Using machine learning to analyze and predict construction task productivity

Abstract

The factors that affect productivity are a major focus in construction. This article proposes a machine learning–based approach to predict task productivity by using a subjective measure (compatibility of personality), together with external and site conditions, and other workers' characteristics. The approach integrates K‐nearest neighbor (KNN), deep neural network (DNN), logistic regression, support vector machine (SVM), and ResNet18 to discover the mapping between input and output variables, alongside rigorous statistical analyses to interpret data. A database including 1977 productivity measures is utilized to train, test, and validate the approach. Results test rules in the masonry industry, which do not seem to have been tested before: Small crews are more productive than large crews; higher compatibility results in higher productivity in easy but not in difficult tasks; the relevance of experience to task productivity may depend on the difficulty of the task.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 31 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.