ARTIFICIAL INTELLIGENT MODELING FOR CONSTRUCTION COST INDICES, ESTIMATION AND PREDICTION

Abstract

Abstract There is a time difference between the cost estimation stage of construction projects and the implementation phase. In addition, the project takes a long time to complete it and the cost of materials varying from time to time in the market so cost estimation play important role in success of any construction project at initial stage therefore the research aims to develop one model to predict the total cost of the construction project and develop twenty-five models to estimate the average prices of construction items of the project with high accuracy by using artificial intelligence techniques such as multiple linear regression analysis (MLR), support vector machine (SVM) and artificial neural networks (ANN). The data set used to build the models is 34 construction projects, and these projects were collected from several government departments in Diyala province. The optimal method based on precision, and enabling to predict the budget of projects was MLR with precision (98.97%) while the optimal method based on correlation factor was ANN with percentage (100%). The optimal method based on precision, which is able to predict the price rate of land fill work item construction works under moisture proof layers item, ordinary concrete for walkways item, reinforced concrete lintel item, reinforced concrete slab item, reinforced concrete stair item, reinforced concrete for the sun bumper item, cement finishing works item, color pigment item and works of placing marble item model was SVM. Price rate of pentellite paints item model was MLR and for excavation the foundation works item, filling with sub-base item, the construction works above moisture proof layers item, construction work of section item, reinforced concrete foundation item, reinforced concrete column item, reinforced concrete beams item, plaster finishing works item , plastic finishing works item , Stone packaging item model, ceramic works for IV floor item model, ceramic work for wall item, flattening (two opposite layers of lime) and flatting (Tiling) tem was ANN. Also, the optimal technique based on correlation factor, it can be used to predict price rate for all items was ANN expect the construction works under moisture proof item model and pentellite layer was MLR.