Modelling the Pull-out Capacity of Ground Anchors Using Multi-objective Feature Selection

Abstract

Pull-out capacity of ground anchors is analogous to axial capacity of piles as both of them apply same methods. These available methods are mostly empirical; therefore, in this paper efficient prediction models for determining the uplift capacity of small ground anchors have been presented using recently developed artificial intelligence (AI) techniques. Multi-objective feature selection (MOFS) has been utilised to find the subset of influential parameters responsible for the pull-out capacity of ground anchors along with the development of prediction equations. MOFS has been applied with artificial neural network and non-dominated sorting genetic algorithm. Prediction models are also presented using two other AI techniques: functional network and multi-variate adaptive regression spline. AI models were compared in terms of different statistical parameters such as mean absolute error, root-mean-square error, correlation coefficient and ranking criterion approach have been implemented to assess the performance of different prediction models.