Time and price of contracted construction as well as their overrun are among highly pronounced issues in the construction practice. Numerous studies indicate that there is a dependency between these parameters at various construction markets. The paper shows the results of a research conducted on a group of 40 projects related to water supply systems carried out in the Federation of Bosnia and Herzegovina from 2001 to 2012. Collected data, obtained through interviewing chief engineers of construction companies, have been used as input parameters for implementation of single linear regression by applying the BTC (Bromilow time-cost) algorithm. The final result is a model in the form of an exponential equation. Whereas it was obtained based on effectuated values, it can be considered appropriate for assessing and testing the construction time in the early planning stage, as well as at work contracting. Since the parameter values in the model depend on circumstances of the specific area under consideration, the application of the model is recommended primarily in the area where the research was conducted. Results of the price overrun and construction overrun interdependence have not resulted in determining the exponential model. The paper also proposes application recommendations and guidelines for further research.

No abstract

Original scientific paper Each contract for a construction project has the costs as an essential element, so the accuracy of forecasting the construction costs can have an impact on the project realization, and also, on the project participants' business. Data for structures (75) were used for modelling with two predictive models: linear regression model (LR) and support vector machine (SVM) model, using Bromilow's model for cost and time relation and predictive modelling software DTREG. The mean absolute percentage error (MAPE) for the SVM model is 0.3% and for the linear regression model is 4.79%. Comparison of the models' results pointed out that the forecasting with SVM was significantly more accurate. Keywords: construction costs; forecasting; linear regression; support vector machine Predviđanje troškova građenja: usporedba točnosti modela linearne regresije i modela podupirućih vektoraIzvorni znanstveni članak Bitni element svakog ugovora za građevinski projekt su troškovi kao bitni element pa točnost predviđanja troškova može utjecati na izvršenje projekta, a također i na poslove onih koji su uključeni u projekt. Podaci za objekte (75) upotrebljeni su za modeliranje s dva modela za predviđanje: model linearne regresije (LR) i model podupirućih vektora (SVM), uz primjenu Bromilowa modela za odnos vremena i troška i softvera DTREG -za prediktivno modeliranje. Srednji apsolutni postotak greške (MAPE) za SVM model je 0.3% a za model linearne regresije 4.7%. Usporedba rezultata modela pokazala je da je predviđanje sa SVM značajno točnije.

No abstract

The need of respecting the construction time as one of the construction contract elements points out that early prediction of construction time is of crucial importance for the construction project participants’ business. Thus, having a model for early prediction of construction time is useful not only for the participants involved in the construction contracting process, but also for other participants in the construction project realization. Regarding that, this paper aims to present a hybrid method for predicting construction time in the early project phase, which is a combination of process-based and data-driven models. Five hybrid models have been developed, and the most accurate one was the BTC-GRNN model, which uses Bromilow’s time-cost (BTC) model as a process-based model and the general regression neural network (GRNN) as a data-driven model. For evaluating the quality of the models, the 10-fold cross-validation method has been used. The mean absolute percentage error (MAPE) of the BTC-GRNN is 3.34% and the coefficient of determination R2, which reflects the global fit of the model, is 93.17%. These results show a drastic improvement of the accuracy in comparison to the model when only data-driven model (GRNN) has been used, where MAPE was 31.8% and R2 was 75.64%. This model can be useful to the investors, the contractors, the project managers, and other project participants for construction time prediction in the early project phases, especially in the phases of bidding and contracting, when many factors, that can determine the construction project realization, are unknown.