This paper presents a solid LSTM-based model for predicting the Pavement Riding Quality Index (RQI) on asphalt highways. The model design and training process are clear, and the paper is well-organized. However, its applicability is limited by a relatively homogeneous dataset (light traffic class, stable RQI), and the model’s generalization capacity in complex or high-stress pavement conditions remains untested. The paper also exceeds the 4-page limit and would benefit from trimming and expansion of the discussion section. Overall, it is a valuable contribution to predictive modeling in pavement management and well-suited for presentation at PFDM.