Advertisement evaluation system based on interval exploration

The invention provides an advertisement evaluation system based on interval exploration, and the system comprises the steps: firstly carrying out the interval division of advertisement value estimation data for an advertisement through an interval exploration module, thereby forming the advertisement value estimation data of a plurality of interval nodes, and further, carrying out the calculation of the advertisement value estimation data; a reinforcement learning module is utilized to construct a value award function, group actions are defined based on advertisement value estimation data after interval exploration, and then an action value function is constructed to select an advertisement evaluation action with the maximum value from the group actions. The advertisement evaluation action with the highest cost performance is found through the advertisement value exploration carried out in the interval exploration by the system provided by the invention, so that the overall evaluation efficiency of the advert