Predictive model for construction labour productivity using hybrid feature selection and principal component analysis

• Author(s) / Creator(s)

  • Ebrahimi, Sara
  • Kazerooni, Matin
  • Sumati, Vuppuluri
  • Fayek, Aminah Robinson

• Construction labour productivity (CLP) is affected by numerous variables made up of subjective and objective factors. Thus, CLP modelling and prediction are a complex task, leading to high computational cost and the risk of overfitting of data. This paper proposes a predictive model for CLP by integrating hybrid feature selection (HFS), as a combination of filter and wrapper methods, with principal component analysis (PCA). This developed HFS-PCA method reduces the dimensionality and complexity of CLP data and obtains better prediction performance by identifying the most predictive factors. Identified factors are utilized as inputs for various classification methods to predict CLP. Finally, prediction errors of the classification methods with and without using the proposed HFS-PCA method are compared, and the most accurate classification method is selected to develop the CLP predictive model. Experimental results show that using HFS-PCA for CLP prediction leads to better performances compared with past studies.

• Date created

  2022-01-01

• Subjects / Keywords

  • Construction labour productivity prediction
  • Hybrid feature selection
  • Principal component analysis
  • Genetic algorithm
  • Support vector machine
  • ReliefF algorithm

• Type of Item

  Article (Draft / Submitted)

• DOI

  https://doi.org/10.7939/r3-595j-3s61

• License

  © Canadian Science Publishing