Prediction of Compaction Parameters of Soils using Artificial Neural Network

This research was conducted for prediction of compaction parameters of soils from its index properties using artificial neural network. The study consists of database of 177 obtained from laboratory measurements. Seven Parameters are mainly considered as input variables to get most accurate results. Plastic limit, Liquid Limit, Plasticity Index, Percentage fines, Percentage sands, Percentage gravels, specific gravity are the input variables and maximum dry density and optimum moisture content were the outputs.The training operation is performed mainly by multilayer perceptron-back propagation algorithm. The network topology was selected after fixing number of hidden neurons. Statistical parameters are used to evaluate the performance of ANN model and also to compare the model with other compaction prediction methods