Mapping the state-of-the-art: bibliometric and systematic analysis of machine learning and AI-based cost and time prediction in construction

Machine Learning (ML) and Artificial Intelligence (AI) have emerged as powerful solutions with their ability to generate accurate predictions. Various ML prediction methods have been increasingly employed for estimating construction project cost and time. This study aims to reveal the patterns, trends, and key insights within the existing literature using ML for cost and time prediction in construction projects. This study provides a state-of-the-art review through a PRISMA-based systematic literature review of 303 studies by conducting bibliometrics and in-depth content analysis. The studies were analyzed based on document metadata (year, authors, keywords, journal, citations, references) and research specifics (project phase, project type, project location, dataset source, sample size, input variables, prediction method, prediction performance, stakeholder perspective). Findings showed a field dominated by cost-estimation studies on building and infrastructure projects, focusing on inception/feasibility and construction phases. The most frequent methods were ANN, LR, SVM and their variants, alongside increasing ensemble and hybrid models.