No open-access location is known for this record yet.
No open-access location is known for this record yet.
Purpose Cost estimation is one of the most significant steps in construction planning, which must be undertaken in the preliminary stages of any project; it is required for all projects to establish the project's budget. Confidence in these initial estimates is low, primarily due to the limited availability of suitable data, which leads the construction projects to frequently end up over budget. This paper investigated the efficacy of artificial neural networks (ANNs) methodologies in overcoming cost estimation problems in the early phases of the building design process. Design/methodology/approach Cost and design data from 104 projects constructed over the past five years in Jordan were used to develop, train and test ANN models. At the detailed design stage, 53 design factors were utilized to develop the first ANN model; then the factors were reduced to 41 and were utilized to develop the second predictive model at the schematic design stage. Finally, 27 design factors available at the concept design stage were utilized for the third ANN model. Findings The models achieved average cost estimation accuracy of 98, 98 and 97% in the detailed, schematic and concept design stages, respectively. Research limitations/implications This paper formulated the aims and objectives to be applicable only in Jordan using historical data of building projects. Originality/value The ANN approach introduced as a management tool is expected to provide the stakeholders in the engineering business with an indispensable tool for predicting the cost with limited data at the early stages of construction projects.
Loading…
Stored only when both papers are in the corpus — 30 citations exist upstream.
Papers that cite the same references — the relational fallback while embeddings are still backfilling.
Imported from OpenAlex (CC0) — nothing here is generated.