Application of Neural Network Models to Improve Prediction Accuracy of Wave Run-up on Antifer Covered Breakwater

Abstract

KEYWORDSWave run-up; Breakwater; Arti cial neural network; Multi-Layer Perceptron (MLP); MSE.Abstract. The primary goal of this study is to present a better way in terms of cost and experimenting duration, instead of using experimental ways for investigating the wave run-up (Ru) over rubble-mound breakwater and examining the e ect of placement pattern of antifer units on the amount of wave run-up. To do so, Arti cial Neural Networks (ANNs) are suggested. For the sake of comparison, the proposed modeling is put into contrast by the ones obtained via other approaches in the literature. The Multi-Layer Perceptron (MLP) is selected as the arti cial neural network is exerted in this study. In the designed neural network, the numbers of inputs and outputs are selected as four and one, respectively. On the other hand, the number of neurons in the single hidden layer of the network should be determined by trial and error considering the Mean Square Error (MSE) of the training and validation samples, which has been chosen as seven in this paper. The regression equations and MSE for the results obtained by ANN are presented in this paper and are compared with other models in the literature. Moreover, the regular placement is preferred to other placement patterns due to its less MSE obtained by ANN.

Integration of Machine Learning Techniques and Control Charts for Multivariate Processes

“…The network structure was comprised of three inputs including temperature, pressure, and humidity and four outputs including Fault Type 1, Fault Type 2, Fault Type 3, and faultless. In order to minimize the mean SE for both training and testing [46], the number of hidden layers for the neurons was set to six. The parameters and functions used in the network are shown in Table 5.…”