A Neural Network Model for Financial Performance Prediction: The Case for Road Works in Bahrain


  • Saraa Naseer Kadhim
  • Kadhim Raheim Erzaij


In construction projects, there are circumstances when contractors meet financial prequalification criteria but show low financial performance in practice. These cases in Bahraini road works add up to complexity in the contractor selection process. Thus, this study considers data from 72 most recent road works contract projects? in Bahrain. Each has the contract amount and the contractor's financial capability record. The use of covariance between these records variable through the Principal Component Analysis reduces them into manageable variables. The resulting variables used to train an Artificial Neural Network ANN to construct criteria-performance mapping. The ANN finds the nonlinear correlation between the FP and contractors' capabilities. The ANN model predicts the FP so decision-makers can efficiently evaluate bidders in the prequalification phase. Then the sensitivity analysis help detects the FP change that corresponds to changes in capabilities. The research findings from the Bahraini case that improvement in some financial capabilities criteria does not reflect equally on the performance of contractors of varying grades.