Development of Predictive Model for the Number of Potholes Using Multi Regression Analysis

Recently, the number of potholes in the principal cities of korea has been increased because of the rapid climate change and the city population concentration. This study is to predict the number of potholes occurred in the Seoul metropolitan area. The prospective models for the pothole can help preventing the risk factors for the pothole. To finding the effect elements for the pothole occurrence, the empirical approach regarding the road engineering and analytical approach with stochastic analysis were used. The design of experiment regarding the surface response method was used to constitute the predictive model. The primary factors were utilized in the constitutive equation model using the multi regression analysis. The empirically predictive model makes little odds in p-value, coefficient of determination value and root mean squared error compared to the analytically predictive model. But, the analytical model has a lower incidence of error than the empirical model when the MPE(Mean Percentage of Error) analysis related to the max- imum and minimum MPE was conducted.