Cost-effective assessment of in-service asphalt pavement condition based on Random Forests and regression analysis

The condition assessment of in-service asphalt pavement plays a key role in pavement maintenance and reha-bilitation. Driven by historical data, the Random Forests algorithm with the Gini importance measurement was employed to identify and extract key detection indicators from the pavement assessment standards. On top of that, a cost-effective method for pavement condition assessment based on the key detection indicators was proposed by decreasing unnecessary data dimensions. A comparison between the proposed method and the traditional method has been made to verify the feasibility of pavement condition assessment. The results show that the pavement assessment results based on the proposed method matched well with those based on the traditional method, which achieved a more than 90% consistency of overall assessment results in validation samples. Hence, they demonstrated that the proposed method utilized fewer pavement detection indicators to reduce the burden of data collection and improve the cost-effectiveness of pavement condition assessment tasks. In the future, it will be a promising alternative to assist pavement maintenance and rehabilitation.