Research Article

GGA-MLP: A Greedy Genetic Algorithm to Optimize Weights and Biases in Multilayer Perceptron

Table 2

MLP architecture and selected attributes for each dataset.

#DatasetMLP architecture (input-hidden-output)Selected attributes for greedy population initialization

1Parkinson22-111-1Average vocal fundamental frequency, minimum vocal fundamental frequency, maximum vocal fundamental frequency
2ILPD5-26-1SGPT alanine aminotransferase, SGOT aspartate aminotransferase, direct bilirubin
3Diabetes8-17-1Blood pressure, insulin, BMI
4Vertebral Column6-31-1Lumbar lordosis angle, sacral slope, pelvic radius
5Spambase57-115-1Length of the longest uninterrupted sequence of capital letters, average length of uninterrupted sequences of capital letters, total number of capital letters in the e-mail 
6QSAR Biodegradation41-83-1SpMax_L: leading eigenvalue from Laplace matrix, J_Dz(e): Balaban-like index from Barysz matrix weighted by Sanderson electronegativity
7Blood Transfusion4-9-1R (recency - months since last donation), F (frequency - total number of donation), M (monetary - total blood donated in c.c.)
8HTRU28-17-1Mean of the DM-SNR curve, standard deviation of the DM-SNR curve, excess kurtosis of the DM-SNR curve, skewness of the DM-SNR curve
9Drug Consumption: Amyl Nitrite12-25-1Nscore, Escore, Oscore, Ascore, Cscore
10Drug Consumption: Ketamine12-25-1Nscore, Escore, Oscore, Ascore, Cscore