 20)(2, 2)12.02018.04612.15214.776