Analysis of Interrelated Distress Types Using Naïve Bayes Method for Provincial Highway Network in Pakistan

Highways play a vital role in a nation’s development. However, due to the focus of road authorities in some countries on the construction of new highways, existing highways are left unattended to. Less severe distresses that develop on these roads after short periods and can be remedied relatively cheaper are left to deteriorate further. This results in higher costs of repair. An understanding of how these distresses evolve would be useful for scheduling maintenance activities and reducing maintenance costs. In this study, the probability of occurrence of a particular distress type due to the presence of another distress types with time was studied by using a probabilistic algorithm known as Bayes classifier. The study was conducted by analyzing data from the provincial highway network of Khyber Pakhtunkhwa province of Pakistan. Data on 12 types of distresses that occurred in pavement aged up to 7 years were utilized in the study. The results showed the probability of occurrence of severe distress types like block cracks increased due to delay in treatment of less severe distress types like edge cutting, alligator cracks, potholes and transverse cracks highlighting the need for preventive maintenance.