Neural Networks for Estimating the Productivity of Concreting Activities

摘要: To overcome the variability and impact of subjective factors on cost concrete-related activities in developing countries, neural networks can offer a guiding tool. In this study, three were developed to estimate productivity, within market, for formwork assembly, steel fixing, concrete pouring activities. Eighteen experts working six projects carefully selected gather data networks. Ninety-two surveys obtained processed use by Commercial software was used perform network calculations. The develop, train, test results framework indicate adequate convergence relatively strong generalization capabilities. When sensitivity analysis input influencing productivity concreting activities, has demonstrated good potential identifying trends such factors.