Modeling of mechanical properties and bond relationship using data mining process Kemal Tus ßat Yücel a , Cengiz Özel b,⇑ a Civil Engineering Department, Faculty of Engineering and Architectural, Suleyman Demirel University, 32260 Isparta, Turkey b Department of Construction Education, Faculty of Technical Education, Suleyman Demirel University, 32260 Isparta, Turkey article info Article history: Received 31 January 2009 Received in revised form 15 October 2010 Accepted 31 August 2011 Available online 14 October 2011 Keywords: Reinforced concrete Data mining Bond strengths Flexural strengths Bond slippage Modeling and prediction abstract Reinforced concrete is a widely used construction material. Its properties depend on the bond between the reinforcing bar and concrete as much as the compressive strength or properties of the reinforcing bar because of component of construction expose to both flexural and bond together compressive loads. In this paper, the bond properties of concretes with different mix designs were investigated according to the results of compressive, flexural, bond, and flexural-bond tests. The data mining (DM) process was used to determine relationships among the test results and DM algorithms. Seventeen modeling tech- niques within WEKA were applied to the experimental data for the prediction of bond properties. The results show that the implemented models were good at predicting the bond properties. The best results were obtained from the RepTree algorithm for bond strength, the Multilayer Perceptron algorithm for flexural-bond strength, the MedSq algorithm for bond slippage, and the Pace Regression for flexural- bond deformation. Bond and flexural-bond can be easily predicted using the compressive strength, flex- ural strength and tensile stress of the rebar. Although a relationship is also existent between these and bond slippage and flexural-bond deformation, these relationships are weaker than the others. These results suggested that the DM algorithms can be used as an alternative approach to predict the bond strength using the results of compressive, flexural, bond, and flexural-bond tests as input parame- ters. The predictions of the bond slippage and flexural-bond deformation models poorly agreed with the experimental results. It can be obtained more successful results for these variables, when DM models with different inputs like the rebar-concrete interface stress together the measured parameters are used. Crown Copyright Ó 2011 Published by Elsevier Ltd. All rights reserved. 1. Introduction The compressive and flexural strength properties of the rein- forcing bar (rebar) are taken as the basis of a construction design. These properties have been easy done and controlled by some tests. However, constructions or buildings are not only exposed to compressive, flexural or tensile loads. Particularly, reinforced constructions are exposed mostly to these loads. In addition to these loads, there are a variety of affects such as the bond or flex- ural-bond and the performance of reinforced concrete structures, which depend on adequate bond strength between the concrete and the rebar. Bond strength is one of the most important proper- ties that control the behavior of reinforced concrete structures. However, the determination of its effects requires special equipment. To determine the bond properties, the bond characteristics be- tween the concrete and reinforcement are commonly used through pull-out, push-in, and related testing methods. The pull-out test is the easiest and oldest of these tests [1–7]. The relationships between obtained data cannot always be lin- ear; sometimes these relationships are non-linear or cannot easily be understood. This paper reviews quantitative models for bond properties. We defined quantitative models as models that use sta- tistical approaches or machine learning based data mining ap- proaches for bond properties. Data mining (DM), also known as Knowledge Discovery Data (KDD), is the process of analysing data from different perspectives and summarizing it into useful information. It allows users to ana- lyse data from many different dimensions or angles, categorize it, and summarize the relationships identified. Technically, DM is the process of finding correlations or patterns among dozens of fields in large relational databases [8]. Data mining is the extrac- tion of implicit, previously unknown, and potentially useful infor- mation from data. The idea is to build computer programs that sift through databases automatically, seeking regularities or patterns. Strong patterns, if found, will likely be generalized to make accu- rate predictions on future data [9]. DM is a multi-disciplinary field and encompasses techniques from a number of fields, including information techniques, statistic 0965-9978/$ - see front matter Crown Copyright Ó 2011 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.advengsoft.2011.09.020 ⇑ Corresponding author. Address: Faculty of Technical Education, Construction Education Department, Suleyman Demirel University, West Campus 32260, Çünür/ Isparta, Turkey. E-mail address: cozel@tef.sdu.edu.tr (C. Özel). Advances in Engineering Software 45 (2012) 54–60 Contents lists available at SciVerse ScienceDirect Advances in Engineering Software journal homepage: www.elsevier.com/locate/advengsoft