Comparing the Random Forest with the Generalized Additive Model to Evaluate the Impacts of Outdoor Ambient Environmental Factors on Scaffolding Construction Productivity

Xin Liu, Yongze Song, Wen Yi, Xiangyu Wang, Junxiang Zhu

Research output: Journal article publicationJournal articleAcademic researchpeer-review

17 Citations (Scopus)

Abstract

The improvement of construction productivity has always been a key concern for both researchers and project managers. Several studies have analyzed construction productivity from different perspectives; however, little research has been conducted to evaluate the impact of outdoor ambient environmental factors on construction productivity, especially at the project level. Therefore, to assess such impacts, a nonparametric regression model - the generalized additive model (GAM) - and a nonlinear machine learning model - random forest (RF) - are comparatively used to assess these contributors on the scaffolding construction performance factor (PF). The meteorological variables used in this study include temperature, humidity, ambient pressure, wind speed and wind direction, specific weather event (clear day, fog, rain, or thunderstorm), and the ultraviolet (UV) index. Results demonstrate that the joint meteorological factors play a key role in construction PF variation, with contribution ranging from 32.50% (GAM) to 59.41% (RF). The better performance of RF and GAM shows that the relationship between outdoor ambient environment and construction productivity is nonlinear and should be built by nonlinear models.

Original languageEnglish
Article number04018037
JournalJournal of Construction Engineering and Management
Volume144
Issue number6
DOIs
Publication statusPublished - 1 Jun 2018

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Building and Construction
  • Industrial relations
  • Strategy and Management

Cite this