Construction claims prediction using ANN models: a case study of the Indian construction industry

Abstract

Claims are a major concern in every construction project. With the increasing complexity and size of the construction projects in India, the number and frequency of claims, also bound to increase, which adversely affects the construction environment. So, an early prediction of these claims is crucial to avoid their costly negative impacts. This paper aimed to develop artificial neural network models to predict the frequency of claims in construction projects. Based on a comprehensive literature review, a total of 39 factors causing claims were identified first, and then refined by Delphi interview with 10 experts. Subsequently, a questionnaire was further developed and disseminated to different owners, contractors, and consultants’ organizations working in the Indian construction sector, which received 206 valid replies. The questionnaire data were then utilized to develop and validate the construction claims frequency prediction models. For this, an artificial neural network approach was used to construct the models. These models will help the construction professionals in predicting the frequency of occurrence for different types of claims throughout the course of the project and enable them to select the best strategy available to avoid or mitigate their occurrence at an early stage.