Table 1 Parameters describing machine learning approaches
From: Machine learning approach for pooled DNA sample calibration
Multi-layer Perceptron (MLP) | |||
|---|---|---|---|
Parameter | Best | Range minimum | Range maximum |
num layers | 2 | 1 | 3 |
nodes per layer | 2 | 1 | 6 |
learning rate | 0.11 | 0.01 | 1.0 |
momentum | 0.15 | 0.01 | 1.0 |
Non-linearity | Sigmoid in all hidden layers | ||
Support Vector Regression (Lib SVM nu-SVR) | |||
nu | 0.092 | 0.01 | 1.0 |
C | 0.027 | 0.01 | 1.0 |
Kernel | Gaussian | ||