De-noising Auto-encoder-based Construction Cost Prediction

Abstract:

High-rise building projects being taken as the example, a study was made of the influencing factors about the construction cost for a reliable identification and reasonable quantification. On the basis of the theory of de-noising auto-encoder under deep learning as well as the neural network, a construction cost prediction model was established for nonlinear engineering projects. A case study was made of the model by a simulation prediction on the Matlab platform, which verified the proposed method for predicting the cost of engineering projects.