I.J. Intelligent Systems and Applications, 2019, 10, 42-53 Published Online October 2019 in MECS (http://www.mecs-press.org/) DOI: 10.5815/ijisa.2019.10.05 Copyright © 2019 MECS I.J. Intelligent Systems and Applications, 2019, 10, 42-53 Prediction of Performance Point of Semi-Rigid Steel Frames Using Artificial Neural Networks Zahra Bahmani a a Department of Computer Engineering, Behbahan Khatam Alanbia University of Technology, Behbahan, Iran E-mail: zahra.bahmani2009@gmail.com Mohammad R. Ghasemi b b Department of Civil Engineering, University of Sistan and Baluchestan, Zahedan, Iran E-mail: mrghasemi@eng.usb.ac.ir. Seyed S. Mousaviamjad c c Department of Civil Engineering, Yazd University, Yazd, Iran E-mail: sajad.mousavi.amjad@gmail.com Sadjad Gharehbaghi d d Department of Civil Engineering, Behbahan Khatam Alanbia University of Technology, Behbahan, Iran E-mail: sgharehbaghi@bkatu.ac.ir Received: 19 April 2019; Accepted: 07 June 2019; Published: 08 October 2019 Abstract—One of the main steps in the performance based seismic analysis and design of structures is determination of performance point where the nonlinear static analysis approach is used. The aim of this paper is to predict the performance point of semi-rigid steel frames using Artificial Neural Networks. As such, to generate data required for the prediction, several semi- rigid steel frames were modeled and their performance point was determined then. Ten input variables including number of bays, number of stories, bays width, moment of inertia of beams, cross sectional area of columns, cross sectional area of braces, rigidity degree of connections and soft story (existence or nonexistence) were considered in the prediction. In addition, the actual results were obtained at the presence of different earthquake intensity levels and soil types. Back Propagation with eleven different algorithms and Radial Basis Function Artificial Neural Networks were used in the prediction. The prediction process was carried out in two steps. In the first step, all samples were used for the prediction and the performance metrics were computed. In the second step, three of the best networks were selected, and the optimum number of samples was found considering a very slight reduction in the accuracy of the networks used. Finally, it was shown that, despite using rather limited number of samples, the generated Artificial Neural Networks accurately predict the performance point of semi-rigid steel frames. Index Terms—Artificial neural networks, prediction, semi-rigid connection, steel structures, performance point. I. INTRODUCTION Past destructive earthquakes (e.g. 1989 Loma Prieta; 1990 Manjil-Rudbar; 1994 Northridge) have left their signature below the documents of the economical and life losses reported then. In accordance with the seismic design codes, some degree of damage is expected for ordinary buildings subjected to design basis earthquakes. Although preventing the natural earthquake occurrence is beyond the human control, the ways of mitigating such losses have been the enforced interest of researchers and engineering communities. The inappropriate performance of structures under the past destructive earthquakes caused the communities to eliminate the unexpected losses then. One of the modern methodologies gaining significant attention is the performance-based seismic design (PBSD) [1]. Over the last 30 years, the conceptual framework of PBSD was developed by the various guidelines published by the well-known engineering associations such as Structural Engineers Association of California [2], Applied Technology Council [3], and Federal Emergency Manage Agency [4]. Conceptually, it is expected to have different seismic performance levels of structures due to various structural seismic behaviors and hazard levels. Because of this, the abovementioned guidelines have introduced different seismic performance levels (or objectives) corresponding to the different hazard levels and thus, each structure of interest should be evaluated based on its relevant performances and hazard levels. Based on the studies so