An easy way to use artificial neural network model for calculating stability number of rock armors

Since Van der Meer proposed a new empirical formula to calculate the stability number of rock armors based on his own experimental data in 1987, the data have been used in the development of artificial neural network (ANN) models. However, the ANN models are seldom used probably because they did not significantly improve the accuracy compared with the empirical formula and/or the engineers are not familiar with the ANN models. To resolve these problems, first we develop an ANN model combined with a principal component analysis (PCA) that markedly improves the accuracy of the model. The by-products of the model such as weights and biases are then used to develop an explicit method to calculate the stability number. The developed method is not as simple as the empirical formula but it can be easily used by an engineer who has basic knowledge of matrix operation without requiring knowledge of ANN and PCA. It is equally accurate as the developed ANN-PCA hybrid model and more accurate than previous empirical formula and ANN models.