Machine learning and technological mediation in cost estimation practice

Due to their complex architecture, methods based on machine learning (ML) lack transparency on how inputs are transformed into outputs. With the growing influence of ML in construction cost estimation, analysing this transformation that is in a great part unperceivable by humans becomes increasingly important. The ways in which technology transforms inputs into outputs are also the focus of the philosophy of technological mediation. The purpose of this study is to understand the mediating role that ML plays in construction cost estimation. An existing method for developing a ML cost estimation model for the tendering phase in construction engineering services was chosen as the subject of an exploratory case study. By applying concepts of technological intentionality, case-based reasoning and digital materials, the kind of technological mediation that ML technology provides between users of this technology and cost estimation in construction practice is discussed.