Chapter VI
Machine Learning
81
VI.6. Analysis and Validation of the ANN Algorithm
In this section, the neural network (NN) model developed by Gent, M. R., Boogaard, H. F., et al.
(2007) is replicated. The features used by this algorithm are displayed in the Table VI-8 .
Table VI-8 Features used for the development of the NN model.
After reproducing this model as described in Gent, M. R., Boogaard, H. F., et al. (2007), the
obtained results are illustrated in Figure VI-8. The í µí±
2
score indicates a value of 0.84 for the
training ensemble and 0.82 for the test ensemble. These scores demonstrate a good correlation
between the measured and predicted values, as supported by the scatter plot.
Figure VI-8 Predictions of the replicated ANN method.
Machine Learning
81
VI.6. Analysis and Validation of the ANN Algorithm
In this section, the neural network (NN) model developed by Gent, M. R., Boogaard, H. F., et al.
(2007) is replicated. The features used by this algorithm are displayed in the Table VI-8 .
Table VI-8 Features used for the development of the NN model.
After reproducing this model as described in Gent, M. R., Boogaard, H. F., et al. (2007), the
obtained results are illustrated in Figure VI-8. The í µí±
2
score indicates a value of 0.84 for the
training ensemble and 0.82 for the test ensemble. These scores demonstrate a good correlation
between the measured and predicted values, as supported by the scatter plot.
Figure VI-8 Predictions of the replicated ANN method.