Skip to main content

Table 1 Different hyperparameter settings for 1D-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 sizefilterkernelmean train_scorestdev train_scoremean test_scorestdev test_score
64(1, 50)80.0690.0310.1670.023
64(1, 50)160.0370.0130.1400.007
64(1, 50)320.0230.0030.1320.006
64(1, 50)640.0130.0020.1280.006
128(1, 50)80.0320.0080.1470.006
128(1, 50)160.0270.0140.1380.014
128(1, 50)320.0110.0030.1210.009
128(1, 50)640.0040.0010.1260.012
512(1, 50)80.0090.0000.1380.008
512(1, 50)160.0060.0010.1270.003
512(1, 50)320.1240.1790.2650.160
512(1, 50)640.0030.0020.1250.008
64(1, 71)80.0720.0090.1770.009
64(1, 71)160.0440.0090.1490.006
64(1, 71)320.0360.0110.1350.009
64(1, 71)640.0160.0040.1240.012
128(1, 71)80.0460.0070.1540.015
128(1, 71)160.0270.0060.1350.015
128(1, 71)320.0140.0020.1290.016
128(1, 71)640.0080.0010.1190.003
512(1, 71)80.0230.0180.1520.023
512(1, 71)160.0090.0080.1320.017
512(1, 71)320.0040.0020.1230.008
512(1, 71)640.0110.0160.1340.015
64(1, 100)80.0880.0100.1720.015
64(1, 100)160.0660.0140.1620.009
64(1, 100)320.0370.0070.1320.009
64(1, 100)640.0240.0090.1280.013
128(1, 100)80.0580.0010.1640.009
128(1, 100)160.0310.0080.1440.014
128(1, 100)320.0190.0040.1280.008
128(1, 100)640.0160.0100.1370.027
512(1, 100)80.0310.0130.1550.014
512(1, 100)160.0090.0010.1350.009