A new interactive model for improving the learning performance of back propagation neural network

The Back Propagation Neural Network (BPNN) has been used widely in construction management, but in fact, the BPNN is limited by a non-optimum weight adjustment manner and negatively influenced the convergence results. For this reason, this paper proposes the Individual Inference Adjusting Learning Rate technique (IIALR) to enhance the learning performance of the BPNN. The mechanism of the weight adjustment in the IIALR is an individual learning rate for each weight. Furthermore, this paper also establishes the Batch-Online Weight Updating Frequency mode (BOWUF) for the IIALR model, so as to adjust the connected weight of the BPNN properly and effectively. Finally, three cases are used to verify that the IIALR model can be more effective than other modifications of the BPNN. The IIALR model is conducive for assisting with the decision making process of construction management. (c) 2007 Elsevier B.V. All rights reserved.