EUVIMEDEUVIMED entdecken
Systematische Übersicht

Systematic Review of Machine Learning–Driven Cost Estimation in Building Information Modeling-Enabled Construction: Trends, Gaps, and Future Research Directions

Emmanuel C. Njemanze, Samson Micheal Idoghor, Efe Okparabenyo

Journal of Trending Research · 29.04.2026 · 1 (1) · 69-81

DOI: 10.69739/jtr.v1i1.1896 ↗

Kurz gefasst

Systematische Übersicht in der Zeitschrift Journal of Trending Research, online erschienen am 29.04.2026. Verfasst von Emmanuel C. Njemanze und 2 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This is not a recent phenomenon, and construction cost overruns are prevalent all over the world, with big projects being as much as 20-45 percent over the budget. Although both of the two disciplines are applicable in cost estimation in their own ways, there is no overarching synthesis of a combination of the two disciplines, which are Machine Learning (ML) and Building Information Modeling (BIM). This study fills this gap by offering a systematic review of the ML-based cost estimation in the BIM-enabled construction that is guided by the principles of PRISMA 2020. After the screening and quality testing (κ = 0.83 title/abstract; κ = 0.79 full-text), they identified 1,847 records (Scopus, Web of Science, and PubMed/ScienceDirect, 20152025). The narrative synthesis was the result of high heterogeneity (I 2 > 75%). This fact demonstrates that the Artificial Neural Networks (n = 18), random forest, and XGBoost are the best in the literature, and deep learning and hybrid models have been outperforming these models since 2019. The high-performance models had a habitual achievement of the degree of R 2 = 0.95 and MAPE = 5% compared to the traditional approaches (MAPE = 1020%). Hybrid models reported a mean R² ≈ 0.96 and MAPE ranges of 3–7%, outperforming shallow models (R² ≈ 0.86–0.92; MAPE 8–18%). Nevertheless, to this day, it has some severe gaps: 87 percent of the studies are on the pre-construction phases, 18 percent are on explainable AI, and 23 percent are externally validated. The results have demonstrated that the accuracy of the estimation is not only dependent on the complexity of the algorithm, but also on the quality of BIM data, interoperability, and maturity in integration. The lifecycle modeling, explainable AI, and standardized benchmarking datasets should be prioritized in future research.

Zitieren

APA

Njemanze, E. C., Idoghor, S. M., & Okparabenyo, E. (2026). Systematic Review of Machine Learning–Driven Cost Estimation in Building Information Modeling-Enabled Construction: Trends, Gaps, and Future Research Directions. Journal of Trending Research, 1(1), 69-81. https://doi.org/10.69739/jtr.v1i1.1896

Vancouver

Njemanze EC, Idoghor SM, Okparabenyo E. Systematic Review of Machine Learning–Driven Cost Estimation in Building Information Modeling-Enabled Construction: Trends, Gaps, and Future Research Directions. Journal of Trending Research. 2026;1(1):69-81. doi:10.69739/jtr.v1i1.1896

BibTeX
@article{Njemanze2026,
  author = {Njemanze, Emmanuel C. and Idoghor, Samson Micheal and Okparabenyo, Efe},
  title = {Systematic Review of Machine Learning–Driven Cost Estimation in Building Information Modeling-Enabled Construction: Trends, Gaps, and Future Research Directions},
  journal = {Journal of Trending Research},
  year = {2026},
  volume = {1},
  number = {1},
  pages = {69--81},
  doi = {10.69739/jtr.v1i1.1896}
}
Kostenlos starten →