Građevinar, Vol. 77 No. 01., 2025.
Preliminary communication
https://doi.org/10.14256/JCE.4029.2024
Novel meta-ensemble modelling approach and comparison of machine-learning models for rebar price estimation
Sahin Tolga Guvel
Abdulkadir Budak
Ibrahim Karataş
Abstract
The early determination of costs in construction projects is crucial for the planning of expenses throughout each investment stage. Making realistic cost calculations is an effective way of preventing cost overruns that may occur in later stages. Rebar price prediction by considering economic indicators significantly affects investment costs and decisions. Therefore, in this study, using historical data for rebar construction material and economic indicators, nine machine-learning algorithms were used to determine the estimated rebar price for 1-, 3-, 6-, 9-, and 12-month lags. The voting meta-ensemble machine-learning algorithm exhibited the best performance for all lag periods investigated. The most successful estimate was obtained for a 3-month lag period. The mean absolute percentage error (MAPE) and coefficient of determination (R2) values for the rebar price estimation during this period were 3.79 % and 95.51 %, respectively.
Keywords
rebar price estimation; construction management; production planning; meta-ensemble; machine learning
Hrčak ID:
329023
URI
Publication date:
18.2.2025.
Visits: 306 *