Corporate Misconduct Prediction with Support Vector Machine in the Construction Industry

Ran Wang, Chia Jung Lee, Shu-Chien Hsu, Cheng Yu Lee

Research output: Journal article publicationJournal articleAcademic researchpeer-review

12 Citations (Scopus)


Corporate misconduct may lead to severe economic loss and even fatal injuries to workers and residents in the construction industry. Previous studies have proven that board composition in organizations can be related to illegal business behaviors. By analyzing board composition data from 45 publicly listed construction companies in Taiwan, this paper provides a tool for predicting corporate misconduct (CM). A support vector machine (SVM) was used to construct such a prediction model, and a logistic regression model was used as a benchmark to assess the performance of the established SVM model. The established SVM model achieved an accuracy rate of 72.22% for predicting the occurrence of CM when applied to all observations in the sample, with a rate of 90% accuracy in predicting misconduct by companies found guilty of doing so in the sample, thus performing better than the logistic regression model. The developed model yields new insights on previous research and can guide stakeholders to reduce the risk of illegal business acts occurring in the construction industry.
Original languageEnglish
Article number04018021
JournalJournal of Management in Engineering
Issue number4
Publication statusPublished - 1 Jul 2018


  • Board composition
  • Corporate misconduct (CM)
  • Support vector machine (SVM)

ASJC Scopus subject areas

  • Industrial relations
  • General Engineering
  • Strategy and Management
  • Management Science and Operations Research


Dive into the research topics of 'Corporate Misconduct Prediction with Support Vector Machine in the Construction Industry'. Together they form a unique fingerprint.

Cite this