Research on Cost Estimation of Power Lines Construction Projects Based on Extreme Learning Machine Method

In order to estimate the power engineering project budget in our country more objectively and scientifically, a mathematical model based on the Extreme Learning Machine (ELM) to estimate the cost of power lines construction projects is proposed in this paper by combining with single hidden layer feedforward neural networks (SLFN) method. On the basis of the analysis on the factors that affect the power line construction cost and relevant cost events of the projects, the relation sets between these factors and construction costs are built, and the ANN model for cost estimation of power lines construction projects based on extreme learning machine method is set up. In addition, the ELM is used to train the artificial neural networks to determinate its weight parameters. The simulation results show that not only the simulation speed of ELM is fast, but also the model has generalization performance relative to BP networks in the field of power line construction cost estimation.