Table 1 Parameters of the best ANN architectures obtained using the GA.

From: A Method Based on Artificial Intelligence To Fully Automatize The Evaluation of Bovine Blastocyst Images

Parameters

ANN Architecture

1

2

3

Neurons

1st layer

80

95

59

Transfer function

between input and 1st layer

tribas

radbas

Tribas

output layer

purelin

purelin

Purelin

Training function

trainscg

traincgf

Trainscg

Success grade 1 (%)

71.4

66.7

65.0

Success grade 2 (%)

71.4

78.3

70.8

Success grade 3 (%)

81.1

80.6

89.3

Total success (%)

76.4

76.4

76.4

mse (mean square error)

0.116

0.082

0.126

Back to article page