Bidding Decision-Support Model for Construction Projects Based on Projection Pursuit Learning Method

Abstract

The decision of whether or not to bid is a crucial task for contractors in the construction industry because multiple attributes with the features of uncertainty and ambiguity can enhance the complexity and difficulty of decision-making. Meanwhile, traditional methods for bidding decisions primarily depend on experience and the subjective evaluation of decision-makers, and it is difficult to develop a more reliable model through data-driven techniques when the historical bidding data is limited. Therefore, this paper proposes a bidding decision-support model for construction projects using the projection pursuit learning method (PPLM), which can explain the characteristic rules in the bidding of contractors through the digging of information of previous bids and help one to make a reasonable and effective decision for bidding a new project. Finally, a case study is provided to demonstrate the applicability of the novel model. The results indicate that the method developed based on the training set can correctly classify 10 testing projects and has a better performance in fitting and predicting the accuracy of bidding decisions when compared with other approaches.