Structural Engineering and Mechanics, Vol. 45, No.3 (2013) 319-332 319 Application of the ANFIS model in deflection prediction of concrete deep beam Mohammad Mohammadhassani 1 , Hossein Nezamabadi-Pour 2a , MohdZamin Jumaat 1a , Mohammed Jameel 1b , S.J.S.Hakim 1c and Majid Zargar 1d 1 Department of Civil Engineering, University of Malaya, 50063, Malaysia 2 Department of Electrical Engineering, Shahid Bahonar University of Kerman, Iran (Received May 6, 2012, Revised November 20, 2012, Accepted December 15, 2012) Abstract. With the ongoing development in the computer science areas of artificial intelligence and computational intelligence, researchers are able to apply them successfully in the construction industry. Given the complexities indeep beam behaviour and the difficulties in accurate evaluation of its deflection, the current study has employed the Adaptive Network-based Fuzzy Inference System (ANFIS) as one of the modelling tools to predict deflection for high strength self compacting concrete (HSSCC) deep beams. In this study, about 3668measured data on eight HSSCC deep beams are considered. Effective input data and the corresponding deflection as output data were recorded at all loading stages up to failure load for all tested deep beams. The results of ANFIS modelling and the classical linear regression were compared and concluded that the ANFIS results are highly accurate, precise and satisfactory. Keywords: ANFIS; deflection; deep beams; fuzzy inference system; linear regression 1. Introduction The use of concrete deep beams has become more prevalent recently. Deep beams have useful applications in many structures, such as offshore structures, tall buildings, foundations, nuclear power plant and several others. Traditional design assumptions, specifically those regarding plane sections that remain the same pre and post bending are not applicable for deep beams. The behaviour and design of deep beams are completely different with normal beams. Predicting the behaviour of deep beam and deflection of its elements are important tasks for structural engineers. There is no clear guidance on deflection prediction for deep beams. There are only few existing studies on deep beam deflection prediction. Mohammadhassani et al. (2013) studied the application of Artificial Neural Network (ANN) for prediction of deflection at the mid span of deep beams. They concluded a high confidence level of ANN in prediction of Corresponding author, Ph.D., E-mail: mmh356@yahoo.com a Professor b Senior Lecturer c Ph.D. Student d Master