Algorithm 1: Neural network | |
Step | Processes involved |
1 | Initialization of synaptic weights |
2 | Preprocess input block with CA and exam scores |
3 | Pass inputs to the real output block (Oreal) |
4 | Compare if Oreal matches Eexpected; then successful |
5 | If Oreal mismatch Eexpected; then repeat step 3 |
6 | Define boundaries for anomalous data points |
7 | Change coinciding synaptic weights with target and repeat step 2 until training set is empty |
8 | Record similar patterns in training set |
9 | If trained process is complete; then evaluate ANN |
10 | Return |