Skip to main content
. 2018 Mar 1;12(2):265–272. doi: 10.1177/1932296818759558

Figure 1.

Figure 1.

(a) Structure of the proposed NNC neural network. (b) Scheme of the software framework implemented to tune NNC hyper-parameters h: block A randomly initializes h values and splits the dataset to defines test and training set; block B assesses the performance of h in a 5-fold CV setting over the training set; block C implements TPE to optimize h; block D selects the best h set and finally; block E evaluates the performance of NNC on the test set.