Advanced Mathematics is a very important foundational theory course in the undergraduate talent cultivation course system of universities. It possesses the ability to cultivate trainees’ abstract thinking processes, logical reasoning abilities, and computational thinking abilities. This present thesis deals with the problem of teaching transformation in Higher Mathematics. This paper carries out deep discussion on the flipped classroom teaching method which is used for Advanced Mathematics course. This text introduces KBRL-Rec, which is an on-line examination question recommendation model that is established on the basis of reinforcement learning. On the foundation of the KBRL-Rec model, a precise teaching intervention framework for the flipped classroom of Advanced Mathematics is built, and an empirical exploration on the effectiveness of the teaching intervention is conducted. The research results show that when make comparison with the random recommendation method, the relevance degree of the exercises that recommended by the KBRL – Rec algorithm increases from 0.0084 to 0.0297. The undulation of difficulty degree descends from 0.2136 to 0.0325. Furthermore, the value of reward becomes more stable, hence the maximum value is comparatively more favorable. The utilization of the constructed generalized frame for accuracy education intervention to carry out accuracy education intervention has obtained better results and shows great application possibility. This present article provides a model for the teaching of Advanced Mathematics by making use of the flipped classroom teaching model, which therefore helps the realization of intelligent and dynamic accurate teaching intervention.
