Indian Journal of Science and Technology
DOI: 10.17485/ijst/2015/v8i22/79321
Year: 2015, Volume: 8, Issue: 22, Pages: 1-7
Original Article
M. H. Jannaty1* , A. Eghbalzadeh2 and S. A. Hosseini 1
1 Department of Technical and Engineering, Tehran Science and Research Branch, Islamic Azad University, Tehran, Iran; [email protected], [email protected]
2 Civil Engineering Department, Advanced Research Institute for Water and Wastewater, Razi University of Kermanshah, Iran; [email protected]
During recent years, data mining and machine learning techniques have been developed in various fields for building intelligent information systems. However, few of the presented methods possess online support capabilities or sufficient flexibility to use large and complicated data sets. For this reason, the present study implemented the Particle Swarm Optimization (PSO) technique to predict scour depths by obtaining appropriate parameters for the neural network model and fuzzy inference system. The test was conducted based on samples obtained from 188 pier scour depths presented by the United States Geological Survey (USGS). The empirical results showed that, due to its minimum Root Mean Square Error (RMSE), the presented model was preferable to the ANFIS model. Moreover, the proposed model produced better solutions than FDOT and HEC-18 equations. The momentum method was implemented to accelerate learning by teaching for increasing the accuracy of short-term predictions.
Keywords: Field Data, PSO-ANFIS, Scour Depth, Single Pier
Subscribe now for latest articles and news.
