Back Propagation Neural Network For Prediction Research Articles

Experimental and Predictive Study on Salt Expansion Behaviors of Frozen Saline-alkali Soil

Saline-alkali soil is typically expansible due to salt expansion. Salt expansion always affected by water migration and salt crystallization. As for the soil in frozen ground, salt expansion behavior is also influenced by the temperature and the salt-water reaction. Thus the expansion behavior of frozen saline-alkali soil is a complex and nonlinear problem, which brings damages to structures and engineering, such as subgrade and infrastructure deformations. In this paper, experiment of salt expansion behaviors is carried on, in which the soil samples are mixed with different content of water, salt. And the influence of the water and salt on the salt expansion is analyzed. To study the salt expansion characteristics, two different kinds of mathematical methods are applied, including the back propagation neural network (BPNN) and support vector machine for regression (SVR). 300 data are used for training, and 35 data are used for testing. The calculated results of BPNN and SVR are compared with the experimental data. The results indicate that SVR is superior to BPNN in prediction. And both methods indicated that water and salt influenced salt expansion strongly, which provides reference in environment engineering and chemical engineering.

Volumetric gas-liquid mass transfer coefficient in an external-loop airlift reactor with inserted membrane

The effects of the inserted membrane in the downcomer of an external-loop airlift reactor, the gas sparger type (single orifice and sinter plate) and added alcohol (ethanol, n-butanol, or n-hexanol) on the volumetric gas-liquid mass transfer coefficient (kLa) were studied. Due to the presence of the membrane in the downcomer, kLa did not change significantly; the differences were smaller than 10%. The highest values of the kLa were obtained using the sinter plate. It was found that the addition of small amounts of alcohol increased the mass transfer. Using our experimental results and the data of other authors, the feed-forward back propagation neural network for prediction of kLa in external-loop airlift reactors with alcohol solutions was proposed.

A Neural Network Model for Driver’s Lane-Changing Trajectory Prediction in Urban Traffic Flow

The neural network may learn and incorporate the uncertainties to predict the driver’s lane-changing behavior more accurately. In this paper, we will discuss in detail the effectiveness of Back-Propagation (BP) neural network for prediction of lane-changing trajectory based on the past vehicle data and compare the results between BP neural network model and Elman Network model in terms of the training time and accuracy. Driving simulator data and NGSIM data were processed by a smooth method and then used to validate the availability of the model. The test results indicate that BP neural network might be an accurate prediction of driver’s lane-changing behavior in urban traffic flow. The objective of this paper is to show the usefulness of BP neural network in prediction of lane-changing process and confirm that the vehicle trajectory is influenced previously by the collected data.

Prediction of density, porosity and hardness in aluminum–copper-based composite materials using artificial neural network

The potential of using feed forward backpropagation neural network in prediction of some physical properties and hardness of aluminium–copper/silicon carbide composites synthesized by compocasting method has been studied in the present work. Two input vectors were used in the construction of proposed network; namely weight percentage of the copper and volume fraction of the reinforced particles. Density, porosity and hardness were the three outputs developed from the proposed network. Effects of addition of copper as alloying element and silicon carbide as reinforcement particles to Al–4 wt.% Mg metal matrix have been investigated by using artificial neural networks. The maximum absolute relative error for predicted values does not exceed 5.99%. Therefore, by using ANN outputs, satisfactory results can be estimated rather than measured and hence reduce testing time and cost.

Feedforward Backpropagation Neural Networks in Prediction of Farmer Risk Preferences

AbstractAn out‐of‐sample prediction of Kansas farmers' responses to five surveyed questions involving risk is used to compare ordered multinomial logistic regression models with feedforward backpropagation neural network models. Although the logistic models often predict more accurately than the neural network models in a mean‐squared error sense, the neural network models are shown to be more accommodating of loss functions associated with a desire to predict certain combinations of categorical responses more accurately than others.